20 Number Game

[jon-nyc]

@Horace said in 20 Number Game:

I guess those odds assume perfect play, which is probably not easy.

I would think those are the odds assuming you simply map the numbers to increments of 50.

[Horace]

@jon-nyc said in 20 Number Game:

@Horace said in 20 Number Game:

I guess those odds assume perfect play, which is probably not easy.

I would think those are the odds assuming you simply map the numbers to increments of 50.

Not a bad start for an approximation of perfect play, but there would be more to it than that.

[jon-nyc]

Yes but my point is those odds probably assume the simple mapping, not some dynamic mapping based on the numbers that have come.

[jon-nyc]

If Klaus were here he could write 3 lines of code in some obscure high level language that would confirm or reject my hypothesis using Monte Carlo methods.

[Horace]

In the reddit thread, someone did code a perfect play algo. Maybe it just assigns numbers to their closest unoccupied bin. But I doubt it. I haven't looked.

[89th]

Here's the python code if you are nerdy enough to understand it:

https://gitlab.com/whweaver/20-number-challenge

[89th]

Got to 16! 4 numbers left. Don't think I'll beat that anytime soon.

[Copper]

10, 8, 15, 12

[Horace]

I just got 21.

[Copper]

Pictures or it didn't happen.