Hacker News Reader

Comment by sillysaurusx | Hacker News Reader

How could the top players be better than any chess engine? It seems like even in 2000 any off the shelf chess engine could beat any human. But I guess not...

It would be neat to see the elo of the engines over time.

EDIT: Heh, deep blue’s famous match happened in 1997. I guess this really is possible. It’s weird growing up with the idea that “every chess engine can beat humans every time”; I wonder when that became true.

View on HN

There's a "no news" zone between a discovery and proliferation. In 1997, deep blue become superhuman.. proving the point. The average consumer engine was still like average tournament player with really weird moves. It took another decade before superhuman engines became widely available to normal people.

That decade gap is where we often mispredict the future. Once a point is proven, we expect instant impact. Reality takes time.

It took 15-20 years after superhuman chess engines were invented before they really impacted chess. Cheating is one impact, but novelty is another one. "Computer moves" are learned by GMs and played in world class tournaments. Contributing novelties is a mark of influence in chess. Even computer strategies are starting to appear in human tournaments.

There was also the little fact that Kasparov was not playing his top chess in the 1997 match - the loss in the sixth game was considered very bizarre at the time (him going for the Caro-Kann, while not having much experience in that opening)

His rematch against Deep Junior (the 2003 computer) ended in a 3-3 tie, so humans were actually able to hold off computer for quite a bit longer than 1997...

Friendly reminder your GTX 1080 is more powerful than "Deeper blue" (not deep blue beat Kasparov, but deeper blue, which was a heavy upgrade) so unless you're better then Kasparov any kid could beat you through cheating.

A more realistic setup is Stockfish on a 4+ core CPU desktop, or maybe on a modern smartphone. GPU chess engines aren't as useful to the majority of machines without a high-end card.

The difference between a super GM and stockfish is that a super GM will play moves the "feel natural" to a human given the current position on the board. A chess engine won't care about what is natural and will maximize the position 15-20 positions down the line. For this reason, they will make sometimes a weird move given the current position that will result in an advantage a lot of moves downstream.

One of the main reasons that Petrosian "got caught" was other GMs (see Hikaru's YouTube video on this) pointing out how unnatural some of his moves felt to them and considering the speed at which he supposedly calculated them.

That is true of traditional chess engines, but no longer true for neural net (self-taught) engines that learn the rules of chess by simulating millions of games against itself.

I've played an early version of one of those (its playing strength is roughly proportional to the amount of training it has done) and it was quite human-like in its behavior, even dropping pieces and making very human-like mistakes.

Look up AlphaZero and the open source port LeelaChess :) Once those become commercially available, it's going to be very hard to detect engine usage.

Super GMs often catch cheaters by creating a stale board in the late mid-game and watch opponents make pointless moves many times in a row. It becomes pretty obvious at that moment.

As someone who only knows the very basics of chess, what do you mean by "stale board"? Searching for that term didn't turn up any interesting results.

He means that a real GM would know that the position is a draw and would offer a draw and stop playing. The cheater doesn't know so he has the engine keep playing forever.

By a stale board I mean a board state where there are no sensible moves in order to create an advantage. i.e. there are no good pawn moves, no good knight or bishop re-positions, castles mostly stuck and the engine pretty much shows no point difference.

I’m not sure, maybe OP means a board where you would customarily ask for a draw as there is no way for you to win and the opponent could only get a stalemate at best.

I think they mean a closed position without many meaningful moves available. Often times real players make small incremental positional improvements and probe weaknesses, and engines take completely different lines.

> EDIT: Heh, deep blue’s famous match happened in 1997. I guess this really is possible. It’s weird growing up with the idea that “every chess engine can beat humans every time”; I wonder when that became true.

There is an interesting section in Hofstaeder's Gödel, Escher, Bach from 1979 where he speculates that, despite the rapid progress of technology, computers might never be able to play chess at a human level!

Reply on news.ycombinator.com