What do people actually do with chess engines these days?

[Peter Berger]

In the online forum of the dead German CSS magazine there was a (kind of) interesting recent heated discussion about chess System Tal extreme that made me think about this question. I am not a user there, so I steal the topic. I am talking users here, not developpers.

1. Analyze positions and games

Safe bet, chess players will do that, and they have very good reasons for it. So you download Stockfish, you maybe download lc0, too. There is no need for any other chess engine at all.

2. You play against them.

Why would you? You can play against all kinds of human players most easily these days, why play against some chess engine instead? Is it fun?

3. You do engine rating lists.

OK - you'll obviously want +loads+ of different chess engines for that and maybe have some additional personal wishes. There are plenty of them engines, so no idea why one more or less should matter in any way.

4. You look at engines solving chess problems or even create chess problems of your own.

OK, so you'll maybe download some other engines that really excel at that.

5. You pit some engines against each other to enjoy watching the resulting games.

I do this myself. It is a bit useless hobby, but it can be fun. As I do it myself, it feels very reasonable to me of course. I don't need any additional tools for that, I'll just go with what I feel like watching.

In the CSS forum, people were demanding additional features each engine should have.

1. They should run on Linux.

This one I can understand. My computers have Windows, but if they were running Linux, I'd clearly be only interested in engines that run on it.

2. They should also be able to use bazillions of processors et al.

Why? This one I +really+ don't get at all. And why are people getting so emotional with this demand? Don't they know that Stockfish is the strongest engine anyway?

Cheers.
Peter

[swami]

1. Yes, the top engine is usually sufficient for our kind of mistakes in games. It mainly points out inaccuracies, mistakes, and blunders.

2. Yes, it is a lot of fun to play against various engines in the same rating range as me, using only an electronic board. I get to see different playing styles, and it motivates me to try and do better. The only issue is deciding whether to allow myself one takeback or no takebacks at all.

Human players, when playing online, or OTB, make blunders, and that can end the game instantly. There isn’t much to learn from it. I don’t get to play similarly rated players all the time, it’s usually with large gaps of 50–150 Elo.

3. Yes, it’s fun to conduct tournaments between engines and see the points tally. I find it especially fun to run marathon format events with a minimum of 101 engines. Watching the games on a big screen is exciting, with the leaders changing every round.

4. This is also quite fun finding out which engine was able to solve a particular problem and which one wasn’t.

5. I don’t do this unless I am testing a new update of the engine.

Overall, computer chess is great fun.

[cpeters]

In the online forum of the dead German CSS magazine there was a (kind of) interesting recent heated discussion about chess System Tal extreme that made me think about this question. I am not a user there, so I steal the topic. I am talking users here, not developpers.

1. Analyze positions and games

Safe bet, chess players will do that, and they have very good reasons for it. So you download Stockfish, you maybe download lc0, too. There is no need for any other chess engine at all.

2. You play against them.

Why would you? You can play against all kinds of human players most easily these days, why play against some chess engine instead? Is it fun?

3. You do engine rating lists.

OK - you'll obviously want +loads+ of different chess engines for that and maybe have some additional personal wishes. There are plenty of them engines, so no idea why one more or less should matter in any way.

4. You look at engines solving chess problems or even create chess problems of your own.

OK, so you'll maybe download some other engines that really excel at that.

5. You pit some engines against each other to enjoy watching the resulting games.

I do this myself. It is a bit useless hobby, but it can be fun. As I do it myself, it feels very reasonable to me of course. I don't need any additional tools for that, I'll just go with what I feel like watching.

Hello Peter!

1. - sometimes, but then everywhere (mobile and at home/the "personal computer")
2. - yes; if you don't exclude emulated dedicated machines. On my phones I tend to solve puzzles (only on the iphone) and like to play against Rodent (because I can abort, pause and continue a game any time) or dedicated machines (only android with mame) or connect the e-board (thanks to Aart Bik)
3./4./5. - not bothered

1. They should run on Linux.

Software I intend to use privately must run on the devices I own and under the operating systems I prefer (Linux, iOS/macOS, *BSD). There would be no "demand" from my side - if something is not available (natively) - why should I use it? Where would be the fun of that?

2. They should also be able to use bazillions of processors et al.

Don't care about your program if you didn't care about portabilty. Doesn't work under arm and I'd like to use on my mobiles/the macbook? Thanks, hard pass. Exception: igel - but there are quirks (beeing free and open source) to compile for this architecture.

Look: The chess-genius is hornalt. Mr. Lang kindly ported it to the mobile platforms. Available both for android and ios. Splendid! Same with the Rodent: Mr. Pawel Koziol programmed it (with or without portability in mind) as free and open source, so that it can run under multiple operating systems and different processors (the bot on lichess lives in a NetBSD-environment (either an aarch64 or amd64 one)). Not that you would care, but I do, as computing (at home) should be fun and I do like Unix-like operating systems.


cheers

[Rebel]

In the online forum of the dead German CSS magazine there was a (kind of) interesting recent heated discussion about chess System Tal extreme that made me think about this question. I am not a user there, so I steal the topic. I am talking users here, not developpers.

I have read it, hilarious

After 40+ years of elo programming something new happened, the release of the EAS rating list, and I fell in love immediately. What reasonable worked with HCE code (Benjamin being the gambit version of ProDeo) worked 10 times better with net programming.

Finally something new and very creative and the bonus is a fantastic playing style and often unbelievable games. Stefan set something totally new in motion and it is my hope that others follow suit. Imagine the SF guys with their hardware and knowledge spend a few months of their time on style programming, taste the creativity and produce something mind blowing.

But as it is, this forum and apparently the CSS forum also are mainly about elo, not much interest in style programming. But the chess world is bigger than computer chess fora, it got the attention from Chessbase, resulting in two articles.

So Stefan, I will always remember you as the promoter and founder of style programming, hence it was a bit sad to read what you wrote elsewhere :

Eine Schande!

Mein Testrun läuft noch. Aber jetzt ist schon klar, daß CS Tal Extreme sich an die Spitze meiner EAS-Rangliste setzen wird, noch vor Patricia 5 xtremeEAS.
Das macht es so unglaublich bitter, daß diese "Engine" eigentlich gar keine ist und man fast nichts mit ihr anfangen kann.

[chrisw]

In the online forum of the dead German CSS magazine there was a (kind of) interesting recent heated discussion about chess System Tal extreme that made me think about this question. I am not a user there, so I steal the topic. I am talking users here, not developpers.

I have read it, hilarious

After 40+ years of elo programming something new happened, the release of the EAS rating list, and I fell in love immediately. What reasonable worked with HCE code (Benjamin being the gambit version of ProDeo) worked 10 times better with net programming.

Finally something new and very creative and the bonus is a fantastic playing style and often unbelievable games. Stefan set something totally new in motion and it is my hope that others follow suit. Imagine the SF guys with their hardware and knowledge spend a few months of their time on style programming, taste the creativity and produce something mind blowing.

But as it is, this forum and apparently the CSS forum also are mainly about elo, not much interest in style programming. But the chess world is bigger than computer chess fora, it got the attention from Chessbase, resulting in two articles.

So Stefan, I will always remember you as the promoter and founder of style programming, hence it was a bit sad to read what you wrote elsewhere :

Eine Schande!

Mein Testrun läuft noch. Aber jetzt ist schon klar, daß CS Tal Extreme sich an die Spitze meiner EAS-Rangliste setzen wird, noch vor Patricia 5 xtremeEAS.
Das macht es so unglaublich bitter, daß diese "Engine" eigentlich gar keine ist und man fast nichts mit ihr anfangen kann.

I read that stuff too. For some reason he appears to hate us. Weird.

[peter]

https://forum.computerschach.de/cgi-bin ... #pid177787

[cpeters]

hilarious

It is, isn't it? As is the like-button (and the counted bitcoinlikes) in the wixsite.

Wixen. Jerking. Cyclejerk.

https://rebelchess.com/
https://rebelchess.org
etc.
would be up for taking.

Just for appearances.

But as it is, this forum and apparently the CSS forum also are mainly about elo, not much interest in style programming

Don't know the specifics. Too complicated for me, obviously. Strength doesn't always matter - but availability and usability does.

greetings

[Rebel]

In the online forum of the dead German CSS magazine there was a (kind of) interesting recent heated discussion about chess System Tal extreme that made me think about this question. I am not a user there, so I steal the topic. I am talking users here, not developpers.

I have read it, hilarious

After 40+ years of elo programming something new happened, the release of the EAS rating list, and I fell in love immediately. What reasonable worked with HCE code (Benjamin being the gambit version of ProDeo) worked 10 times better with net programming.

Finally something new and very creative and the bonus is a fantastic playing style and often unbelievable games. Stefan set something totally new in motion and it is my hope that others follow suit. Imagine the SF guys with their hardware and knowledge spend a few months of their time on style programming, taste the creativity and produce something mind blowing.

But as it is, this forum and apparently the CSS forum also are mainly about elo, not much interest in style programming. But the chess world is bigger than computer chess fora, it got the attention from Chessbase, resulting in two articles.

So Stefan, I will always remember you as the promoter and founder of style programming, hence it was a bit sad to read what you wrote elsewhere :

Eine Schande!

Mein Testrun läuft noch. Aber jetzt ist schon klar, daß CS Tal Extreme sich an die Spitze meiner EAS-Rangliste setzen wird, noch vor Patricia 5 xtremeEAS.
Das macht es so unglaublich bitter, daß diese "Engine" eigentlich gar keine ist und man fast nichts mit ihr anfangen kann.

I read that stuff too. For some reason he appears to hate us. Weird.

I was surprised as well because after all we have given him what he wanted, massive EAS points, recognition of his invention.

[Uri Blass]

1)I prefer to use also other engines thar are nore stockfish and lc0 for analysis because:
a)Stockfish does not find good moves when one side has a big advantage(I proved it when stockfish lost matches with queen odds of who win in less moves against old Wasp.

b)Sometimes stockfish give evaluations that I do not understand and if I want to see some tactical mistakes I made I perfer t use more matertialistic engine so I see based on big changes in the evaluation what are the tactical mistakes in the game.

c)I do nor have a good hardware for lc0

2)I do not play against engines but I think playing agqainst engines may be productive for improvement in chess(almost never tried it so do not know)

If you have some opening preperation and you want to hide it but train playing the opening preperation then training against humans mean that some human see your opening preperation when if you train nly against engines then nobody can see your games.

Note that training against engines does not mean training against the strongest engines and you may choose engines with different styles or engines who try to play more similiar to humans in some rating target(not sure how much engines are good in it)

3)Engines list can be in different conditions.
For some reason I do not see people make queen odds rating list with the rule that the winner of 2 engines is the engine who win in less moves or a different queen odds list when you have some strong engines without queen d1 and some weak engines,

People tested even the random mover for rating lists so I see no problem to find weak engines that are not always going to win inspite of the fact that the opponent starts without queen d1(you can use also strong engines with fixed small number of nodes in the team of weak engines)

I also see no finding mates rating list when for this purpose you can use some pgn and give engines to analyze every position in the pgn for some different fixed times and order the engines by the number of mates they find or the engine that can find the shortest mate more often.

[ovenel]

If you have some opening preperation and you want to hide it but train playing the opening preperation then training against humans mean that some human see your opening preperation when if you train nly against engines then nobody can see your games.

Note that training against engines does not mean training against the strongest engines and you may choose engines with different styles or engines who try to play more similiar to humans in some rating target(not sure how much engines are good in it)

This is typically why I play against a computer. Every so often, I fall back into chess and try to learn a bit more opening theory. When reading about a particular opening, I like to start off by playing against an engine, referring to whatever literature I'm learning from and taking my time to consult those sources. I don't really have any people to play or study with, so my options are either an engine or games online, but only the engine is a viable option here if I want to be able to refer to notes about the opening.