Engine's "ELO" revisited

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

User avatar
Ponti
Posts: 493
Joined: Wed Mar 15, 2006 6:13 am
Location: Curitiba - PR - BRAZIL

Re: Engine's "ELO" revisited

Post by Ponti »

carldaman wrote: Thu Dec 19, 2019 10:31 pm
MikeB wrote: Thu Dec 19, 2019 6:56 am
MikeB wrote: Thu Dec 19, 2019 5:42 am
carldaman wrote: Thu Dec 19, 2019 12:18 am
bob wrote: Wed Dec 18, 2019 5:20 am
Try the most recent version, 25.4... It has an "elo" command.
Hello Bob, where can v25.4 be obtained?
Thanks,
Carl
That would be me - I got distracted for a few days. Will put it on Github in a few minutes.
Here you go:

https://github.com/MichaelB7/Crafty
Thanks Mike & Bob, enjoy your holidays! :)
Are there any Windows 64 bit binaries, or do I have to compile it myself ?
A. Ponti
AMD Ryzen 1800x, Windows 10.
FIDE current ratings: standard 1913, rapid 1931
User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Engine's "ELO" revisited

Post by xr_a_y »

Canoike wrote: Sat Dec 21, 2019 11:12 pm
xr_a_y wrote: Thu Dec 19, 2019 8:53 am If you give a try to Minic level function from 1 to 100, I d be glad to have feedback.
I compiled minic with the syzygy option and got a Linux executable called minic_WITH_SYZYGY_linux_x64. The last line of the JSON file is as follows :
"syzygyPath": "/home/user/jeux/echecs/syzygy/"
1) I must use the UCI protocol in order to be able to start a classic game, Winboard 1 or 2 protocol do not allow to play.
2) I tested the endgame with bishop and knight vs a lonely king. The score is about +99.8 but it cannot mate.
Thanks a lot for testing and feedback. I have a few questions and remarks.

- When you say you got a so-called executable, did you choose the name by yourself, if not, there is probably something wrong with the way you are activating SYZYGY in Minic.

To be sure TB are activated, run Minic from command line with the -quiet 0 option and look for this in the log

Code: Select all

# Info  2019-12-22 10:41:52-355: Init tb from path /ssd/Minic/TB/syzygy
# Info  2019-12-22 10:41:52-357: MAX_TB_MEN: 5
You can also run a little test this way

Code: Select all

./your_minic_exe -analyze "3k4/8/8/8/8/8/8/3KBN2 w - - 0 1" 60 -quiet 0
and look for this

Code: Select all

# Info  2019-12-22 10:46:02-484: TBHit1 8355
# Info  2019-12-22 10:46:02-484: TBHit2 66
- I know there are some GUI that do no work well with Minic but I'd like to fix that. With what GUI did you have issue with xboard ?

- If SYZYGY were working, you won't get a 99.8 score immediately, it would be a 68+20=88 score ("known eval win + TB win" inside Minic) and only get a 99 when the mate is sure.

- In fact with bishop and knight against lone king, Minic do not need TB, it already have the necessary knowledge

With SYZYGY activated, it play like this at very fast TC

[pgn]
[Variant "From Position"]
[FEN "3k4/8/8/8/8/8/8/3KNB2 w - - 0 1"]
1. Kd2 {+68,41/43} Kd7 {-68,41/40} 2. Ke3 {+68,41/39} Ke7 {-68,41/48} 3. Ke4 {+68,41/48} Kf6 {-68,41/49} 4. Nd3 {+68,41/58} Ke7 {-68,41/45} 5. Kd5 {+68,51/47} Kf6 {-68,51/37} 6. Bh3 {+68,51/54} Kg7 {-68,51/47} 7. Ke6 {+68,61/46} Kh8 {-68,51/45} 8. Kf6 {+68,61/54} Kh7 {-68,51/48} 9. Bf5+ {+68,61/55} Kh8 {-68,61/47} 10. Be4 {+68,61/51} Kg8 {-68,71/48} 11. Nb2 {+68,61/56} Kf8 {-68,71/52} 12. Bd5 {+68,81/47} Ke8 {-68,71/54} 13. Be6 {+98,78/39} Kd8 {-68,81/44} 14. Nd3 {+99,65/47} Kc7 {-68,81/44} 15. Ke7 {+99,67/37} Kc6 {-68,91/44} 16. Bb3 {+99,71/38} Kb5 {-98,86/50} 17. Kd6 {+99,73/36} Kb6 {-99,72/50} 18. Ba4 {+99,75/34} Ka6 {-99,74/36} 19. Bc6 {+99,77/37} Kb6 {-99,78/40} 20. Ne5 {+99,79/36} Ka5 {-99,80/41} 21. Kc5 {+99,81/31} Ka6 {-99,82/40} 22. Nf7 {+99,83/30} Ka7 {-99,86/46} 23. Kb5 {+99,87/43} Kb8 {-99,88/42} 24. Kb6 {+99,89/40} Kc8 {-99,90/45} 25. Bb5 {+99,91/51} Kb8 {-99,92/46} 26. Ba6 {+99,93/47} Ka8 {-99,94/121} 27. Ne5 {+99,95/95} Kb8 {-99,96/121} 28. Nd7+ {+99,97/121} Ka8 {-99,98/121} 29. Bb7# {+99,99/121}
[/pgn]

And without

[pgn]
[Variant "From Position"]
[FEN "3k4/8/8/8/8/8/8/3KNB2 w - - 0 1"]
1. Kd2 {+68,51/43} Kd7 {-68,41/40} 2. Ke3 {+68,51/43} Ke7 {-68,41/48} 3. Ke4 {+68,51/39} Kf6 {-68,41/48} 4. Nd3 {+68,51/42} Ke7 {-68,41/45} 5. Bg2 {+68,51/48} Kf7 {-68,41/43} 6. Ke5 {+68,51/51} Kf8 {-68,41/46} 7. Bd5 {+68,71/42} Kg7 {-68,41/53} 8. Nf4 {+68,81/41} Kh6 {-68,51/44} 9. Kf6 {+68,81/54} Kh7 {-68,61/54} 10. Bc4 {+68,71/49} Kh6 {-68,61/50} 11. Bb3 {+68,81/53} Kh7 {-68,61/55} 12. Ng6 {+68,81/55} Kh6 {-68,61/56} 13. Bg8 {+68,91/52} Kh5 {-68,91/43} 14. Ne5 {+68,91/53} Kh4 {-68,91/48} 15. Kf5 {+68,91/56} Kg3 {-68,91/46} 16. Ng4 {+68,91/50} Kg2 {-68,91/51} 17. Bc4 {+68,91/54} Kg3 {-70,28/52} 18. Be2 {+68,91/39} Kh3 {-99,70/53} 19. Kf4 {+99,81/50} Kg2 {-99,82/42} 20. Bc4 {+99,83/54} Kh3 {-99,84/42} 21. Ne3 {+99,85/56} Kh4 {-99,86/38} 22. Nf5+ {+99,87/51} Kh3 {-99,84/27} 23. Bf1+ {+99,89/56} Kh2 {-99,90/50} 24. Kf3 {+99,91/54} Kg1 {-99,92/53} 25. Ne3 {+99,93/63} Kh1 {-99,94/121} 26. Kf2 {+99,95/116} Kh2 {-99,96/121} 27. Ng4+ {+99,97/120} Kh1 {-99,98/121} 28. Bg2# {+99,99/121}
[/pgn]

At too fast TC, without TB, it is possible sometimes that Minic cannot do it before 50 moves rules, but this is quite rare.
Uri
Posts: 473
Joined: Thu Dec 27, 2007 9:34 pm

Re: Engine's "ELO" revisited

Post by Uri »

In my opinion Chess is for rich people only because only rich people can afford to buy such massively powerful computers.

I'm still using a 4 quad core computer because I simply can't afford to buy a faster computer.

I'm also living in poverty in general and I may also become homeless sometime in the future because I'm dealing with various legal and health problems currently.
User avatar
Ponti
Posts: 493
Joined: Wed Mar 15, 2006 6:13 am
Location: Curitiba - PR - BRAZIL

Re: Engine's "ELO" revisited

Post by Ponti »

xr_a_y wrote: Thu Dec 19, 2019 8:53 am If you give a try to Minic level function from 1 to 100, I d be glad to have feedback.
What is the best version to run with Ryzen 7 1800X ?
x64_avx2_bmi2 <======= I guess this is it ?
x64_sse4.2
or
x64_x86-64 ?
A. Ponti
AMD Ryzen 1800x, Windows 10.
FIDE current ratings: standard 1913, rapid 1931
mehmet1921
Posts: 12
Joined: Tue Sep 10, 2019 9:58 am
Full name: mehmet karaman

Re: Engine's "ELO" revisited

Post by mehmet1921 »

Uri wrote: Sun Dec 22, 2019 10:57 am In my opinion Chess is for rich people only because only rich people can afford to buy such massively powerful computers.

I'm still using a 4 quad core computer because I simply can't afford to buy a faster computer.

I'm also living in poverty in general and I may also become homeless sometime in the future because I'm dealing with various legal and health problems currently.
Chess is all people because Stockfish in a Pentium cpu is enough to defeat mankind.

But to become a good chess engine/ opening book tester a powerful system is necessary.
User avatar
xr_a_y
Posts: 1871
Joined: Sat Nov 25, 2017 2:28 pm
Location: France

Re: Engine's "ELO" revisited

Post by xr_a_y »

Ponti wrote: Sun Dec 22, 2019 8:14 pm
xr_a_y wrote: Thu Dec 19, 2019 8:53 am If you give a try to Minic level function from 1 to 100, I d be glad to have feedback.
What is the best version to run with Ryzen 7 1800X ?
x64_avx2_bmi2 <======= I guess this is it ?
x64_sse4.2
or
x64_x86-64 ?
You're right.
Canoike
Posts: 125
Joined: Tue Jan 17, 2012 8:08 pm

mimic : syzygy

Post by Canoike »

The compilation did not handle the syzygy egtb because I had not properly understood the doc : "activate WITH_SYZYGY definition at compile time". I had simply typed "./build.sh WITH_SYZYGY". This is the reason why the output executable named was "minic_WITH_SYZYGY_linux_x64".
I went to the "minic.cc" file and removed the two slashes in order to have this line : "#define WITH_SYZYGY". I compiled again "./build.sh" and now when using the command line and testing 2 knights vs 1 pawn (difficult ending, no engine has the knowledge for it: :D )

Code: Select all

# Info  2019-12-22 22:22:32-942: From config file, syzygyPath : /home/user/jeux/echecs/syzygy/
# Info  2019-12-22 22:22:33-796: Init tb from path /home/user/jeux/echecs/syzygy/
# Info  2019-12-22 22:22:33-819: MAX_TB_MEN: 5
# Info  2019-12-22 22:22:33-819: Reading fen 8/4k3/3p4/3K4/8/2N1N3/8/8 w - - 0 1
# Info  2019-12-22 22:22:33-829: TBHit1 204
# Info  2019-12-22 22:22:33-829: TBHit2 41
seems to work fine. Moreover I had compiled Fathom and by reading the instructions, I ran successfully the test.

Concerning the GUI, I used Xboard and Arena for Linux. I got a strange result. From the FEN position given above, in a 5 minutes ending, the evaluation was around +24 then minic chose to play inaccurate moves giving too much space to the weak King. The evaluation dropped at +3. Of course it was unable to mate.
User avatar
Ponti
Posts: 493
Joined: Wed Mar 15, 2006 6:13 am
Location: Curitiba - PR - BRAZIL

Re: Engine's "ELO" revisited

Post by Ponti »

ydebilloez wrote: Fri Dec 20, 2019 10:22 am Try any of these XBoard/UCI engines:
These should remain bad even at fastest CPU. Start from bottom for weakest.
I got trouble in getting the binaries for a lot of them, good luck.

1: Fimbulwinter 5.02
2: DarkFusch 0.09
3: Qutechess 1.01t rev3 ja
4: Pierre 1.7 ja
5: Kace 0.82 ja
6: Raffaela 0.14
7: Gringo 1.49b
8: Kcchess pre-alpha2 rev 1.13
9: Xadreco 5.7
10: Strategicdeep 1.25
11: Dreamer 0.2.0
12: Delfi 1000 ELO
13: Giuchess 1.0b2dc
14: Butterfly 0.1.7_alpha
15: Nanook 0.17
16: Bigbook 3.1
17: Dika 0.4209
18: Ochess rev2 ja
19: Mfchess 1.3
20: Luzhin 0.1
21: Youk 1.05
22: Belofte 0.2.8
23: Pyotr Am 0.6
24: Cheops 1.1
25: Dendron 0.42b ja
26: Geko 0.43
27: Neg 0.3d
28: Sachy 0.2.01
29: Prechess 0.7.8
30: Microchess1976
31: Tri-OS
32: Ram 2.0
33: Omar 3.1
34: Ace v0.1
35: Neurone I
36: Cpp1
37: Brutus RND
38: Deep Duke 1.00ja
39: NSVChess 0.14
40: Pos-1.20
NIce list !
In fact I was searching for an engine with "elo strengh" feature (like crafty, prodeo, shredder... ). I've downloaded some of these... I can beat Fimbulwinter without much effort. I've played Pos-1.20 at freechess.org, checkmated it in 3 moves.
A. Ponti
AMD Ryzen 1800x, Windows 10.
FIDE current ratings: standard 1913, rapid 1931
User avatar
Ponti
Posts: 493
Joined: Wed Mar 15, 2006 6:13 am
Location: Curitiba - PR - BRAZIL

Re: Engine's "ELO" revisited

Post by Ponti »

xr_a_y wrote: Thu Dec 19, 2019 8:53 am If you give a try to Minic level function from 1 to 100, I d be glad to have feedback.
How strong is it at 50% ?

[pgn]
[Event "Blitz 3 min+ 2 sec/move"]
[Site "Ryzen 1800x"]
[Date "2019.12.23"]
[Round "1"]
[White "Ponti,A."]
[Black "Minic 1.20 (C) @50%"]
[Result "1-0"]
[BlackElo "?"]
[ECO "B15"]
[Opening "Caro-Kann"]
[Time "00:23:40"]
[Variation "Gurgenidze: 4.e5"]
[WhiteElo "1931"]
[TimeControl "180+2"]
[Termination "normal"]
[PlyCount "80"]
[WhiteType "human"]
[BlackType "program"]

1. e4 g6 2. d4 c6 3. Nc3 d5 4. e5 a5 5. Be3 Be6 6. Qd2 Bd7 7. h4 e6 8. h5
gxh5 9. Be2 h6 10. Bxh5 Ne7 11. g4 b5 12. Nh3 Na6 13. Ne2 Bg7 14. f4 Qc7
15. Ng3 Nb4 16. a3 Na6 17. O-O Qd8 18. f5 exf5 19. gxf5 Qb6 20. f6 Bxf6 21.
Rxf6 Bxh3 22. Rxf7 Rg8 23. Kh2 Bg4 24. Raf1 Kd7 25. Qf2 Be6 26. Rh7 Qc7 27.
Nf5 Bxf5 28. Qxf5+ Kd8 29. Qe6 Ra7 30. Rff7 Qd7 31. Qxh6 Kc8 32. Rxe7 Qxe7
33. Rxe7 Rxe7 34. Bg5 Rd7 35. Bg4 Nb8 36. Bxd7+ Kb7 37. Bh3 Kb6 38. Bf6 Re8
39. e6 Rg8 40. Bg7 c5 {Black resigns} 1-0
[/pgn]
A. Ponti
AMD Ryzen 1800x, Windows 10.
FIDE current ratings: standard 1913, rapid 1931
User avatar
Ovyron
Posts: 4556
Joined: Tue Jul 03, 2007 4:30 am

Re: Engine's "ELO" revisited

Post by Ovyron »

Uri wrote: Sun Dec 22, 2019 10:57 am I'm still using a 4 quad core computer because I simply can't afford to buy a faster computer.
If it can reach depth 22 within 1 minute then it should be fast enough for any purpose, you don't need to be rich.