News for hackers. Stuff that matters.
Re:To approximate 52! by hand, compute 54! and divide by 3000 by vinchuco on Sunday June 05, 2016 @02:14PM (#11842247)
I was tempted to just plug in 52 into Stirling's formula n!~~((n/e)^n)sqrt(2pi*n) but the result isn't as 'nice'.

What makes 54 special and how would you know to try it instead?

What makes a number computationally 'nice' for humans? Is it just less computation steps under relaxed precision?

Is there a collection of heuristics for approximation?(This might be just Numerical Analysis) Or better yet, ome sort of approximate calculator software that gives error bounds?

I've heard this book is good (from HN) but haven't gotten to reading it: Street-fighting mathematics. https://mitpress.mit.edu/sites/default/files/titles/free_dow...

Re:To approximate 52! by hand, compute 54! and divide by 3000 by ColinWright on Sunday June 05, 2016 @02:20PM (#11842283)
It's nice because 2.7 is close to e, and 2.7 is a nice factor of 54, which is close to 52.

It's worth starting from the fact that e^3 ~ 20 and see if you can make that work. I'd suggest calculating 50! using e^3 ~ 20 (somehow!) - that's the next version I'll be doing. It's an interesting exercise.