3 ms·
> write a program that plays every possible tic-tac-toe game Last time I interviewed in the Valley, some mofo actually asked me how many TicTacToe games end in
by dxbydt 6y ago
> write a program that plays every possible tic-tac-toe game
Last time I interviewed in the Valley, some mofo actually asked me how many TicTacToe games end in 5 moves ? 9.5%. Here's my MonteCarlo solution if you can't be bothered with this garbage. Sorry that repo is 7 years old. Never got a chance to use it again.
https://github.com/krishnanraman/TicTacToe https://github.com/krishnanraman/TicTacToe
Makes me so glad I quit the field & went into stats. My classmate just graduated with his stats phd & I asked him how the job interview went. Apparently they asked him if he was a frequentist or a Bayesian. So he said he was a Fisherian, and they had a good laugh. After a few preliminaries on p-values, credible intervals vs confidence intervals, 2 way anova, standard proofs on Cochran's theorem, Cramer-Rao & how to simulate multi-variate gaussian, everybody went for lunch. He was out the door under an hour with a job offer.
- joshuamorton 6y ago> 2 way anova, standard proofs on Cochran's theorem, Cramer-Rao & how to simulate multi-variate gaussian I don't see how regurgitating standard proofs is far and away different from regurgitating (or solving) algorithmic coding problems. "Implement a priority queue" is to software engineering what "prove Cochrans' theorem" is to stats (arguably the second is more difficult/less reasonable).