Thanks for posting these Haiku - very interesting. I looked Houdini up and it had about a 3300 rating already. Wonder if computers will someday have a 4000 rating or what's the upper limit. Course if it keeps beating 2400 rated players, does its rating go up about 8 pts every time? So that theoretically it could indeed go up to 4000?
Houdini playing the King's Gambit
Sort:
just for kicks it would be funny to see what would happen if houdini played something like e4 e5 qh5 nc6 qxf7. I know there is someone on these forums where that is their favorite opening.
The differences in the ratings between the two players determines how many points you gain. At some point between 3000 and 4000, the upper-rated player would no longer gain anything from defeating the lower rated player. We see that in live chess: if you're a 1400 player, you'll gain at least 10-15 points from defeating someone in the 2000s, but if you defeat someone with a rating at, say, 800, you'd be lucky to gain anything!
As for reaching a 4000 rating, that would depend on how well we are playing already! At some point, you would have to achieve perfection in order to become any better, and at that point, your highest rating would be determined by the ratings of your competitors. We'll see if we manage to reach that point or not... I'm not sure if I'm willing to bet on it just yet. :)
I'll post both the Bongcloud and The Idiot's Opening (a variation to the Scholar's Mate, in which White forgets to move the Bishop!) soon. Life's interrupting.
Haiku575 wrote:
In correspondence chess, he actually might be able to. It all depends on how far and accurately he analyzes. He could theoretically draw or even win, but it would take a LOT of work. To even the playing field, I could agree to limit Houdini's playing strength, which I can do so by either limiting the search depth (how far it looks ahead) or the time it takes to consider each position (self-explanatory). For example, each game I posted above had both players thinking for six seconds each move. I chose this limit to save time. If you want to play against Houdini with it slightly limited in its abilities, I'll be happy to do so. However, since I do want to keep things interesting, I won't set limits that will make it play like a fool.
Sounds tempting, but I'd probably lose even if you only gave Houdini 1s/move.
Seems like there's one opening that allows Dragon to win, after all... though it took forever. I kept expecting Black to accidentally stalemate it.
browni3141 wrote:
Haiku575 wrote:
In correspondence chess, he actually might be able to. It all depends on how far and accurately he analyzes. He could theoretically draw or even win, but it would take a LOT of work. To even the playing field, I could agree to limit Houdini's playing strength, which I can do so by either limiting the search depth (how far it looks ahead) or the time it takes to consider each position (self-explanatory). For example, each game I posted above had both players thinking for six seconds each move. I chose this limit to save time. If you want to play against Houdini with it slightly limited in its abilities, I'll be happy to do so. However, since I do want to keep things interesting, I won't set limits that will make it play like a fool.
Sounds tempting, but I'd probably lose even if you only gave Houdini 1s/move.
Yes, but at least you could say you tried!
Class F, you mean. :D
This opening's another old favorite.
Could we do it as an unrated chess.com game? Chess.com allows for the use of engines if both sides agree to it and it is unrated.
Could I play full strength Houdini giving me piece odds? You can analyze as long as you want. I want to play under conditions that I feel actually give me a chance.
try the jerome gambit for houdini
All right then. Let's do what Morphy used to do, and get rid of the Rook on a1 :P
Really? I meant a minor piece, but if that's the one you want to give up, then I'm fine with that. After I win this game can we play one that is fair to Houdini?
There's no convenient way to start without a piece, so how about we just use conditionals to get to the position?
Sure
Houdini plays an anti-computer opening:
The name of the above opening is (unofficially) the Nemeth Gambit, named after the person who had an unbeaten score with it against all chess programs. I actually like him using it better than Houdini:
As requested, the Jerome gambit:
Hmm... what would happen if I made it play the Najdorf, I wonder? Let's find out:
Parham opening? Sure: