Easier to win against computer than human beings?

Sort:
Avatar of NikkiLikeChikki
It’s so easy to fix, too. You just look at Anand’s games, see what he plays against certain openings, and play those at the same rate. Andrea Botez should never play anything but the London… ever! Why doesn’t Morphy ever play the king’s gambit? Criminal!

It’s not complicated.
Avatar of p8q
NikkiLikeChikki wrote:
It’s so easy to fix, too. You just look at Anand’s games, see what he plays against certain openings, and play those at the same rate. Andrea Botez should never play anything but the London… ever! Why doesn’t Morphy ever play the king’s gambit? Criminal!

It’s not complicated.

hehehe, that's funny, the king's gambit XD, Morphy-bot doesn't play it and i find nowadays human players on live that play it all the time. The world is going crazy! shock.png

Surprisingly enough, i just read in Karpov's book "Find the right plan" (page 13th) that Morphy's style (also Philidor's) was very positional and that's why they always won (they were also strong in tactics, off course). They just didn't say it or didn't realize. His contemporaries didn't realize either.

That's why engines should be good at mimicking Morphy's style.

I wondered if there exist engines able to play romantic chess style, but the only thing i could find is to set a personality myself in Chessmaster11 by setting the parameter "Material/positional" all the way to material, also adding some randomness and full attacker. Openning book "Gambit":

i set maximum strength, max search depth.... i don't know if i did it right.

I'm winning all the time, so i don't know if it's maybe because i did it right and positional play beats easily material or i win because i did something really wrong, or maybe because engines can't play strong romantic style unless they are programmed as such from the core.

If anyone knows about an engine that plays romantic style, please tell us. OpenTal 1.1 engine is kinda cool, but not romantic enough (still doesn't bring me flowers or invite to dinner hahah).

 

Avatar of TheJugglingBear
p8q wrote:
TheJugglingBear wrote:

Mines the other way around. I can beat 1000-1300 or 1400ish players but man if I cant beat the ai David to save my life. Blunders then they plays 5 turn deep chess all of sudden playing master moves. 

That's interesting, i felt chess.com bots under 2000 much easier than humans here under 1200. Maybe it's because i played bots at 20 or 30 minutes games and humans always i play at 5|5 minutes.

So, maybe it's because of time control. At what time control do you beat humans 1300-1400 and what time control do you lose vs David (1400) bot?

 

Admittedly I play too quickly against the ai. No more than a 15 min game and that’s being generous. People can range from 10 min to 3days for a move depending on how much time I have also I have beaten him since this post. So happy lol. Now on to the next ones 

Avatar of p8q
TheJugglingBear wrote:
 

Admittedly I play too quickly against the ai. No more than a 15 min game and that’s being generous. People can range from 10 min to 3days for a move depending on how much time I have also I have beaten him since this post. So happy lol. Now on to the next ones 

I don't see in your archive any games vs bots, maybe you didn't know that you can save in your archive the games you play with bots? there's an option to save it.

Avatar of TheJugglingBear
p8q wrote:
TheJugglingBear wrote:
 

Admittedly I play too quickly against the ai. No more than a 15 min game and that’s being generous. People can range from 10 min to 3days for a move depending on how much time I have also I have beaten him since this post. So happy lol. Now on to the next ones 

I don't see in your archive any games vs bots, maybe you didn't know that you can save in your archive the games you play with bots? there's an option to save it.

I did not know that and it really bugged me that I couldnt go back and analyze them ;-; ill have to look for the option I play about 10 games a day against various ones trying different strats and openings.  Currently working on beating Ali.

Avatar of Robotofdeath395

Depends who your playing l say computers

Avatar of MasterMatthew52

Engine opponents aren't the rating they claim to be. A 1200 engine might drop a bishop in a horrible way, but a typical 1200 player wouldn't make that bad of a move. Beating a computer at "1200" is nothing like beating a player at 1200

Avatar of p8q
MasterMatthew52 wrote:

Engine opponents aren't the rating they claim to be. A 1200 engine might drop a bishop in a horrible way, but a typical 1200 player wouldn't make that bad of a move. Beating a computer at "1200" is nothing like beating a player at 1200

Depends which engine are you playing:

Chessmaster gives you a piece all the way to 1800 rating. Personalities under 1600 rating most of the time gives you the knight by Nxf2 in the opening and you take his knight with your king. Tasha personality does it 95% of the times. Other personalities also do Nxf2 too, always the same pointless sacrifice. But after that sacrifice get ready to play vs Fide Master strength until the engine does the next blunder. I've seen chess.com 1400 players losing vs Chessmaster 1200 rating perssonalities over and over again, despite of the Knight sacrifice.

Chess.com low rated engines gift you pieces too, IIRC, it's a long time i don't play low rated c.c bots, now i only play 2000+ bots. I think chess.com bots uses Komodo at different strength levels as chess engine.

However, Shredder or Rodent IV doesn't give you any piece away, their mistakes are more strategic or tactical. At least when they are at strength 1300+.

I think that difference is because some engines strength is reduced by nodes per second limitation, so with a few nodes per second the engine doesn't give pieces away, but you will see tactical strategical mistakes/blunders.

And the strength of the other type of engines that gift pieces, the strength is reduced by other ways... i suppose they are playing full strength in terms of nodes per second count, but they are gifting evaluation points to the opponent and that's why they choose to give material away, but i don't know if it's done in that way exactly, so don't quote me on that. Fact is that above 1800 rating Chessmaster doesn't give you any piece anymore, or it seldom does it, because the evaluation points it's programmed to give away to simulate low strength is inferior than material value (i suppose).

On the other hand, 1200 players sometimes give pieces away (I'm talking about blitz), not always though. I know it because that's how i lose most of the time XD hehehe

 

Avatar of p8q

By the way, i'm afraid engines rating are actually the rating they claim to be. I know it because in a forum, one of Chessmaster 11 directors said the way they rated the different personalities is by playing hundreds of games vs rated club players.

I also think it could be true because CCRL rating list is accurate, since there's engines that played in real tournaments vs human Grand Masters and engines rated 2600 in CCRL won the tournament performing at 2938 rating not losing a single game. You can see it here:

https://www.hiarcs.com/Games/Mercosur2009/mercosur09.htm

The rest of CCRL engines ratings are calibrated with these engines that beated human tournaments according to Elo formula. I checked it out and i measured ratings CCRL list vs Chessmaster personalities in Arena and they are pretty accurate, only +-100 points error.

The strongest personalities in Chessmaster (historical GM personalites) rating was determined and checked by real Grand Masters, including Larry Christiansen and Joshua Waitzkin. That's why former USA champion Larry Christiansen lost vs Chessmaster.

But at the same time i don't believe it, because i don't find it difficult to beat chessmaster 2000+ rated engines and here in chess.com 1100 players destroy me with total accuracy and precision. I suppose it's because i play engines at slow time controls and in chess.com i only play blitz or bullet from years ago. So i could be very bad at fast chess and not so bad at slow chess. Anyways, I also have read many forums where weak players claim the same thing, beating grand master ratings in chessmaster, for example here:

https://www.chess.com/forum/view/general/chessmaster-personalities-1

So, there's something strange that i don't understand. Chessmaster ratings must be wrong i suppose, but at the same time they are perfectly right and accurate.