Back to the design catalogFactorial gives the textbooks' answers
We ran the worked examples of the classic books and papers through Factorial, and checked every answer they print against ours.
The effect of temperature on yield
The book 23Factorial 23.0
See the whole exampleWhen an answer differs, we say why
And where we do not know why yet, we say so.
- A misprint in the bookWe show the page, what the book prints, and the value its own data give.
- The book rounds soonerWorked by hand, a book rounds part-way through. Factorial carries every digit.
- Another methodThe book uses another method or convention, which we name.
- Not in Factorial yetA few methods, such as ridge analysis, are still to come. Each book's page lists them.
Pick your book
- Matches
- Very close
- Close
- Differs
Loading the cases...
Show the numbers
How closely each case reproduces
| Tier | What it means | Cases | Numbers | The list |
|---|
A case is placed by its worst number, so one large difference puts a case in the last tier even when every other number it prints agrees; the Numbers column counts each number in its own tier. A number the source prints in error, or a quiz option its answer key marks wrong, does not count against a case: the library's value is the right one.
Why numbers differ, and which are not recomputed
For all the sources together. The catalogue records the cause of every number that differs from its source and of every printed number the library does not recompute.
By source
Each row counts one source's cases in each tier, and the share of its printed numbers that agree with the library. A difference is between the library and what a source prints, whatever its cause: the source's own rounding, a misprint, or a method that differs from the library's. The tiers describe how far the library reproduces each source, not the quality of the source. Choose a source, or a kind of source, to see only its cases on this page.
How far off the differences are
Every case
Why do numbers differ? Try:
Not in the library yet
The printed numbers the library has no function for, grouped by what is missing, those that would recompute the most numbers first. The list follows the search and the source chosen above.