How strong are retro chess games' engines?
 Nothing Ever Happens Shirt $21.68 |
 Tip Your Landlord Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
How strong are retro chess games' engines?
How strong are retro chess games' engines?
Nothing Ever Happens Shirt $21.68 |
Tip Your Landlord Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
Consoles had to think for really long time when playing at higher levels. Atari 2600 chess has a meme highest difficulty where the opponent thinks for multiple hours to play its first turn because it has to bruteforce thorough all possible plays before taking a decision. That's why the most powerful co-processor on SNES, a 21 MHz 32-bit ARM chip was used only once in a shogi game of all things.
even in draughts it takes quite a long time
Not to mention that was the sequel, the previous version used the second-most powerful SNES co-processor, also an ARM CPU.
I wonder how much putting a CPU that was several times more powerful than the entire SNES itself on a cart inflated it's costs back then, since those games are about $5 at most these days.
My dad had the Atari computer version of Chess. It was pretty shit. I used to have some Windows 3.x chess games that were also a joke.
The one with the windows entertainment pack is fine for the lay man like me.
I don't think you could beat Deep Blue.
I was trained by the best
Deep Blue was a supercomputer in 1996. I doubt an NES cart from 1991 had access to that technology.
still retro
It was available via polyphase quantum diamondylium channeled by element 116 booster
Taboo: The Sixth Sense had access to something far more powerful: the spirit realm
Maybe some chess games were implemented similarly
I doubt anyone here would even be given the chance to play against it you moron.
A $100 smart phone with Stockfish could beat Deep Blue. Chess engines that can beat the best players in the world aren't exactly rare these days.
What's your point? Nothing you said contradicts the statement that the vast majority of people couldn't beat DeepBlue if they tried.
My point is that if somebody wants to have the experience of playing against a computer that they can't beat then they should just pull out their phone.
Ok, that's not really what the conversation was about though.
What was the first commercially available chess program that was impossible for a human to beat?
I dunno, I'd just be using google-fu at this point. Some time between 1996 and 2010.
>Some time between 1996 and 2010.
Sounds like a zoomer’s conception of time
I dont understand how they make a bad chess AI. What are they doing if, lets say, the AI finds a great move right away? Just let it pick a random mistake from a list of frickups?
Back in the day they basically had the AI consider potential moves for x amount of time and then pick whatever the most optimal one it could find is. You don't win at chess by making a single great move followed by a bunch of halfassed ones.
They probably create an ordering of potential moves weighted on score, and for the medium difficulty they pick something that's in the top 80%ish (with random variance) rather than the absolute best scores.
If you play against Stockfish on chess.com, you can check a box to enable showing what the computer is thinking.
It evaluates moves by looking at each move for each piece it can move on the board. The computer then plays ahead a few turns for each of these moves, assuming you'll play perfect counters to each of it's moves.
After doing all that, it ranks the moves base on how strong it's position was after those simulated turns.
For Stockfish on chess.com, it'll usually show you it's top 3 moves, and it's predictions for the next 10 turns based on those moves.
Setting the difficulty level basically adjusts a probabilistic model of which of those top 3 moves the computer will choose.
If you set it on easy, the computer will mostly choose the worst of the top 3 moves, and if you set it on hard, it will mostly choose the best possible move.
Stockfish also scores how aggressive or defensive moves are, and you can also tweak that so that it will give preference to one or the other.
It can also do fun things like analyze your moves to tell if you've mostly been playing defensively or offensively and adjust it's move preferences to better counter you.
That one N64 chess game, if you try and play F4 as white it will always counter with E5, which is crazy like fox news. Human players don't want to give up their king pawn and play D5. It's some moronic gambit. If you transpose to king's gambit the computer will give away its queen in like two moves, guaranteed, it always did it.
Granted this is the computer's "bad AI". If I actually adjusted the difficulty I always lost. And I consider myself a decent player.