Engines with limited strength

[chesskobra]

Holidays are over now, but maybe for the next holidays: Komodo, Dragon, Crafty, Phalanx, Wasp have skill levels. Lc0 and Rubichess have nps setting.

[tapio]

Any experiences with more humanly playing engines? I'm a weak player capable to even lose against Rodents weak School personalities. Patricia 4 at lowest skill still beats me. It's a concentration issue, my blunders can be horrible.

I think Rodent has some code to deal wirh human blunders. It usually instantly counter blunders, which I always see after a proper engine analysis. I don't like the instant counter blunder.

Do you know other engines with have some interesting blunder/bad movement management included? I wish developers put more thought into such things instead of boring elo achievements withgames which nohumanbeing can judge anyway.

[chesskobra]

Have you tried phalanx? It has -z option to randomize moves in centipawns, -e for easy levels from 0 to 100, where 0 is the default for best play, and -n for limiting nps.

[tapio]

Have you tried phalanx? It has -z option to randomize moves in centipawns, -e for easy levels from 0 to 100, where 0 is the default for best play, and -n for limiting nps.

Nice options, but unfortunately I'm on Android. uci options aren't exposed there in the Phalanx builds of Archimedes, somethings wrong with the compiles. (Phalanx XXV and XXII)

[chesskobra]

I have no idea about running phalanx on android. I didn't even know that there was an android build. But the options I mentioned are command line options on linux, which I can also configure in the .xboardrc file.

[Archimedes]

Have you tried phalanx? It has -z option to randomize moves in centipawns, -e for easy levels from 0 to 100, where 0 is the default for best play, and -n for limiting nps.

Nice options, but unfortunately I'm on Android. uci options aren't exposed there in the Phalanx builds of Archimedes, somethings wrong with the compiles. (Phalanx XXV and XXII)

Phalanx only supports the XBoard protocol.

[jefk]

phalanx apparently made a comeback and it looks interesting,
as mentioned here:
https://www.chess.com/forum/view/chess- ... lanx-xxiii

for my gambit book testing i most often use xboard engines
but haven't tried phalanx yet.

Positive suggestion:

instead of all theses xboard commands/options (for weakened play)
it would be more userfriendly if it had an .ini file (like
several other engines) where you can state eg. Elo, etc.

[Ras]

The CT800 engine (2700 CCRL) supports UCI_Elo: https://www.ct800.net/download
Should also scale with the move time. That's a little buggy on shorter time controls, but will still stay within 50 Elo difference, no nothing major.
Note that the Shredder GUI respects the Elo setting during normal play, but seems to ignore it in analysis.

[Whiskers]

Any experiences with more humanly playing engines? I'm a weak player capable to even lose against Rodents weak School personalities. Patricia 4 at lowest skill still beats me. It's a concentration issue, my blunders can be horrible.

I think Rodent has some code to deal wirh human blunders. It usually instantly counter blunders, which I always see after a proper engine analysis. I don't like the instant counter blunder.

Do you know other engines with have some interesting blunder/bad movement management included? I wish developers put more thought into such things instead of boring elo achievements withgames which nohumanbeing can judge anyway.

I’m 90% sure that Patricia had a bug where some GUIs would explicitly set both skill level AND uci_elo instead of leaving the other one at default, which means you might have been playing full strength Patricia!


It’s fixed and pushed to main now.

[jdart]

Hi,
I ran a large swiss tournament (64 competitors, 10 rounds); you might be interested in reading the outcome here.

One challenge in implementing UCI_Elo is that computer ratings from lists are generally on a different scale than human ratings, and computer ratings differ across testers.

I could not make Arasan play at low levels without introducing blunders, but blunders are frequent in human games at these low rating levels. I have some ideas about adjusting the low ELO behavior, but to make it beatable by low-rated players, I have found it necessary to produce moves that have a large score deviation from the optimal move.