Hello, R friends
My student unearthed this quirk that might interest you. I wondered if this might be a bug in the R interpreter. If not a bug, it certainly stands as a good example of the dangers of floating point numbers in computing. What do you think? > 100*(23/40) [1] 57.5 > (100*23)/40 [1] 57.5 > round(100*(23/40)) [1] 57 > round((100*23)/40) [1] 58 The result in the 2 rounds should be the same, I think. Clearly some digital number devil is at work. I *guess* that when you put in whole numbers and group them like this (100*23), the interpreter does integer math, but if you group (23/40), you force a fractional division and a floating point number. The results from the first 2 calculations are not actually 57.5, they just appear that way. Before you close the books, look at this: > aa <- 100*(23/40) > bb <- (100*23)/40 > all.equal(aa,bb) [1] TRUE > round(aa) [1] 57 > round(bb) [1] 58 I'm putting this one in my collection of "difficult to understand" numerical calculations. If you have seen this before, I'm sorry to waste your time. pj -- Paul E. Johnson http://pj.freefaculty.org Director, Center for Research Methods and Data Analysis http://crmda.ku.edu To write to me directly, please address me at pauljohn at ku.edu. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
This is FAQ 7.31. It is not a bug, it is the unavoidable problem of accurately representing floating point numbers with a finite number of bits of precision. Look at the following:
> a <- 100*(23/40) > b <- (100*23)/40 > print(a,digits=20) [1] 57.499999999999993 > print(b,digits=20) [1] 57.5 > Your example with all.equal evaluates TRUE because all.equal uses a 'fuzz factor'. From the all.equal man page "all.equal(x, y) is a utility to compare R objects x and y testing 'near equality'." Hope this is helpful, Dan Daniel Nordlund, PhD Research and Data Analysis Division Services & Enterprise Support Administration Washington State Department of Social and Health Services > -----Original Message----- > From: R-help [mailto:[hidden email]] On Behalf Of Paul > Johnson > Sent: Thursday, April 20, 2017 2:56 PM > To: R-help > Subject: [R] Interesting quirk with fractions and rounding > > Hello, R friends > > My student unearthed this quirk that might interest you. > > I wondered if this might be a bug in the R interpreter. If not a bug, > it certainly stands as a good example of the dangers of floating point > numbers in computing. > > What do you think? > > > 100*(23/40) > [1] 57.5 > > (100*23)/40 > [1] 57.5 > > round(100*(23/40)) > [1] 57 > > round((100*23)/40) > [1] 58 > > The result in the 2 rounds should be the same, I think. Clearly some > digital number devil is at work. I *guess* that when you put in whole > numbers and group them like this (100*23), the interpreter does > integer math, but if you group (23/40), you force a fractional > division and a floating point number. The results from the first 2 > calculations are not actually 57.5, they just appear that way. > > Before you close the books, look at this: > > > aa <- 100*(23/40) > > bb <- (100*23)/40 > > all.equal(aa,bb) > [1] TRUE > > round(aa) > [1] 57 > > round(bb) > [1] 58 > > I'm putting this one in my collection of "difficult to understand" > numerical calculations. > > If you have seen this before, I'm sorry to waste your time. > > pj > -- > Paul E. Johnson http://pj.freefaculty.org > Director, Center for Research Methods and Data Analysis > http://crmda.ku.edu > > To write to me directly, please address me at pauljohn at ku.edu. > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting- > guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
I might add that things that *look* like integers in R are not really
integers, unless you explicitly label them as such: > str(20) num 20 > str(20.5) num 20.5 > str(20L) int 20 > I think that Python 2 will do integer arithmetic on things that look like integers: $ python2 . . . >>> 30 / 20 1 >>> But that behavior has changed in Python 3: $ python3 . . . >>> 30 / 20 1.5 >>> -- Mike On Thu, Apr 20, 2017 at 3:20 PM, Nordlund, Dan (DSHS/RDA) <[hidden email]> wrote: > This is FAQ 7.31. It is not a bug, it is the unavoidable problem of accurately representing floating point numbers with a finite number of bits of precision. Look at the following: > >> a <- 100*(23/40) >> b <- (100*23)/40 >> print(a,digits=20) > [1] 57.499999999999993 >> print(b,digits=20) > [1] 57.5 >> > > Your example with all.equal evaluates TRUE because all.equal uses a 'fuzz factor'. From the all.equal man page > > "all.equal(x, y) is a utility to compare R objects x and y testing 'near equality'." > > > Hope this is helpful, > > Dan > > Daniel Nordlund, PhD > Research and Data Analysis Division > Services & Enterprise Support Administration > Washington State Department of Social and Health Services > > >> -----Original Message----- >> From: R-help [mailto:[hidden email]] On Behalf Of Paul >> Johnson >> Sent: Thursday, April 20, 2017 2:56 PM >> To: R-help >> Subject: [R] Interesting quirk with fractions and rounding >> >> Hello, R friends >> >> My student unearthed this quirk that might interest you. >> >> I wondered if this might be a bug in the R interpreter. If not a bug, >> it certainly stands as a good example of the dangers of floating point >> numbers in computing. >> >> What do you think? >> >> > 100*(23/40) >> [1] 57.5 >> > (100*23)/40 >> [1] 57.5 >> > round(100*(23/40)) >> [1] 57 >> > round((100*23)/40) >> [1] 58 >> >> The result in the 2 rounds should be the same, I think. Clearly some >> digital number devil is at work. I *guess* that when you put in whole >> numbers and group them like this (100*23), the interpreter does >> integer math, but if you group (23/40), you force a fractional >> division and a floating point number. The results from the first 2 >> calculations are not actually 57.5, they just appear that way. >> >> Before you close the books, look at this: >> >> > aa <- 100*(23/40) >> > bb <- (100*23)/40 >> > all.equal(aa,bb) >> [1] TRUE >> > round(aa) >> [1] 57 >> > round(bb) >> [1] 58 >> >> I'm putting this one in my collection of "difficult to understand" >> numerical calculations. >> >> If you have seen this before, I'm sorry to waste your time. >> >> pj >> -- >> Paul E. Johnson http://pj.freefaculty.org >> Director, Center for Research Methods and Data Analysis >> http://crmda.ku.edu >> >> To write to me directly, please address me at pauljohn at ku.edu. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/posting- >> guide.html >> and provide commented, minimal, self-contained, reproducible code. > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
Use all.equal(tolerance=0, aa, bb) to check for exact equality:
> aa <- 100*(23/40) > bb <- (100*23)/40 > all.equal(aa,bb) [1] TRUE > all.equal(aa,bb,tolerance=0) [1] "Mean relative difference: 1.235726e-16" > aa < bb [1] TRUE The numbers there are rounded to 52 binary digits (16+ decimal digits) for storage and rounding is not a linear or associative operation. Think of doing arithmetic by hand where you store all numbers, include intermediate results, with only 2 significant decimal digits: (3 * 1) / 3 -> 3 / 3 -> 1 3 * (1/3) -> 3 * 0.33 -> 0.99 Bill Dunlap TIBCO Software wdunlap tibco.com On Thu, Apr 20, 2017 at 2:56 PM, Paul Johnson <[hidden email]> wrote: > Hello, R friends > > My student unearthed this quirk that might interest you. > > I wondered if this might be a bug in the R interpreter. If not a bug, > it certainly stands as a good example of the dangers of floating point > numbers in computing. > > What do you think? > > > 100*(23/40) > [1] 57.5 > > (100*23)/40 > [1] 57.5 > > round(100*(23/40)) > [1] 57 > > round((100*23)/40) > [1] 58 > > The result in the 2 rounds should be the same, I think. Clearly some > digital number devil is at work. I *guess* that when you put in whole > numbers and group them like this (100*23), the interpreter does > integer math, but if you group (23/40), you force a fractional > division and a floating point number. The results from the first 2 > calculations are not actually 57.5, they just appear that way. > > Before you close the books, look at this: > > > aa <- 100*(23/40) > > bb <- (100*23)/40 > > all.equal(aa,bb) > [1] TRUE > > round(aa) > [1] 57 > > round(bb) > [1] 58 > > I'm putting this one in my collection of "difficult to understand" > numerical calculations. > > If you have seen this before, I'm sorry to waste your time. > > pj > -- > Paul E. Johnson http://pj.freefaculty.org > Director, Center for Research Methods and Data Analysis > http://crmda.ku.edu > > To write to me directly, please address me at pauljohn at ku.edu. > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/ > posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > [[alternative HTML version deleted]] ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by Michael Hannon-2
Also note that we see the same thing in Ruby:
irb(main):001:0> 100*(23/40) => 0 irb(main):002:0> 100.0*(23.0/40.0) => 57.49999999999999 irb(main):003:0> (100.0*23.0)/40.0 => 57.5 and in C: hpages@latitude:~$ cat test.c #include <stdio.h> main() { printf("%.15f\n", 100.0 * (23.0 / 40.0)); printf("%.15f\n", (100.0 * 23.0) / 40.0); } hpages@latitude:~$ gcc test.c hpages@latitude:~$ ./a.out 57.499999999999993 57.500000000000000 These rounding errors are intrinsically related to how computers do floating point arithmetic. H. On 04/20/2017 03:34 PM, Michael Hannon wrote: > I might add that things that *look* like integers in R are not really > integers, unless you explicitly label them as such: > >> str(20) > num 20 > >> str(20.5) > num 20.5 > >> str(20L) > int 20 >> > > I think that Python 2 will do integer arithmetic on things that look > like integers: > > $ python2 > . > . > . >>>> 30 / 20 > 1 >>>> > > But that behavior has changed in Python 3: > > $ python3 > . > . > . >>>> 30 / 20 > 1.5 >>>> > > -- Mike > > > On Thu, Apr 20, 2017 at 3:20 PM, Nordlund, Dan (DSHS/RDA) > <[hidden email]> wrote: >> This is FAQ 7.31. It is not a bug, it is the unavoidable problem of accurately representing floating point numbers with a finite number of bits of precision. Look at the following: >> >>> a <- 100*(23/40) >>> b <- (100*23)/40 >>> print(a,digits=20) >> [1] 57.499999999999993 >>> print(b,digits=20) >> [1] 57.5 >>> >> >> Your example with all.equal evaluates TRUE because all.equal uses a 'fuzz factor'. From the all.equal man page >> >> "all.equal(x, y) is a utility to compare R objects x and y testing 'near equality'." >> >> >> Hope this is helpful, >> >> Dan >> >> Daniel Nordlund, PhD >> Research and Data Analysis Division >> Services & Enterprise Support Administration >> Washington State Department of Social and Health Services >> >> >>> -----Original Message----- >>> From: R-help [mailto:[hidden email]] On Behalf Of Paul >>> Johnson >>> Sent: Thursday, April 20, 2017 2:56 PM >>> To: R-help >>> Subject: [R] Interesting quirk with fractions and rounding >>> >>> Hello, R friends >>> >>> My student unearthed this quirk that might interest you. >>> >>> I wondered if this might be a bug in the R interpreter. If not a bug, >>> it certainly stands as a good example of the dangers of floating point >>> numbers in computing. >>> >>> What do you think? >>> >>>> 100*(23/40) >>> [1] 57.5 >>>> (100*23)/40 >>> [1] 57.5 >>>> round(100*(23/40)) >>> [1] 57 >>>> round((100*23)/40) >>> [1] 58 >>> >>> The result in the 2 rounds should be the same, I think. Clearly some >>> digital number devil is at work. I *guess* that when you put in whole >>> numbers and group them like this (100*23), the interpreter does >>> integer math, but if you group (23/40), you force a fractional >>> division and a floating point number. The results from the first 2 >>> calculations are not actually 57.5, they just appear that way. >>> >>> Before you close the books, look at this: >>> >>>> aa <- 100*(23/40) >>>> bb <- (100*23)/40 >>>> all.equal(aa,bb) >>> [1] TRUE >>>> round(aa) >>> [1] 57 >>>> round(bb) >>> [1] 58 >>> >>> I'm putting this one in my collection of "difficult to understand" >>> numerical calculations. >>> >>> If you have seen this before, I'm sorry to waste your time. >>> >>> pj >>> -- >>> Paul E. Johnson https://urldefense.proofpoint.com/v2/url?u=http-3A__pj.freefaculty.org&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=mCvk8xBulYqnlArSRFzabdeCvvmH9UFJ0kxrxsNC0SE&e= >>> Director, Center for Research Methods and Data Analysis >>> https://urldefense.proofpoint.com/v2/url?u=http-3A__crmda.ku.edu&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=rtQq4X4vty_dsRAWEsD_-ZZh7UDdnlLKyBllAl3i5eY&e= >>> >>> To write to me directly, please address me at pauljohn at ku.edu. >>> >>> ______________________________________________ >>> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >>> https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=efyHNH41uiy_mo_2RYV1aQU3sMi2yKoAfK7LMdoK5eM&e= >>> PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_posting-2D&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=ZmY65_nfgCDuDwZFHVL0QcFn1ttVyyj8UY4XADgtDm0&e= >>> guide.html >>> and provide commented, minimal, self-contained, reproducible code. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=efyHNH41uiy_mo_2RYV1aQU3sMi2yKoAfK7LMdoK5eM&e= >> PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_posting-2Dguide.html&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=0CaeomvoF9oO4LmfDFSjbfMQn4TufgUA-mqqIP7VFno&e= >> and provide commented, minimal, self-contained, reproducible code. > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=efyHNH41uiy_mo_2RYV1aQU3sMi2yKoAfK7LMdoK5eM&e= > PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_posting-2Dguide.html&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=h2uDuHofumYjJ48NvOaART10-esHowpTtp37cGtD3yQ&s=0CaeomvoF9oO4LmfDFSjbfMQn4TufgUA-mqqIP7VFno&e= > and provide commented, minimal, self-contained, reproducible code. > -- Hervé Pagès Program in Computational Biology Division of Public Health Sciences Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N, M1-B514 P.O. Box 19024 Seattle, WA 98109-1024 E-mail: [hidden email] Phone: (206) 667-5791 Fax: (206) 667-1319 ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
Hi
The problem is that people using Excel or probably other such spreadsheets do not encounter this behaviour as Excel silently rounds all your calculations and makes approximate comparison without telling it does so. Therefore most people usually do not have any knowledge of floating point numbers representation. Cheers Petr -----Original Message----- From: R-help [mailto:[hidden email]] On Behalf Of Paul Johnson Sent: Thursday, April 20, 2017 11:56 PM To: R-help <[hidden email]> Subject: [R] Interesting quirk with fractions and rounding Hello, R friends My student unearthed this quirk that might interest you. I wondered if this might be a bug in the R interpreter. If not a bug, it certainly stands as a good example of the dangers of floating point numbers in computing. What do you think? > 100*(23/40) [1] 57.5 > (100*23)/40 [1] 57.5 > round(100*(23/40)) [1] 57 > round((100*23)/40) [1] 58 The result in the 2 rounds should be the same, I think. Clearly some digital number devil is at work. I *guess* that when you put in whole numbers and group them like this (100*23), the interpreter does integer math, but if you group (23/40), you force a fractional division and a floating point number. The results from the first 2 calculations are not actually 57.5, they just appear that way. Before you close the books, look at this: > aa <- 100*(23/40) > bb <- (100*23)/40 > all.equal(aa,bb) [1] TRUE > round(aa) [1] 57 > round(bb) [1] 58 I'm putting this one in my collection of "difficult to understand" numerical calculations. If you have seen this before, I'm sorry to waste your time. pj -- Paul E. Johnson http://pj.freefaculty.org Director, Center for Research Methods and Data Analysis http://crmda.ku.edu To write to me directly, please address me at pauljohn at ku.edu. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. ________________________________ Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou určeny pouze jeho adresátům. Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie vymažte ze svého systému. Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi či zpožděním přenosu e-mailu. V případě, že je tento e-mail součástí obchodního jednání: - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany příjemce s dodatkem či odchylkou. - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve výslovným dosažením shody na všech jejích náležitostech. - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich existence je adresátovi či osobě jím zastoupené známá. This e-mail and any documents attached to it may be confidential and are intended only for its intended recipients. If you received this e-mail by mistake, please immediately inform its sender. Delete the contents of this e-mail with all attachments and its copies from your system. If you are not the intended recipient of this e-mail, you are not authorized to use, disseminate, copy or disclose this e-mail in any manner. The sender of this e-mail shall not be liable for any possible damage caused by modifications of the e-mail or by delay with transfer of the email. In case that this e-mail forms part of business dealings: - the sender reserves the right to end negotiations about entering into a contract in any time, for any reason, and without stating any reasoning. - if the e-mail contains an offer, the recipient is entitled to immediately accept such offer; The sender of this e-mail (offer) excludes any acceptance of the offer on the part of the recipient containing any amendment or variation. - the sender insists on that the respective contract is concluded only upon an express mutual agreement on all its aspects. - the sender of this e-mail informs that he/she is not authorized to enter into any contracts on behalf of the company except for cases in which he/she is expressly authorized to do so in writing, and such authorization or power of attorney is submitted to the recipient or the person represented by the recipient, or the existence of such authorization is known to the recipient of the person represented by the recipient. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
We all agree it is a problem with digital computing, not unique to R. I
don't think that is the right place to stop. What to do? The round example arose in a real funded project where 2 R programs differed in results and cause was that one person got 57 and another got 58. The explanation was found, but its less clear how to prevent similar in future. Guidelines, anyone? So far, these are my guidelines. 1. Insert L on numbers to signal that you really mean INTEGER. In R, forgetting the L in a single number will usually promote whole calculation to floats. 2. S3 variables are called 'numeric' if they are integer or double storage. So avoid "is.numeric" and prefer "is.double". 3. == is a total fail on floats 4. Run print with digits=20 so we can see the less rounded number. Perhaps start sessions with "options(digits=20)" 5. all.equal does what it promises, but one must be cautious. Are there math habits we should follow? For example, Is it generally true in R that (100*x)/y is more accurate than 100*(x/y), if x > y? (If that is generally true, couldn't the R interpreter do it for the user?) I've seen this problem before. In later editions of the game theory program Gambit, extraordinary effort was taken to keep values symbolically as integers as long as possible. Avoid division until the last steps. Same in Swarm simulations. Gary Polhill wrote an essay about the Ghost in the Machine along those lines, showing accidents from trusting floats. I wonder now if all uses of > or < with numeric variables are suspect. Oh well. If everybody posts their advice, I will write a summary. Paul Johnson University of Kansas On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > Hi > > The problem is that people using Excel or probably other such spreadsheets > do not encounter this behaviour as Excel silently rounds all your > calculations and makes approximate comparison without telling it does so. > Therefore most people usually do not have any knowledge of floating point > numbers representation. > > Cheers > Petr > > -----Original Message----- > From: R-help [mailto:[hidden email]] On Behalf Of Paul > Johnson > Sent: Thursday, April 20, 2017 11:56 PM > To: R-help <[hidden email]> > Subject: [R] Interesting quirk with fractions and rounding > > Hello, R friends > > My student unearthed this quirk that might interest you. > > I wondered if this might be a bug in the R interpreter. If not a bug, it > certainly stands as a good example of the dangers of floating point numbers > in computing. > > What do you think? > > > 100*(23/40) > [1] 57.5 > > (100*23)/40 > [1] 57.5 > > round(100*(23/40)) > [1] 57 > > round((100*23)/40) > [1] 58 > > The result in the 2 rounds should be the same, I think. Clearly some > digital number devil is at work. I *guess* that when you put in whole > numbers and group them like this (100*23), the interpreter does integer > math, but if you group (23/40), you force a fractional division and a > floating point number. The results from the first 2 calculations are not > actually 57.5, they just appear that way. > > Before you close the books, look at this: > > > aa <- 100*(23/40) > > bb <- (100*23)/40 > > all.equal(aa,bb) > [1] TRUE > > round(aa) > [1] 57 > > round(bb) > [1] 58 > > I'm putting this one in my collection of "difficult to understand" > numerical calculations. > > If you have seen this before, I'm sorry to waste your time. > > pj > -- > Paul E. Johnson http://pj.freefaculty.org > Director, Center for Research Methods and Data Analysis > http://crmda.ku.edu > > To write to me directly, please address me at pauljohn at ku.edu. > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/ > posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > > ________________________________ > Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou > určeny pouze jeho adresátům. > Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě > neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie > vymažte ze svého systému. > Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email > jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. > Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi > či zpožděním přenosu e-mailu. > > V případě, že je tento e-mail součástí obchodního jednání: > - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření > smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. > - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; > Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany > příjemce s dodatkem či odchylkou. > - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve > výslovným dosažením shody na všech jejích náležitostech. > - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za > společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn > nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto > emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich > existence je adresátovi či osobě jím zastoupené známá. > > This e-mail and any documents attached to it may be confidential and are > intended only for its intended recipients. > If you received this e-mail by mistake, please immediately inform its > sender. Delete the contents of this e-mail with all attachments and its > copies from your system. > If you are not the intended recipient of this e-mail, you are not > authorized to use, disseminate, copy or disclose this e-mail in any manner. > The sender of this e-mail shall not be liable for any possible damage > caused by modifications of the e-mail or by delay with transfer of the > email. > > In case that this e-mail forms part of business dealings: > - the sender reserves the right to end negotiations about entering into a > contract in any time, for any reason, and without stating any reasoning. > - if the e-mail contains an offer, the recipient is entitled to > immediately accept such offer; The sender of this e-mail (offer) excludes > any acceptance of the offer on the part of the recipient containing any > amendment or variation. > - the sender insists on that the respective contract is concluded only > upon an express mutual agreement on all its aspects. > - the sender of this e-mail informs that he/she is not authorized to enter > into any contracts on behalf of the company except for cases in which > he/she is expressly authorized to do so in writing, and such authorization > or power of attorney is submitted to the recipient or the person > represented by the recipient, or the existence of such authorization is > known to the recipient of the person represented by the recipient. > [[alternative HTML version deleted]] ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
A good part of the problem in the specific case you initially presented
is that some non-integer numbers have an exact representation in the binary floating point arithmetic being used. Basically, if the fractional part is of the form 1/2^k for some integer k > 0, there is an exact representation in the binary floating point scheme. > options(digits=20) > (100*23)/40 [1] 57.5 > 100*(23/40) [1] 57.499999999999992895 So the two operations give a slightly different result because the fractional part of the division of 100*23 by 40 is 0.5. So the first operations gives, exactly, 57.5 while the second operation does not because 23/40 has no exact representation. But, change the example's divisor from 40 to 30 [the fractional part from 1/2 to 2/3]: > (100*23)/30 [1] 76.666666666666671404 > 100*(23/30) [1] 76.666666666666671404 Now the two operations give the same answer to the full precision available. So, it isn't "generally true true in R that (100*x)/y is more accurate than 100*(x/y), if x > y." The key (in your example) is a property of the way that floating point arithmetic is implemented. ---JRG On 04/21/2017 08:19 AM, Paul Johnson wrote: > We all agree it is a problem with digital computing, not unique to R. I > don't think that is the right place to stop. > > What to do? The round example arose in a real funded project where 2 R > programs differed in results and cause was that one person got 57 and > another got 58. The explanation was found, but its less clear how to > prevent similar in future. Guidelines, anyone? > > So far, these are my guidelines. > > 1. Insert L on numbers to signal that you really mean INTEGER. In R, > forgetting the L in a single number will usually promote whole calculation > to floats. > 2. S3 variables are called 'numeric' if they are integer or double storage. > So avoid "is.numeric" and prefer "is.double". > 3. == is a total fail on floats > 4. Run print with digits=20 so we can see the less rounded number. Perhaps > start sessions with "options(digits=20)" > 5. all.equal does what it promises, but one must be cautious. > > Are there math habits we should follow? > > For example, Is it generally true in R that (100*x)/y is more accurate than > 100*(x/y), if x > y? (If that is generally true, couldn't the R > interpreter do it for the user?) > > I've seen this problem before. In later editions of the game theory program > Gambit, extraordinary effort was taken to keep values symbolically as > integers as long as possible. Avoid division until the last steps. Same in > Swarm simulations. Gary Polhill wrote an essay about the Ghost in the > Machine along those lines, showing accidents from trusting floats. > > I wonder now if all uses of > or < with numeric variables are suspect. > > Oh well. If everybody posts their advice, I will write a summary. > > Paul Johnson > University of Kansas > > On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > >> Hi >> >> The problem is that people using Excel or probably other such spreadsheets >> do not encounter this behaviour as Excel silently rounds all your >> calculations and makes approximate comparison without telling it does so. >> Therefore most people usually do not have any knowledge of floating point >> numbers representation. >> >> Cheers >> Petr >> >> -----Original Message----- >> From: R-help [mailto:[hidden email]] On Behalf Of Paul >> Johnson >> Sent: Thursday, April 20, 2017 11:56 PM >> To: R-help <[hidden email]> >> Subject: [R] Interesting quirk with fractions and rounding >> >> Hello, R friends >> >> My student unearthed this quirk that might interest you. >> >> I wondered if this might be a bug in the R interpreter. If not a bug, it >> certainly stands as a good example of the dangers of floating point numbers >> in computing. >> >> What do you think? >> >>> 100*(23/40) >> [1] 57.5 >>> (100*23)/40 >> [1] 57.5 >>> round(100*(23/40)) >> [1] 57 >>> round((100*23)/40) >> [1] 58 >> >> The result in the 2 rounds should be the same, I think. Clearly some >> digital number devil is at work. I *guess* that when you put in whole >> numbers and group them like this (100*23), the interpreter does integer >> math, but if you group (23/40), you force a fractional division and a >> floating point number. The results from the first 2 calculations are not >> actually 57.5, they just appear that way. >> >> Before you close the books, look at this: >> >>> aa <- 100*(23/40) >>> bb <- (100*23)/40 >>> all.equal(aa,bb) >> [1] TRUE >>> round(aa) >> [1] 57 >>> round(bb) >> [1] 58 >> >> I'm putting this one in my collection of "difficult to understand" >> numerical calculations. >> >> If you have seen this before, I'm sorry to waste your time. >> >> pj >> -- >> Paul E. Johnson http://pj.freefaculty.org >> Director, Center for Research Methods and Data Analysis >> http://crmda.ku.edu >> >> To write to me directly, please address me at pauljohn at ku.edu. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/ >> posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. >> >> ________________________________ >> Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou >> určeny pouze jeho adresátům. >> Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě >> neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie >> vymažte ze svého systému. >> Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email >> jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. >> Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi >> či zpožděním přenosu e-mailu. >> >> V případě, že je tento e-mail součástí obchodního jednání: >> - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření >> smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. >> - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; >> Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany >> příjemce s dodatkem či odchylkou. >> - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve >> výslovným dosažením shody na všech jejích náležitostech. >> - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za >> společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn >> nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto >> emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich >> existence je adresátovi či osobě jím zastoupené známá. >> >> This e-mail and any documents attached to it may be confidential and are >> intended only for its intended recipients. >> If you received this e-mail by mistake, please immediately inform its >> sender. Delete the contents of this e-mail with all attachments and its >> copies from your system. >> If you are not the intended recipient of this e-mail, you are not >> authorized to use, disseminate, copy or disclose this e-mail in any manner. >> The sender of this e-mail shall not be liable for any possible damage >> caused by modifications of the e-mail or by delay with transfer of the >> email. >> >> In case that this e-mail forms part of business dealings: >> - the sender reserves the right to end negotiations about entering into a >> contract in any time, for any reason, and without stating any reasoning. >> - if the e-mail contains an offer, the recipient is entitled to >> immediately accept such offer; The sender of this e-mail (offer) excludes >> any acceptance of the offer on the part of the recipient containing any >> amendment or variation. >> - the sender insists on that the respective contract is concluded only >> upon an express mutual agreement on all its aspects. >> - the sender of this e-mail informs that he/she is not authorized to enter >> into any contracts on behalf of the company except for cases in which >> he/she is expressly authorized to do so in writing, and such authorization >> or power of attorney is submitted to the recipient or the person >> represented by the recipient, or the existence of such authorization is >> known to the recipient of the person represented by the recipient. >> > > [[alternative HTML version deleted]] > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
I suggest you read some basic books on numerical analysis and/or talk
with a numerical analyst. You are (like most of us) an amateur at this sort of thing trying to reinvent wheels. If you are concerned with details, talk with experts. Don't assume what you don't know. This list is *not* a reliable source of such expertise, although there *are* individuals in the R universe with considerable knowledge who may or may not choose to respond. Cheers, Bert Bert Gunter "The trouble with having an open mind is that people keep coming along and sticking things into it." -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Fri, Apr 21, 2017 at 5:19 AM, Paul Johnson <[hidden email]> wrote: > We all agree it is a problem with digital computing, not unique to R. I > don't think that is the right place to stop. > > What to do? The round example arose in a real funded project where 2 R > programs differed in results and cause was that one person got 57 and > another got 58. The explanation was found, but its less clear how to > prevent similar in future. Guidelines, anyone? > > So far, these are my guidelines. > > 1. Insert L on numbers to signal that you really mean INTEGER. In R, > forgetting the L in a single number will usually promote whole calculation > to floats. > 2. S3 variables are called 'numeric' if they are integer or double storage. > So avoid "is.numeric" and prefer "is.double". > 3. == is a total fail on floats > 4. Run print with digits=20 so we can see the less rounded number. Perhaps > start sessions with "options(digits=20)" > 5. all.equal does what it promises, but one must be cautious. > > Are there math habits we should follow? > > For example, Is it generally true in R that (100*x)/y is more accurate than > 100*(x/y), if x > y? (If that is generally true, couldn't the R > interpreter do it for the user?) > > I've seen this problem before. In later editions of the game theory program > Gambit, extraordinary effort was taken to keep values symbolically as > integers as long as possible. Avoid division until the last steps. Same in > Swarm simulations. Gary Polhill wrote an essay about the Ghost in the > Machine along those lines, showing accidents from trusting floats. > > I wonder now if all uses of > or < with numeric variables are suspect. > > Oh well. If everybody posts their advice, I will write a summary. > > Paul Johnson > University of Kansas > > On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > >> Hi >> >> The problem is that people using Excel or probably other such spreadsheets >> do not encounter this behaviour as Excel silently rounds all your >> calculations and makes approximate comparison without telling it does so. >> Therefore most people usually do not have any knowledge of floating point >> numbers representation. >> >> Cheers >> Petr >> >> -----Original Message----- >> From: R-help [mailto:[hidden email]] On Behalf Of Paul >> Johnson >> Sent: Thursday, April 20, 2017 11:56 PM >> To: R-help <[hidden email]> >> Subject: [R] Interesting quirk with fractions and rounding >> >> Hello, R friends >> >> My student unearthed this quirk that might interest you. >> >> I wondered if this might be a bug in the R interpreter. If not a bug, it >> certainly stands as a good example of the dangers of floating point numbers >> in computing. >> >> What do you think? >> >> > 100*(23/40) >> [1] 57.5 >> > (100*23)/40 >> [1] 57.5 >> > round(100*(23/40)) >> [1] 57 >> > round((100*23)/40) >> [1] 58 >> >> The result in the 2 rounds should be the same, I think. Clearly some >> digital number devil is at work. I *guess* that when you put in whole >> numbers and group them like this (100*23), the interpreter does integer >> math, but if you group (23/40), you force a fractional division and a >> floating point number. The results from the first 2 calculations are not >> actually 57.5, they just appear that way. >> >> Before you close the books, look at this: >> >> > aa <- 100*(23/40) >> > bb <- (100*23)/40 >> > all.equal(aa,bb) >> [1] TRUE >> > round(aa) >> [1] 57 >> > round(bb) >> [1] 58 >> >> I'm putting this one in my collection of "difficult to understand" >> numerical calculations. >> >> If you have seen this before, I'm sorry to waste your time. >> >> pj >> -- >> Paul E. Johnson http://pj.freefaculty.org >> Director, Center for Research Methods and Data Analysis >> http://crmda.ku.edu >> >> To write to me directly, please address me at pauljohn at ku.edu. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/ >> posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. >> >> ________________________________ >> Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou >> určeny pouze jeho adresátům. >> Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě >> neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie >> vymažte ze svého systému. >> Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email >> jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. >> Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi >> či zpožděním přenosu e-mailu. >> >> V případě, že je tento e-mail součástí obchodního jednání: >> - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření >> smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. >> - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; >> Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany >> příjemce s dodatkem či odchylkou. >> - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve >> výslovným dosažením shody na všech jejích náležitostech. >> - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za >> společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn >> nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto >> emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich >> existence je adresátovi či osobě jím zastoupené známá. >> >> This e-mail and any documents attached to it may be confidential and are >> intended only for its intended recipients. >> If you received this e-mail by mistake, please immediately inform its >> sender. Delete the contents of this e-mail with all attachments and its >> copies from your system. >> If you are not the intended recipient of this e-mail, you are not >> authorized to use, disseminate, copy or disclose this e-mail in any manner. >> The sender of this e-mail shall not be liable for any possible damage >> caused by modifications of the e-mail or by delay with transfer of the >> email. >> >> In case that this e-mail forms part of business dealings: >> - the sender reserves the right to end negotiations about entering into a >> contract in any time, for any reason, and without stating any reasoning. >> - if the e-mail contains an offer, the recipient is entitled to >> immediately accept such offer; The sender of this e-mail (offer) excludes >> any acceptance of the offer on the part of the recipient containing any >> amendment or variation. >> - the sender insists on that the respective contract is concluded only >> upon an express mutual agreement on all its aspects. >> - the sender of this e-mail informs that he/she is not authorized to enter >> into any contracts on behalf of the company except for cases in which >> he/she is expressly authorized to do so in writing, and such authorization >> or power of attorney is submitted to the recipient or the person >> represented by the recipient, or the existence of such authorization is >> known to the recipient of the person represented by the recipient. >> > > [[alternative HTML version deleted]] > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
The subject is messy. I vaguely remember learning this stuff on my
first numerical analysis course over 40 years ago. The classic reference material (much newer, only 25 years old) is: What Every Computer Scientist Should Know About Floating-Point Arithmetic, David Goldberg, ACM Computing Surveys, Vol 23, No 1, 1991. Available here: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.22.6768 George I suggest you read some basic books on numerical analysis and/or talk > with a numerical analyst. You are (like most of us) an amateur at this > sort of thing trying to reinvent wheels. If you are concerned with > details, talk with experts. Don't assume what you don't know. This > list is *not* a reliable source of such expertise, although there > *are* individuals in the R universe with considerable knowledge who > may or may not choose to respond. > Cheers, > Bert > Bert Gunter > "The trouble with having an open mind is that people keep coming along > and sticking things into it." > -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Fri, Apr 21, 2017 at 5:19 AM, Paul Johnson <pauljohn32 at gmail.com <https://stat.ethz.ch/mailman/listinfo/r-help>> wrote: >* We all agree it is a problem with digital computing, not unique to R. I *>* don't think that is the right place to stop. *>>* What to do? The round example arose in a real funded project where 2 R *>* programs differed in results and cause was that one person got 57 and *>* another got 58. The explanation was found, but its less clear how to *>* prevent similar in future. Guidelines, anyone? *>>* So far, these are my guidelines. *>>* 1. Insert L on numbers to signal that you really mean INTEGER. In R, *>* forgetting the L in a single number will usually promote whole calculation *>* to floats. *>* 2. S3 variables are called 'numeric' if they are integer or double storage. *>* So avoid "is.numeric" and prefer "is.double". *>* 3. == is a total fail on floats *>* 4. Run print with digits=20 so we can see the less rounded number. Perhaps *>* start sessions with "options(digits=20)" *>* 5. all.equal does what it promises, but one must be cautious. *>>* Are there math habits we should follow? *>>* For example, Is it generally true in R that (100*x)/y is more accurate than *>* 100*(x/y), if x > y? (If that is generally true, couldn't the R *>* interpreter do it for the user?) *>>* I've seen this problem before. In later editions of the game theory program *>* Gambit, extraordinary effort was taken to keep values symbolically as *>* integers as long as possible. Avoid division until the last steps. Same in *>* Swarm simulations. Gary Polhill wrote an essay about the Ghost in the *>* Machine along those lines, showing accidents from trusting floats. *>>* I wonder now if all uses of > or < with numeric variables are suspect. *>>* Oh well. If everybody posts their advice, I will write a summary. *>>* Paul Johnson *>* University of Kansas *>>* On Apr 21, 2017 12:02 AM, "PIKAL Petr" <petr.pikal at precheza.cz <https://stat.ethz.ch/mailman/listinfo/r-help>> wrote: *>>>* Hi *>>>>* The problem is that people using Excel or probably other such spreadsheets *>>* do not encounter this behaviour as Excel silently rounds all your *>>* calculations and makes approximate comparison without telling it does so. *>>* Therefore most people usually do not have any knowledge of floating point *>>* numbers representation. *>>>>* Cheers *>>* Petr *>>>>* -----Original Message----- *>>* From: R-help [mailto:r-help-bounces at r-project.org <https://stat.ethz.ch/mailman/listinfo/r-help>] On Behalf Of Paul *>>* Johnson *>>* Sent: Thursday, April 20, 2017 11:56 PM *>>* To: R-help <r-help at r-project.org <https://stat.ethz.ch/mailman/listinfo/r-help>> *>>* Subject: [R] Interesting quirk with fractions and rounding *>>>>* Hello, R friends *>>>>* My student unearthed this quirk that might interest you. *>>>>* I wondered if this might be a bug in the R interpreter. If not a bug, it *>>* certainly stands as a good example of the dangers of floating point numbers *>>* in computing. *>>>>* What do you think? *>>>>* > 100*(23/40) *>>* [1] 57.5 *>>* > (100*23)/40 *>>* [1] 57.5 *>>* > round(100*(23/40)) *>>* [1] 57 *>>* > round((100*23)/40) *>>* [1] 58 *>>>>* The result in the 2 rounds should be the same, I think. Clearly some *>>* digital number devil is at work. I *guess* that when you put in whole *>>* numbers and group them like this (100*23), the interpreter does integer *>>* math, but if you group (23/40), you force a fractional division and a *>>* floating point number. The results from the first 2 calculations are not *>>* actually 57.5, they just appear that way. *>>>>* Before you close the books, look at this: *>>>>* > aa <- 100*(23/40) *>>* > bb <- (100*23)/40 *>>* > all.equal(aa,bb) *>>* [1] TRUE *>>* > round(aa) *>>* [1] 57 *>>* > round(bb) *>>* [1] 58 *>>>>* I'm putting this one in my collection of "difficult to understand" *>>* numerical calculations. *>>>>* If you have seen this before, I'm sorry to waste your time. *>>>>* pj *>>* -- *>>* Paul E. Johnson http://pj.freefaculty.org <http://pj.freefaculty.org> *>>* Director, Center for Research Methods and Data Analysis *>>* http://crmda.ku.edu <http://crmda.ku.edu> *>>>>* To write to me directly, please address me at pauljohn at ku.edu <http://ku.edu>. *>>>>* ______________________________________________ *>>* R-help at r-project.org <https://stat.ethz.ch/mailman/listinfo/r-help> mailing list -- To UNSUBSCRIBE and more, see *>>* https://stat.ethz.ch/mailman/listinfo/r-help <https://stat.ethz.ch/mailman/listinfo/r-help> *>>* PLEASE do read the posting guide http://www.R-project.org/ <http://www.R-project.org/> *>>* posting-guide.html *>>* and provide commented, minimal, self-contained, reproducible code. *>>>>* ________________________________ *>>* Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou *>>* určeny pouze jeho adresátům. *>>* Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě *>>* neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie *>>* vymažte ze svého systému. *>>* Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email *>>* jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. *>>* Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi *>>* či zpožděním přenosu e-mailu. *>>>>* V případě, že je tento e-mail součástí obchodního jednání: *>>* - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření *>>* smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. *>>* - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; *>>* Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany *>>* příjemce s dodatkem či odchylkou. *>>* - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve *>>* výslovným dosažením shody na všech jejích náležitostech. *>>* - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za *>>* společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn *>>* nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto *>>* emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich *>>* existence je adresátovi či osobě jím zastoupené známá. *>>>>* This e-mail and any documents attached to it may be confidential and are *>>* intended only for its intended recipients. *>>* If you received this e-mail by mistake, please immediately inform its *>>* sender. Delete the contents of this e-mail with all attachments and its *>>* copies from your system. *>>* If you are not the intended recipient of this e-mail, you are not *>>* authorized to use, disseminate, copy or disclose this e-mail in any manner. *>>* The sender of this e-mail shall not be liable for any possible damage *>>* caused by modifications of the e-mail or by delay with transfer of the *>>* email. *>>>>* In case that this e-mail forms part of business dealings: *>>* - the sender reserves the right to end negotiations about entering into a *>>* contract in any time, for any reason, and without stating any reasoning. *>>* - if the e-mail contains an offer, the recipient is entitled to *>>* immediately accept such offer; The sender of this e-mail (offer) excludes *>>* any acceptance of the offer on the part of the recipient containing any *>>* amendment or variation. *>>* - the sender insists on that the respective contract is concluded only *>>* upon an express mutual agreement on all its aspects. *>>* - the sender of this e-mail informs that he/she is not authorized to enter *>>* into any contracts on behalf of the company except for cases in which *>>* he/she is expressly authorized to do so in writing, and such authorization *>>* or power of attorney is submitted to the recipient or the person *>>* represented by the recipient, or the existence of such authorization is *>>* known to the recipient of the person represented by the recipient. *>>> [[alternative HTML version deleted]] ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
Your guideline #1 is invalid for R... compare 5L/3L to 5L %/% 3L. If you want to avoid automatic conversion to double then you have to be cautious which operators/functions you apply to them... merely throwing in L everywhere is not going to help.
#2 refers to S3, but that is a completely different concept. I am not sure why you think you should "prefer is.double", since is.numeric is perfectly useful for everyday analysis. I think that numbers that represent input data can be numeric (either integer or double) but should be treated as though they were always approximate. That means don't spend time printing them out with 20 digits of precision except for illustration... just always assume that they could be slightly different than they look when printed. Which means (for example) that you have to be cautious about whether you want to ROUND or TRUNCATE when binning. If you do this you will stop feeling surprised by this kind of result. Of all groups of people who have to deal with floating point numbers, shouldn't statisticians be able to work effectively with approximate values? Numbers that represent internally-generated indexes and factors should be created as and treated like integers. All you have to do is remember which of these two categories each numeric vector has and you will tend to avoid abusing them and being surprised by the results. I don't think (100*x)/y is necessarily more accurate than 100*(x/y)... they just have different approximations in some cases. Each sub-expression has the same number of bits in the mantissa, so the uncertainty propagates the same in both cases. Of more concern is things like c1+c2*x+c3*x*x versus c1+x*(c2+c3*x), but these are numerical analysis concerns for which there are whole courses to explain the issue, and are not specific to R or on-topic here. -- Sent from my phone. Please excuse my brevity. On April 21, 2017 5:19:10 AM PDT, Paul Johnson <[hidden email]> wrote: >We all agree it is a problem with digital computing, not unique to R. I >don't think that is the right place to stop. > >What to do? The round example arose in a real funded project where 2 R >programs differed in results and cause was that one person got 57 and >another got 58. The explanation was found, but its less clear how to >prevent similar in future. Guidelines, anyone? > >So far, these are my guidelines. > >1. Insert L on numbers to signal that you really mean INTEGER. In R, >forgetting the L in a single number will usually promote whole >calculation >to floats. >2. S3 variables are called 'numeric' if they are integer or double >storage. >So avoid "is.numeric" and prefer "is.double". >3. == is a total fail on floats >4. Run print with digits=20 so we can see the less rounded number. >Perhaps >start sessions with "options(digits=20)" >5. all.equal does what it promises, but one must be cautious. > >Are there math habits we should follow? > >For example, Is it generally true in R that (100*x)/y is more accurate >than >100*(x/y), if x > y? (If that is generally true, couldn't the R >interpreter do it for the user?) > >I've seen this problem before. In later editions of the game theory >program >Gambit, extraordinary effort was taken to keep values symbolically as >integers as long as possible. Avoid division until the last steps. Same >in >Swarm simulations. Gary Polhill wrote an essay about the Ghost in the >Machine along those lines, showing accidents from trusting floats. > >I wonder now if all uses of > or < with numeric variables are suspect. > >Oh well. If everybody posts their advice, I will write a summary. > >Paul Johnson >University of Kansas > >On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > >> Hi >> >> The problem is that people using Excel or probably other such >spreadsheets >> do not encounter this behaviour as Excel silently rounds all your >> calculations and makes approximate comparison without telling it does >so. >> Therefore most people usually do not have any knowledge of floating >point >> numbers representation. >> >> Cheers >> Petr >> >> -----Original Message----- >> From: R-help [mailto:[hidden email]] On Behalf Of Paul >> Johnson >> Sent: Thursday, April 20, 2017 11:56 PM >> To: R-help <[hidden email]> >> Subject: [R] Interesting quirk with fractions and rounding >> >> Hello, R friends >> >> My student unearthed this quirk that might interest you. >> >> I wondered if this might be a bug in the R interpreter. If not a bug, >it >> certainly stands as a good example of the dangers of floating point >numbers >> in computing. >> >> What do you think? >> >> > 100*(23/40) >> [1] 57.5 >> > (100*23)/40 >> [1] 57.5 >> > round(100*(23/40)) >> [1] 57 >> > round((100*23)/40) >> [1] 58 >> >> The result in the 2 rounds should be the same, I think. Clearly some >> digital number devil is at work. I *guess* that when you put in whole >> numbers and group them like this (100*23), the interpreter does >integer >> math, but if you group (23/40), you force a fractional division and a >> floating point number. The results from the first 2 calculations are >not >> actually 57.5, they just appear that way. >> >> Before you close the books, look at this: >> >> > aa <- 100*(23/40) >> > bb <- (100*23)/40 >> > all.equal(aa,bb) >> [1] TRUE >> > round(aa) >> [1] 57 >> > round(bb) >> [1] 58 >> >> I'm putting this one in my collection of "difficult to understand" >> numerical calculations. >> >> If you have seen this before, I'm sorry to waste your time. >> >> pj >> -- >> Paul E. Johnson http://pj.freefaculty.org >> Director, Center for Research Methods and Data Analysis >> http://crmda.ku.edu >> >> To write to me directly, please address me at pauljohn at ku.edu. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/ >> posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. >> >> ________________________________ >> Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a >jsou >> určeny pouze jeho adresátům. >> Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě >> neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho >kopie >> vymažte ze svého systému. >> Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento >> jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. >> Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou >modifikacemi >> či zpožděním přenosu e-mailu. >> >> V případě, že je tento e-mail součástí obchodního jednání: >> - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření >> smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. >> - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně >přijmout; >> Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze >strany >> příjemce s dodatkem či odchylkou. >> - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve >> výslovným dosažením shody na všech jejích náležitostech. >> - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za >> společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně >zmocněn >> nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi >tohoto >> emailu případně osobě, kterou adresát zastupuje, předloženy nebo >jejich >> existence je adresátovi či osobě jím zastoupené známá. >> >> This e-mail and any documents attached to it may be confidential and >are >> intended only for its intended recipients. >> If you received this e-mail by mistake, please immediately inform its >> sender. Delete the contents of this e-mail with all attachments and >its >> copies from your system. >> If you are not the intended recipient of this e-mail, you are not >> authorized to use, disseminate, copy or disclose this e-mail in any >manner. >> The sender of this e-mail shall not be liable for any possible damage >> caused by modifications of the e-mail or by delay with transfer of >the >> email. >> >> In case that this e-mail forms part of business dealings: >> - the sender reserves the right to end negotiations about entering >into a >> contract in any time, for any reason, and without stating any >reasoning. >> - if the e-mail contains an offer, the recipient is entitled to >> immediately accept such offer; The sender of this e-mail (offer) >excludes >> any acceptance of the offer on the part of the recipient containing >any >> amendment or variation. >> - the sender insists on that the respective contract is concluded >only >> upon an express mutual agreement on all its aspects. >> - the sender of this e-mail informs that he/she is not authorized to >enter >> into any contracts on behalf of the company except for cases in which >> he/she is expressly authorized to do so in writing, and such >authorization >> or power of attorney is submitted to the recipient or the person >> represented by the recipient, or the existence of such authorization >is >> known to the recipient of the person represented by the recipient. >> > > [[alternative HTML version deleted]] > >______________________________________________ >[hidden email] mailing list -- To UNSUBSCRIBE and more, see >https://stat.ethz.ch/mailman/listinfo/r-help >PLEASE do read the posting guide >http://www.R-project.org/posting-guide.html >and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by PaulJohnson32gmail
On 04/21/2017 05:19 AM, Paul Johnson wrote:
> We all agree it is a problem with digital computing, not unique to R. I > don't think that is the right place to stop. > > What to do? The round example arose in a real funded project where 2 R > programs differed in results and cause was that one person got 57 and > another got 58. The explanation was found, but its less clear how to > prevent similar in future. Guidelines, anyone? Note that the error is amplified by the use of round() at the end of the calculation. Whoever wrote that code should know that by doing this the result will possibly differ by 1 across users. So, ideally, s/he should document this in his/her function to minimize the "surprise effect" on the user side. Another option is to not round() the final result and leave that responsibility to the end user. That's what I would do. H. > > So far, these are my guidelines. > > 1. Insert L on numbers to signal that you really mean INTEGER. In R, > forgetting the L in a single number will usually promote whole calculation > to floats. > 2. S3 variables are called 'numeric' if they are integer or double storage. > So avoid "is.numeric" and prefer "is.double". > 3. == is a total fail on floats > 4. Run print with digits=20 so we can see the less rounded number. Perhaps > start sessions with "options(digits=20)" > 5. all.equal does what it promises, but one must be cautious. > > Are there math habits we should follow? > > For example, Is it generally true in R that (100*x)/y is more accurate than > 100*(x/y), if x > y? (If that is generally true, couldn't the R > interpreter do it for the user?) > > I've seen this problem before. In later editions of the game theory program > Gambit, extraordinary effort was taken to keep values symbolically as > integers as long as possible. Avoid division until the last steps. Same in > Swarm simulations. Gary Polhill wrote an essay about the Ghost in the > Machine along those lines, showing accidents from trusting floats. > > I wonder now if all uses of > or < with numeric variables are suspect. > > Oh well. If everybody posts their advice, I will write a summary. > > Paul Johnson > University of Kansas > > On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > >> Hi >> >> The problem is that people using Excel or probably other such spreadsheets >> do not encounter this behaviour as Excel silently rounds all your >> calculations and makes approximate comparison without telling it does so. >> Therefore most people usually do not have any knowledge of floating point >> numbers representation. >> >> Cheers >> Petr >> >> -----Original Message----- >> From: R-help [mailto:[hidden email]] On Behalf Of Paul >> Johnson >> Sent: Thursday, April 20, 2017 11:56 PM >> To: R-help <[hidden email]> >> Subject: [R] Interesting quirk with fractions and rounding >> >> Hello, R friends >> >> My student unearthed this quirk that might interest you. >> >> I wondered if this might be a bug in the R interpreter. If not a bug, it >> certainly stands as a good example of the dangers of floating point numbers >> in computing. >> >> What do you think? >> >>> 100*(23/40) >> [1] 57.5 >>> (100*23)/40 >> [1] 57.5 >>> round(100*(23/40)) >> [1] 57 >>> round((100*23)/40) >> [1] 58 >> >> The result in the 2 rounds should be the same, I think. Clearly some >> digital number devil is at work. I *guess* that when you put in whole >> numbers and group them like this (100*23), the interpreter does integer >> math, but if you group (23/40), you force a fractional division and a >> floating point number. The results from the first 2 calculations are not >> actually 57.5, they just appear that way. >> >> Before you close the books, look at this: >> >>> aa <- 100*(23/40) >>> bb <- (100*23)/40 >>> all.equal(aa,bb) >> [1] TRUE >>> round(aa) >> [1] 57 >>> round(bb) >> [1] 58 >> >> I'm putting this one in my collection of "difficult to understand" >> numerical calculations. >> >> If you have seen this before, I'm sorry to waste your time. >> >> pj >> -- >> Paul E. Johnson https://urldefense.proofpoint.com/v2/url?u=http-3A__pj.freefaculty.org&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=iJyBC6GWjCxIv2vLzCBFL9KSP7jdm7aNxv00SLVJ7P0&e= >> Director, Center for Research Methods and Data Analysis >> https://urldefense.proofpoint.com/v2/url?u=http-3A__crmda.ku.edu&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=enQ7hFGHB0hmVSmc-Ey3B_421i62QzFsaiN8faqRiSc&e= >> >> To write to me directly, please address me at pauljohn at ku.edu. >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=lhpXJFzjB4p_pEFeUijjWa8Xd7IQ9ZOIzh5grUd7DQg&e= >> PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=ekbaqk4VTBEMyCFXosKWt561I1fiM0OVZDw5p8spUaI&e= >> posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. >> >> ________________________________ >> Tento e-mail a jakékoliv k němu připojené dokumenty jsou důvěrné a jsou >> určeny pouze jeho adresátům. >> Jestliže jste obdržel(a) tento e-mail omylem, informujte laskavě >> neprodleně jeho odesílatele. Obsah tohoto emailu i s přílohami a jeho kopie >> vymažte ze svého systému. >> Nejste-li zamýšleným adresátem tohoto emailu, nejste oprávněni tento email >> jakkoliv užívat, rozšiřovat, kopírovat či zveřejňovat. >> Odesílatel e-mailu neodpovídá za eventuální škodu způsobenou modifikacemi >> či zpožděním přenosu e-mailu. >> >> V případě, že je tento e-mail součástí obchodního jednání: >> - vyhrazuje si odesílatel právo ukončit kdykoliv jednání o uzavření >> smlouvy, a to z jakéhokoliv důvodu i bez uvedení důvodu. >> - a obsahuje-li nabídku, je adresát oprávněn nabídku bezodkladně přijmout; >> Odesílatel tohoto e-mailu (nabídky) vylučuje přijetí nabídky ze strany >> příjemce s dodatkem či odchylkou. >> - trvá odesílatel na tom, že příslušná smlouva je uzavřena teprve >> výslovným dosažením shody na všech jejích náležitostech. >> - odesílatel tohoto emailu informuje, že není oprávněn uzavírat za >> společnost žádné smlouvy s výjimkou případů, kdy k tomu byl písemně zmocněn >> nebo písemně pověřen a takové pověření nebo plná moc byly adresátovi tohoto >> emailu případně osobě, kterou adresát zastupuje, předloženy nebo jejich >> existence je adresátovi či osobě jím zastoupené známá. >> >> This e-mail and any documents attached to it may be confidential and are >> intended only for its intended recipients. >> If you received this e-mail by mistake, please immediately inform its >> sender. Delete the contents of this e-mail with all attachments and its >> copies from your system. >> If you are not the intended recipient of this e-mail, you are not >> authorized to use, disseminate, copy or disclose this e-mail in any manner. >> The sender of this e-mail shall not be liable for any possible damage >> caused by modifications of the e-mail or by delay with transfer of the >> email. >> >> In case that this e-mail forms part of business dealings: >> - the sender reserves the right to end negotiations about entering into a >> contract in any time, for any reason, and without stating any reasoning. >> - if the e-mail contains an offer, the recipient is entitled to >> immediately accept such offer; The sender of this e-mail (offer) excludes >> any acceptance of the offer on the part of the recipient containing any >> amendment or variation. >> - the sender insists on that the respective contract is concluded only >> upon an express mutual agreement on all its aspects. >> - the sender of this e-mail informs that he/she is not authorized to enter >> into any contracts on behalf of the company except for cases in which >> he/she is expressly authorized to do so in writing, and such authorization >> or power of attorney is submitted to the recipient or the person >> represented by the recipient, or the existence of such authorization is >> known to the recipient of the person represented by the recipient. >> > > [[alternative HTML version deleted]] > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=lhpXJFzjB4p_pEFeUijjWa8Xd7IQ9ZOIzh5grUd7DQg&e= > PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_posting-2Dguide.html&d=DwIGaQ&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=1JaAU8JTsAs-1eiwcWXgFMF6An1cjDHTSOjin674VRk&s=Gg8ShT8jT0TS19y42_4tYW23RMUgqeewiCRAW-TbXT0&e= > and provide commented, minimal, self-contained, reproducible code. > -- Hervé Pagès Program in Computational Biology Division of Public Health Sciences Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N, M1-B514 P.O. Box 19024 Seattle, WA 98109-1024 E-mail: [hidden email] Phone: (206) 667-5791 Fax: (206) 667-1319 ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
I've been following this thread with interest. A nice
collection of things to watch out for, if you don't want the small arithmetic errors due to finite-length digital representations of fractions to cause trouble! However, as well as these small discrepancies, major malfunctions can also result. Back on Dec 22, 2013, I posted a Christmas Greetings message to R-help: Season's Greetings (and great news ... )! which starts: Greetings All! With the Festive Season fast approaching, I bring you joy with the news (which you will surely wish to celebrate) that R cannot do arithmetic! Usually, this is manifest in a trivial way when users report puzzlement that, for instance, sqrt(pi)^2 == pi # [1] FALSE which is the result of a (generally trivial) rounding or truncation error: sqrt(pi)^2 - pi # [1] -4.440892e-16 But for some very simple calculations R goes off its head. And the example given is: Consider a sequence generated by the recurrence relation x[n+1] = 2*x[n] if 0 <= x[n] <= 1/2 x[n+1] = 2*(1 - x[n]) if 1/2 < x[n] <= 1 (for 0 <= x[n] <= 1). This has equilibrium points (x[n+1] = x[n]) at x[n] = 0 and at x[n] = 2/3: 2/3 -> 2*(1 - 2/3) = 2/3 It also has periodic points, e.g. 2/5 -> 4/5 -> 2/5 (period 2) 2/9 -> 4/9 -> 8/9 -> 2/9 (period 3) The recurrence relation can be implemented as the R function nextx <- function(x){ if( (0<=x)&(x<=1/2) ) {x <- 2*x} else {x <- 2*(1 - x)} } Now have a look at what happens when we start at the equilibrium point x = 2/3: N <- 1 ; x <- 2/3 while(x > 0){ cat(sprintf("%i: %.9f\n",N,x)) x <- nextx(x) ; N <- N+1 } cat(sprintf("%i: %.9f\n",N,x)) For a while [run it and see!], this looks as though it's doing what the arithmetic would lead us to expect: the first 24 results will all be printed as 0.666666667, which looks fine as 2/3 to 9 places. But then the "little errors" start to creep in: N=25: 0.666666666 N=28: 0.666666672 N=46: 0.667968750 N=47: 0.664062500 N=48: 0.671875000 N=49: 0.656250000 N=50: 0.687500000 N=51: 0.625000000 N=52: 0.750000000 N=53: 0.500000000 N=54: 1.000000000 N=55: 0.000000000 What is happening is that, each time R multiplies by 2, the binary representation is shifted up by one and a zero bit is introduced at the bottom end. At N=53, the first binary bit of 'x' is 1, and all the rest are 0, so now 'x' is exactly 0.5 = 1/2, hence the final two are also exact results; 53 is the Machine$double.digits = 53 binary places. So this normally "almost" trivial feature can, for such a simple calculation, lead to chaos or catastrophe (in the literal technical sense). For more detail, including an extension of the above, look at the original posting in the R-help archives for Dec 22, 2013: From: (Ted Harding) <[hidden email]> Subject: [R] Season's Greetings (and great news ... )! Date: Sun, 22 Dec 2013 09:59:16 -0000 (GMT) (Apologies, but I couldn't track down the URL for this posting in the R-help archives; there were a few follow-ups). I gave this as an example to show that the results of the "little" arithmetic errors (such as have recently been discussed from many aspects) can, in certain contexts, destroy a computation. So be careful to consider what can happen in the particular context you are working with. There are ways to dodge the issue -- such as using the R interface to the 'bc' calculator, which computes arithmetic expressions in a way which is quite different from the fixed-finite-length binary representation and algorithms used, not only by R, but also by many other numerical computation software suites Best wishes to all, Ted. ------------------------------------------------- E-Mail: (Ted Harding) <[hidden email]> Date: 21-Apr-2017 Time: 22:03:15 This message was sent by XFMail ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
On 04/21/2017 02:03 PM, (Ted Harding) wrote:
> I've been following this thread with interest. A nice > collection of things to watch out for, if you don't > want the small arithmetic errors due to finite-length > digital representations of fractions to cause trouble! > > However, as well as these small discrepancies, major > malfunctions can also result. > > Back on Dec 22, 2013, I posted a Christmas Greetings > message to R-help: > > Season's Greetings (and great news ... )! > > which starts: > > Greetings All! > With the Festive Season fast approaching, I bring you joy > with the news (which you will surely wish to celebrate) > that R cannot do arithmetic! > > Usually, this is manifest in a trivial way when users report > puzzlement that, for instance, > > sqrt(pi)^2 == pi > # [1] FALSE > > which is the result of a (generally trivial) rounding or > truncation error: > > sqrt(pi)^2 - pi > # [1] -4.440892e-16 > > But for some very simple calculations R goes off its head. > > And the example given is: > > Consider a sequence generated by the recurrence relation > > x[n+1] = 2*x[n] if 0 <= x[n] <= 1/2 > x[n+1] = 2*(1 - x[n]) if 1/2 < x[n] <= 1 > > (for 0 <= x[n] <= 1). > > This has equilibrium points (x[n+1] = x[n]) at x[n] = 0 > and at x[n] = 2/3: > > 2/3 -> 2*(1 - 2/3) = 2/3 > > It also has periodic points, e.g. > > 2/5 -> 4/5 -> 2/5 (period 2) > 2/9 -> 4/9 -> 8/9 -> 2/9 (period 3) > > The recurrence relation can be implemented as the R function > > nextx <- function(x){ > if( (0<=x)&(x<=1/2) ) {x <- 2*x} else {x <- 2*(1 - x)} > } > > Now have a look at what happens when we start at the equilibrium > point x = 2/3: > > N <- 1 ; x <- 2/3 > while(x > 0){ > cat(sprintf("%i: %.9f\n",N,x)) > x <- nextx(x) ; N <- N+1 > } > cat(sprintf("%i: %.9f\n",N,x)) > > For a while [run it and see!], this looks as though it's doing what > the arithmetic would lead us to expect: the first 24 results will all > be printed as 0.666666667, which looks fine as 2/3 to 9 places. > > But then the "little errors" start to creep in: > > N=25: 0.666666666 > N=28: 0.666666672 > N=46: 0.667968750 > N=47: 0.664062500 > N=48: 0.671875000 > N=49: 0.656250000 > N=50: 0.687500000 > N=51: 0.625000000 > N=52: 0.750000000 > N=53: 0.500000000 > N=54: 1.000000000 > N=55: 0.000000000 > > What is happening is that, each time R multiplies by 2, the binary > representation is shifted up by one and a zero bit is introduced > at the bottom end. > > At N=53, the first binary bit of 'x' is 1, and all the rest are 0, > so now 'x' is exactly 0.5 = 1/2, hence the final two are also exact > results; 53 is the Machine$double.digits = 53 binary places. > > So this normally "almost" trivial feature can, for such a simple > calculation, lead to chaos or catastrophe (in the literal technical > sense). The only surprise is to end up at the other equilibrium point (0), not that the iterations didn't stabilize around or converge to equilibrium point 2/3. That's because: 1) You didn't start *exactly* at equilibrium point 2/3 2) The iterating process is not supposed to converge but to diverge from the equilibrium point. It's its mathematical nature and has nothing to do with rounding errors. So if you don't start exactly at 2/3, the iterations will take you further apart, even on a hypothetical computer that uses exact representation of any arbitrary real number (i.e. no rounding errors). With such an instable algorithm, all bets are off: you can end up to the other equilibrium point or not end up anywhere (chaos) or hit a periodic point and go in cycles from there. But hey, shouldn't people examine the mathematical properties of their iterative numerical algorithms before implementing them? Like convergence, stability etc... H. > > For more detail, including an extension of the above, look at the > original posting in the R-help archives for Dec 22, 2013: > > From: (Ted Harding) <[hidden email]> > Subject: [R] Season's Greetings (and great news ... )! > Date: Sun, 22 Dec 2013 09:59:16 -0000 (GMT) > > (Apologies, but I couldn't track down the URL for this posting > in the R-help archives; there were a few follow-ups). > > I gave this as an example to show that the results of the "little" > arithmetic errors (such as have recently been discussed from many > aspects) can, in certain contexts, destroy a computation. > > So be careful to consider what can happen in the particular > context you are working with. > > There are ways to dodge the issue -- such as using the R interface > to the 'bc' calculator, which computes arithmetic expressions in > a way which is quite different from the fixed-finite-length binary > representation and algorithms used, not only by R, but also by many > other numerical computation software suites > > Best wishes to all, > Ted. > > ------------------------------------------------- > E-Mail: (Ted Harding) <[hidden email]> > Date: 21-Apr-2017 Time: 22:03:15 > This message was sent by XFMail > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://urldefense.proofpoint.com/v2/url?u=https-3A__stat.ethz.ch_mailman_listinfo_r-2Dhelp&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=ILkgA-W4NkL_PpalU1VtS9iO60moznxYVBGfU5QbyNw&s=puQK7QZvG7lZuOYvo9D6UANEgWylkApF-xIvpLMGVhQ&e= > PLEASE do read the posting guide https://urldefense.proofpoint.com/v2/url?u=http-3A__www.R-2Dproject.org_posting-2Dguide.html&d=DwICAg&c=eRAMFD45gAfqt84VtBcfhQ&r=BK7q3XeAvimeWdGbWY_wJYbW0WYiZvSXAJJKaaPhzWA&m=ILkgA-W4NkL_PpalU1VtS9iO60moznxYVBGfU5QbyNw&s=zohn8q-ufE7UhlQDQGc3KP-lsbM4O62SzwBXQ4S7I0g&e= > and provide commented, minimal, self-contained, reproducible code. > -- Hervé Pagès Program in Computational Biology Division of Public Health Sciences Fred Hutchinson Cancer Research Center 1100 Fairview Ave. N, M1-B514 P.O. Box 19024 Seattle, WA 98109-1024 E-mail: [hidden email] Phone: (206) 667-5791 Fax: (206) 667-1319 ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
In reply to this post by JRG
On Apr 21, 2017 12:01 PM, "JRG" <[hidden email]> wrote:
A good part of the problem in the specific case you initially presented is that some non-integer numbers have an exact representation in the binary floating point arithmetic being used. Basically, if the fractional part is of the form 1/2^k for some integer k > 0, there is an exact representation in the binary floating point scheme. > options(digits=20) > (100*23)/40 [1] 57.5 > 100*(23/40) [1] 57.499999999999992895 So the two operations give a slightly different result because the fractional part of the division of 100*23 by 40 is 0.5. So the first operations gives, exactly, 57.5 while the second operation does not because 23/40 has no exact representation. Thanks for answering. This case seemed fun because it was not a contrived example. We found this one by comparing masses of report tables from 2 separate programs. It happened 1 time in about 10,000 calculations. Guidelines for R coders, though, would be welcome. So far, all I am sure of is 1 Don't use == for floating point numbers. Your 1/2^k point helps me understand why == does seem to work correctly sometimes. I wonder if we should be suspicious of >=. Imagine the horror if a= w/x > b=y/z in fractions, but digitally a < b. Blech. Can that happen? But, change the example's divisor from 40 to 30 [the fractional part from 1/2 to 2/3]: > (100*23)/30 [1] 76.666666666666671404 > 100*(23/30) [1] 76.666666666666671404 Now the two operations give the same answer to the full precision available. So, it isn't "generally true true in R that (100*x)/y is more accurate than 100*(x/y), if x > y." The good news here is that round() gives same answer in both cases:) I am looking for a case where the first method is less accurate than the second. I expect that multiplying integers before dividing is never less accurate. Sometimes it is more accurate. ` Following your 1/2^k insight, you see why multiplying first is helpful in some cases. Question is will situation get worse. But Bert is right. I have to read more books. I studied Golub and van Loan and came away with healthy fear of matrix inversion. But when you look at user contributed regression packages, what do you find? Matrix inversion and lots of X'X. Paul Johnson University of Kansask The key (in your example) is a property of the way that floating point arithmetic is implemented. ---JRG On 04/21/2017 08:19 AM, Paul Johnson wrote: > We all agree it is a problem with digital computing, not unique to R. I > don't think that is the right place to stop. > > What to do? The round example arose in a real funded project where 2 R > programs differed in results and cause was that one person got 57 and > another got 58. The explanation was found, but its less clear how to > prevent similar in future. Guidelines, anyone? > > So far, these are my guidelines. > > 1. Insert L on numbers to signal that you really mean INTEGER. In R, > forgetting the L in a single number will usually promote whole calculation > to floats. > 2. S3 variables are called 'numeric' if they are integer or double > So avoid "is.numeric" and prefer "is.double". > 3. == is a total fail on floats > 4. Run print with digits=20 so we can see the less rounded number. Perhaps > start sessions with "options(digits=20)" > 5. all.equal does what it promises, but one must be cautious. > > Are there math habits we should follow? > > For example, Is it generally true in R that (100*x)/y is more accurate than > 100*(x/y), if x > y? (If that is generally true, couldn't the R > interpreter do it for the user?) > > I've seen this problem before. In later editions of the game theory program > Gambit, extraordinary effort was taken to keep values symbolically as > integers as long as possible. Avoid division until the last steps. Same in > Swarm simulations. Gary Polhill wrote an essay about the Ghost in the > Machine along those lines, showing accidents from trusting floats. > > I wonder now if all uses of > or < with numeric variables are suspect. > > Oh well. If everybody posts their advice, I will write a summary. > > Paul Johnson > University of Kansas > > On Apr 21, 2017 12:02 AM, "PIKAL Petr" <[hidden email]> wrote: > >> Hi >> >> The problem is that people using Excel or probably other such >> do not encounter this behaviour as Excel silently rounds all your >> calculations and makes approximate comparison without telling it does so. >> Therefore most people usually do not have any knowledge of floating point >> numbers representation. >> >> Cheers >> Petr >> ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
For over 4 decades I've had to put up with people changing my codes because
I use equalities of floating point numbers in tests for convergence. (Note that tests of convergence are a subset of tests for termination -- I'll be happy to explain that if requested.) Then I get "your program isn't working" and find that is is NOT my program, but a crippled version thereof. But I don't use equality tests (i.e., ==) blindly. My tests are of the form if ( (x_new + offset) == (x_old + offset) ) { # we cannot get more progress Now it is possible to imagine some weird cases where this can fail, so an additional test is needed on, say, maximum iterations to avoid trouble. But I've not seen such cases (or perhaps never noticed one, though I run with very large maximum counters). The test works by having offset as some modest value. For single precision I use 10 or 16, for double around 100. When x_new and x_old are near zero, the bit pattern is dominated by offset and we get convergence. When x_new is big, then it dominates. Why do I do this? The reason is that in the early 1970s there were many, many different floating point arithmetics with all sorts of choices of radix and length of mantissa. (Are "radix" and "mantissa" taught to computer science students any more?). If my programs were ported between machines there was a good chance any tolerances would be wrong. And users -- for some reason at the time engineers in particular -- would say "I only need 2 digits, I'll use 1e-3 as my tolerance". And "my" program would then not work. Sigh. For some reason, the nicely scaled offset did not attract the attention of the compulsive fiddlers. So equality in floating point is not always "wrong", though it should be used with some attention to what is going on. Apologies to those (e.g., Peter D.) who have heard this all before. I suspect there are many to whom it is new. John Nash On 2017-04-23 12:52 AM, Paul Johnson wrote: > > 1 Don't use == for floating point numbers. > ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
> On 23 Apr 2017, at 14:49 , J C Nash <[hidden email]> wrote: > > > So equality in floating point is not always "wrong", though it should be used > with some attention to what is going on. > > Apologies to those (e.g., Peter D.) who have heard this all before. I suspect > there are many to whom it is new. Peter D. still insists on never trusting exact equality, though. There was at least one case in the R sources where age-old code got itself into a condition where a residual terme that provably should decrease on every iteration oscillated between two values of 1-2 ulp in magnitude without ever reaching 0. The main thing is that you cannot trust optimising compilers these days. There is, e.g., no guarantee that a compiler will not transform (x_new + offset) == (x_old + offset) to (x_new + offset) - (x_old + offset) == 0 to (x_new - x_old) + (offset - offset) == 0 to.... well, you get the point. -pd -- Peter Dalgaard, Professor, Center for Statistics, Copenhagen Business School Solbjerg Plads 3, 2000 Frederiksberg, Denmark Phone: (+45)38153501 Office: A 4.23 Email: [hidden email] Priv: [hidden email] ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
Yes. I should have mentioned "optimizing" compilers, and I can agree with "never
trusting exact equality", though I consider conscious use of equality tests useful. Optimizing compilers have bitten me once or twice. Unfortunately, a lot of floating-point work requires attention to detail. In the situation of testing for convergence, the alternatives to the test I propose require quite a lot of code, with variants for different levels of precision e.g., single, double, quad. There's no universal answer and we do have to look "under the hood". A particularly nasty instance (now fortunately long gone) was the Tektronix 4051 graphics station, where the comparisons automatically included a "fuzz". There was a FUZZ command to set this. Sometimes the "good old days" weren't! Today's equivalent is when there is an upstream change to an "optimization" that changes the manner of computation, as in Peter's examples. If we specify x <- y * (a / b), then we should not get x <- (y * a) / b. A slightly related case concerns eigenvectors / singular vectors when there are degenerate values (i.e., two or more equal eigenvalues). The vectors are then determined only to lie in a (hyper)plane. A large computer contracting firm spent two weeks of high-priced but non-numerical help trying to find the "error" in either an IBM or Univac program because they gave very different eigenvectors. And in my own field of function minimization / nonlinear least squares, it is quite common to have multiple minima or a plateau. Does anyone know if R has a test suite to check some of these situations? If not, I'll be happy to participate in generating some. They would not need to be run very often, and could be useful as a didactic tool as well as checking for differences in platforms. JN On 2017-04-23 09:37 AM, peter dalgaard wrote: > >> On 23 Apr 2017, at 14:49 , J C Nash <[hidden email]> wrote: >> >> >> So equality in floating point is not always "wrong", though it should be used >> with some attention to what is going on. >> >> Apologies to those (e.g., Peter D.) who have heard this all before. I suspect >> there are many to whom it is new. > > Peter D. still insists on never trusting exact equality, though. There was at least one case in the R sources where age-old code got itself into a condition where a residual terme that provably should decrease on every iteration oscillated between two values of 1-2 ulp in magnitude without ever reaching 0. The main thing is that you cannot trust optimising compilers these days. There is, e.g., no guarantee that a compiler will not transform > > (x_new + offset) == (x_old + offset) > > to > > (x_new + offset) - (x_old + offset) == 0 > > to > > (x_new - x_old) + (offset - offset) == 0 > > to.... well, you get the point. > > -pd > ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
John,
I would be happy to participate in designing the test suite you suggest. About a year ago I revised FAQ 7.31, based on my talk at the Aalberg R conference. It now points, in addition to the Goldberg paper that has been referenced there for a long time, to my appendix on precision. Here is the link from the FAQ A discussion with many easily followed examples is in Appendix G "Computational Precision and Floating Point Arithmetic", pages 753-771 of _Statistical Analysis and Data Display: An Intermediate Course with Examples in R_, Richard M. Heiberger and Burt Holland (Springer 2015, second edition). This appendix is a free download from <http://link.springer.com/content/pdf/bbm%3A978-1-4939-2122-5%2F1.pdf>. The appendix is based on the paper I gave at the Aalberg R Conference. It uses the Rmpfr package to illustrate exactly what is happening at the level of the machine arithmetic. The investigation and discussion of the effect of optimization on floating point arithmetic should use the Rmpfr tools. Rich On Sun, Apr 23, 2017 at 10:06 AM, J C Nash <[hidden email]> wrote: > Yes. I should have mentioned "optimizing" compilers, and I can agree with > "never > trusting exact equality", though I consider conscious use of equality tests > useful. > Optimizing compilers have bitten me once or twice. Unfortunately, a lot of > floating-point work requires attention to detail. In the situation of > testing > for convergence, the alternatives to the test I propose require quite a lot > of > code, with variants for different levels of precision e.g., single, double, > quad. > > There's no universal answer and we do have to look "under the hood". A > particularly > nasty instance (now fortunately long gone) was the Tektronix 4051 graphics > station, > where the comparisons automatically included a "fuzz". There was a FUZZ > command to > set this. Sometimes the "good old days" weren't! Today's equivalent is when > there > is an upstream change to an "optimization" that changes the manner of > computation, > as in Peter's examples. If we specify x <- y * (a / b), then we should not > get > x <- (y * a) / b. > > A slightly related case concerns eigenvectors / singular vectors when there > are > degenerate values (i.e., two or more equal eigenvalues). The vectors are > then > determined only to lie in a (hyper)plane. A large computer contracting firm > spent > two weeks of high-priced but non-numerical help trying to find the "error" > in either > an IBM or Univac program because they gave very different eigenvectors. > > And in my own field of function minimization / nonlinear least squares, it > is quite > common to have multiple minima or a plateau. > > Does anyone know if R has a test suite to check some of these situations? If > not, > I'll be happy to participate in generating some. They would not need to be > run > very often, and could be useful as a didactic tool as well as checking for > differences in platforms. > > JN > > On 2017-04-23 09:37 AM, peter dalgaard wrote: >> >> >>> On 23 Apr 2017, at 14:49 , J C Nash <[hidden email]> wrote: >>> >>> >>> So equality in floating point is not always "wrong", though it should be >>> used >>> with some attention to what is going on. >>> >>> Apologies to those (e.g., Peter D.) who have heard this all before. I >>> suspect >>> there are many to whom it is new. >> >> >> Peter D. still insists on never trusting exact equality, though. There was >> at least one case in the R sources where age-old code got itself into a >> condition where a residual terme that provably should decrease on every >> iteration oscillated between two values of 1-2 ulp in magnitude without ever >> reaching 0. The main thing is that you cannot trust optimising compilers >> these days. There is, e.g., no guarantee that a compiler will not transform >> >> (x_new + offset) == (x_old + offset) >> >> to >> >> (x_new + offset) - (x_old + offset) == 0 >> >> to >> >> (x_new - x_old) + (offset - offset) == 0 >> >> to.... well, you get the point. >> >> -pd >> > > ______________________________________________ > [hidden email] mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
Thanks Richard.
I've some stuff too, but I need to look it up. A few years ago I built a small test spreadsheet for Gnumeric when working with Jody Goldberg. In the early 2000s, Jody contacted R (I think Duncan Murdoch) to ask if it was OK for Gnumeric to use R's distribution function approximations (Yes). Someone can correct me if wrong, but apparently a few weeks later Jody sent some improvements, and I believe these were incorporated. Good open-source improvements like that deserve to be recorded in a better way than I have done here. But in any event, I can dig out that material, as well as some of Kahan's PARANOIA materials. Suggest taking the discussion off-line from here on unless we come up with some important bugs or improvements or ... Best, JN On 2017-04-23 11:42 AM, Richard M. Heiberger wrote: > John, > > I would be happy to participate in designing the test suite you suggest. > > About a year ago I revised FAQ 7.31, based on my talk at the Aalberg R > conference. It now points, in addition to the Goldberg paper that has > been referenced there for a long time, to my appendix on precision. > Here is the link from the FAQ > > A discussion with many easily followed examples is in Appendix G > "Computational Precision and Floating Point Arithmetic", pages 753-771 > of _Statistical Analysis and Data Display: An Intermediate Course with > Examples in R_, Richard M. Heiberger and Burt Holland (Springer 2015, > second edition). This appendix is a free download from > <http://link.springer.com/content/pdf/bbm%3A978-1-4939-2122-5%2F1.pdf>. > > The appendix is based on the paper I gave at the Aalberg R > Conference. It uses the Rmpfr package to illustrate exactly what > is happening at the level of the machine arithmetic. The > investigation and discussion of the effect of optimization on > floating point arithmetic should use the Rmpfr tools. > > > Rich > > On Sun, Apr 23, 2017 at 10:06 AM, J C Nash <[hidden email]> wrote: >> Yes. I should have mentioned "optimizing" compilers, and I can agree with >> "never >> trusting exact equality", though I consider conscious use of equality tests >> useful. >> Optimizing compilers have bitten me once or twice. Unfortunately, a lot of >> floating-point work requires attention to detail. In the situation of >> testing >> for convergence, the alternatives to the test I propose require quite a lot >> of >> code, with variants for different levels of precision e.g., single, double, >> quad. >> >> There's no universal answer and we do have to look "under the hood". A >> particularly >> nasty instance (now fortunately long gone) was the Tektronix 4051 graphics >> station, >> where the comparisons automatically included a "fuzz". There was a FUZZ >> command to >> set this. Sometimes the "good old days" weren't! Today's equivalent is when >> there >> is an upstream change to an "optimization" that changes the manner of >> computation, >> as in Peter's examples. If we specify x <- y * (a / b), then we should not >> get >> x <- (y * a) / b. >> >> A slightly related case concerns eigenvectors / singular vectors when there >> are >> degenerate values (i.e., two or more equal eigenvalues). The vectors are >> then >> determined only to lie in a (hyper)plane. A large computer contracting firm >> spent >> two weeks of high-priced but non-numerical help trying to find the "error" >> in either >> an IBM or Univac program because they gave very different eigenvectors. >> >> And in my own field of function minimization / nonlinear least squares, it >> is quite >> common to have multiple minima or a plateau. >> >> Does anyone know if R has a test suite to check some of these situations? If >> not, >> I'll be happy to participate in generating some. They would not need to be >> run >> very often, and could be useful as a didactic tool as well as checking for >> differences in platforms. >> >> JN >> >> On 2017-04-23 09:37 AM, peter dalgaard wrote: >>> >>> >>>> On 23 Apr 2017, at 14:49 , J C Nash <[hidden email]> wrote: >>>> >>>> >>>> So equality in floating point is not always "wrong", though it should be >>>> used >>>> with some attention to what is going on. >>>> >>>> Apologies to those (e.g., Peter D.) who have heard this all before. I >>>> suspect >>>> there are many to whom it is new. >>> >>> >>> Peter D. still insists on never trusting exact equality, though. There was >>> at least one case in the R sources where age-old code got itself into a >>> condition where a residual terme that provably should decrease on every >>> iteration oscillated between two values of 1-2 ulp in magnitude without ever >>> reaching 0. The main thing is that you cannot trust optimising compilers >>> these days. There is, e.g., no guarantee that a compiler will not transform >>> >>> (x_new + offset) == (x_old + offset) >>> >>> to >>> >>> (x_new + offset) - (x_old + offset) == 0 >>> >>> to >>> >>> (x_new - x_old) + (offset - offset) == 0 >>> >>> to.... well, you get the point. >>> >>> -pd >>> >> >> ______________________________________________ >> [hidden email] mailing list -- To UNSUBSCRIBE and more, see >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html >> and provide commented, minimal, self-contained, reproducible code. ______________________________________________ [hidden email] mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. |
Free forum by Nabble | Edit this page |