Critter 0.90 1cpu tests

Discussion of computer chess matches and engine tournaments.

Moderator: Ras

User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Critter 0.90 1cpu tests

Post by beachknight »

Code: Select all

    Engine               Score Cr
01: Critter 0.90 1cpu    0.0/0 · 
01: Houdini 1.03a 1cpu   0.0/0   
01: Rybka 4 1cpu         0.0/0   
01: IvanHoe T52A 1cpu    0.0/0   
01: Tankist 1.7 w        0.0/0   
01: FireBird 1.0.1 1cpu  0.0/0   
01: Stockfish 1.9.1 1cpu 0.0/0   
01: Saros 2.3c4 w        0.0/0   
01: Sjeng ct10 1cpu w    0.0/0   
01: Naum 4.2 1cpu        0.0/0   
01: Gull 1.1             0.0/0   

0 of 60 games played
Name of the tournament: CET30001
Site/ Country: CET_Antalya_TUR, Türkiye
Level: Blitz 3/3
hi, merhaba, hallo HT
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Critter 0.90 1cpu [64 bit] tests

Post by beachknight »

Book: RpC-SF-0.9A11.abk (limited to 7 moves or 14 halfmoves)

Critter 0.8 1cpu [64 bit] = 3183 (343 games)

Expected ratings increase : 40 - 50 (according to Richard Vida)

Critter 0.90 1cpu expected rating : 3243 - 3253
hi, merhaba, hallo HT
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Critter 0.90 1cpu [64 bit] tests

Post by beachknight »

Code: Select all

    Engine               Score    Cr
01: Critter 0.90 1cpu    15.5/30 ··· 
02: Houdini 1.03a 1cpu   2.5/3   =11 
03: Rybka 4 1cpu         2.0/3   =1= 
03: FireBird 1.0.1 1cpu  2.0/3   110 
05: IvanHoe T52A 1cpu    1.5/3   1=0 
05: Stockfish 1.9.1 1cpu 1.5/3   === 
05: Sjeng ct10 1cpu w    1.5/3   0=1 
08: Tankist 1.7 w        1.0/3   ==0 
08: Naum 4.2 1cpu        1.0/3   =0= 
08: Gull 1.1             1.0/3   100 
11: Saros 2.3c4 w        0.5/3   =00 

30 of 60 games played
Name of the tournament: CET30001
Site/ Country: CET_Antalya_TUR, Türkiye
Level: Blitz 3/3
hi, merhaba, hallo HT
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Critter 0.90 1cpu [64 bit] tests

Post by beachknight »

Code: Select all

    Engine               Score       Cr
01: Critter 0.90 1cpu    33.5/60 ······ 
02: Houdini 1.03a 1cpu   5.0/6   =1111= 
03: Rybka 4 1cpu         3.5/6   =1==== 
03: FireBird 1.0.1 1cpu  3.5/6   110=== 
05: Tankist 1.7 w        3.0/6   ==01== 
06: IvanHoe T52A 1cpu    2.0/6   1=000= 
06: Stockfish 1.9.1 1cpu 2.0/6   ===00= 
06: Saros 2.3c4 w        2.0/6   =00=01 
06: Sjeng ct10 1cpu w    2.0/6   0=1=00 
06: Naum 4.2 1cpu        2.0/6   =0==0= 
11: Gull 1.1             1.5/6   1000=0 

60 games played / Tournament finished
Name of the tournament: CET30001
Site/ Country: CET_Antalya_TUR, Türkiye
Level: Blitz 3/3
hi, merhaba, hallo HT
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Critter 0.90 1cpu [64 bit] tests

Post by beachknight »

Code: Select all

CET30001 CET_Antalya_TUR  2010

                                      1      2      3      4      5      6      7      8      9      0      1      
1   Critter 0.90 1cpu     3200   +79  ****** ½0000½ ½0½½½½ 001½½½ ½½10½½ 0½111½ ½½½11½ ½11½10 1½0½11 ½1½½1½ 0111½1  33.5/60
2   Houdini 1.03a 1cpu    3327  +152  ½1111½ ******                                                                  5.0/6
3   Rybka 4 1cpu          3316   -57  ½1½½½½        ******                                                           3.5/6  117.25
4   FireBird 1.0.1 1cpu   3270   -11  110½½½               ******                                                    3.5/6  117.25
5   Tankist 1.7 w         3276   -76  ½½01½½                      ******                                             3.0/6
6   IvanHoe T52A 1cpu     3315  -235  1½000½                             ******                                      2.0/6  67.00
7   Stockfish 1.9.1 1cpu  3259  -179  ½½½00½                                    ******                               2.0/6  67.00
8   Saros 2.3c4 w         3211  -131  ½00½01                                           ******                        2.0/6  67.00
9   Sjeng ct10 1cpu w     3150   -70  0½1½00                                                  ******                 2.0/6  67.00
10  Naum 4.2 1cpu         3137   -57  ½0½½0½                                                         ******          2.0/6  67.00
11  Gull 1.1              3131  -121  1000½0                                                                ******   1.5/6

hi, merhaba, hallo HT
User avatar
beachknight
Posts: 3533
Joined: Tue Jan 09, 2007 8:33 pm
Location: Antalya, Turkey

Re: Critter 0.90 1cpu [64 bit] tests

Post by beachknight »

Start rating of Critter 0.90 1cpu : 3250
hi, merhaba, hallo HT
Engin
Posts: 1001
Joined: Mon Jan 05, 2009 7:40 pm
Location: Germany
Full name: Engin Üstün

Re: Critter 0.90 1cpu [64 bit] tests

Post by Engin »

wow, how come this huge strength improvements?

critter grows up in the last years rapidly, then i wondor how he doing this, i think it is impossible to doing such improvements in a short time, may he working very hard on critter or he is cheating and cloning from other sources.

OR he can MAGIC !?

i am doing this hobby, but have only reached about 2700 elo , in this way i am now a little demotivated from computerchess :(
Dann Corbit
Posts: 12791
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Critter 0.90 1cpu [64 bit] tests

Post by Dann Corbit »

Engin wrote:wow, how come this huge strength improvements?

critter grows up in the last years rapidly, then i wondor how he doing this, i think it is impossible to doing such improvements in a short time, may he working very hard on critter or he is cheating and cloning from other sources.

OR he can MAGIC !?

i am doing this hobby, but have only reached about 2700 elo , in this way i am now a little demotivated from computerchess :(
Every engine has value. Do you know my second favorite engine for personal play? It's Golem. If you put up a screen and had someone relay you the moves in words, you would swear you were playing a human.

You have produced a huge pile of useful computer chess tools. I can tell you that I, personally, am very grateful for your efforts. There are probably many people that only look at the Elo value on some list to decide what to like. Those same people will have no idea if it is TSCP or Rybka 4 that is beating their pants off.

IMO-YMMV
Jouni
Posts: 3645
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: Critter 0.90 1cpu [64 bit] tests

Post by Jouni »

I am also interested how Critter development is done - has Richard written anything on programming forum? I have feeling, that today you need 10s of computers AND testers to verify smallest improvement. And Vida does couple of changes and +100 !! IMPRESSIVE definitely.

Jouni
Dann Corbit
Posts: 12791
Joined: Wed Mar 08, 2006 8:57 pm
Location: Redmond, WA USA

Re: Critter 0.90 1cpu [64 bit] tests

Post by Dann Corbit »

Engin wrote:wow, how come this huge strength improvements?

critter grows up in the last years rapidly, then i wondor how he doing this, i think it is impossible to doing such improvements in a short time, may he working very hard on critter or he is cheating and cloning from other sources.

OR he can MAGIC !?

i am doing this hobby, but have only reached about 2700 elo , in this way i am now a little demotivated from computerchess :(
As for the "How" of it, I think the following:
Almost all of the chess programmers are examining the code of strong engines like Fruit, Stockfish, and IvanHoe. If you are just reading the code for understanding the ideas, in my opinion there is totally nothing wrong with that. And if you use the ideas (I am not talking about cut and paste) there is nothing wrong with that either.

Literally every chess engine on earth is, in some sense, a clone. That is because nobody, not even Dr. Hyatt, has invented every idea in their chess engines. Sometimes these ideas are lifted from chess papers written by chess programmers. Sometimes these ideas are lifted from examination of someone else's code. Sometimes these ideas are lifted from reading responses in boards like this one. Sometimes these ideas are lifted in conversations with other chess programmers. Sometimes these ideas are lifted in collaboration with other chess programmers.

There is nothing wrong with that. Now, personal satisfaction can come from:
1. Understanding an algorithm and writing your own version.
2. Subsequent improvement of that algorithm with your own fresh ideas.
3. Invention of a totally novel way of doing things.
And many other ways.

If someone were to grab 50K lines of chess code, ram in a stick and stir it around, and then slap their name on it, where will the personal satisfaction come from that? I totally don't understand it.

If you were to take the low road and tweak some giant killer, probably a group of people would hail you as a hero. But you will still need to answer to your own heart at some point.

So I recommend the Spike Lee quotation: "Always do the right thing."

What I think is the right thing is not right for everyone. Some would be deeply offended at reading someone else's code, understanding it and writing my own version. If they put a patent on their algorithm, and publish that fact, I would not duplicate the idea (and I would also not like their choice, but would obey it). Right and wrong are clouded in this arena by the fierce competition. I understand this also. If you have labored for 5 years to write your own chess engine from the ground up and someone else steals someone else's code, change literally 5 lines, and then beat you in a big public contest then I think you have a right to be flaming mad.

Here is my bottom line:
No matter what you do, some people will hate you for it and some people will love you for it. Some will hail you as a hero and some will pan you as a demon. So be true to your own heart and I think you will find the path of chess programming happiness.

P.S.
It's not a path to riches.