I am posting this for Larry Kaufma, who is traveling now:
" KomodoChess.com has released Dragon 3.1, which today won the World Computer Chess Championship after a tiebreaking match with 
Lc0. It is essentially the same version that won the WCCC, except for adjustments to limited Elo levels. It fixes some issues in 
Dragon 3 (occasional time forfeits in standard mode and crashes in MCTS mode), and so is free for anyone who bought Dragon 3 (or got
it by subscription). It has a newer net and search improvements compared to Dragon 3, which make it about ten elo stronger with 
normal openings (at CCRL blitz tc, one thread), twelve elo stronger at FRC (chess960), and twenty elo stronger with "unbalanced human 
openings". It also has a user-settable minimum time per move called "Skill Time" when using the LimitStrength option, and now allows for Elo setting up 
to 3499 (3500 means full strength). As usual, it is free for subscribers, and 20% discounted for purchasers of Dragon 2 thru 2.6. No 
price changes from Dragon 3 "
Note I will be sending out copies to rating list testers starting tomorrow.
			
			
									
						
										
						Dragon 3.1 Released at KomodoChess.com
Moderator: Ras
- 
				Chessqueen
- Posts: 5685
- Joined: Wed Sep 05, 2018 2:16 am
- Location: Moving
- Full name: Jorge Picado
- 
				peter
- Posts: 3419
- Joined: Sat Feb 16, 2008 7:38 am
- Full name: Peter Martan
Re: Dragon 3.1 Released at KomodoChess.com
Thanks a lot, Mark!
At the moment only Dragon 3 is offered for download to me, or is it the same name as for the link to click?
Download Dragon 3 ist says,
regards
Peter.
			
						- 
				criko
- Posts: 155
- Joined: Wed Feb 28, 2018 2:50 pm
Re: Dragon 3.1 Released at KomodoChess.com
the name of  the zip is Dragon3 ... but the zip contains version 3.1
			
			
									
						
										
						- 
				Fritz 0
- Posts: 151
- Joined: Fri Mar 11, 2022 12:10 pm
- Full name: Branislav Đošić
Re: Dragon 3.1 Released at KomodoChess.com
In which way are the Elo levels adjusted? Are they made stronger or weaker? And in which range?mjlef wrote: ↑Sat Jul 30, 2022 4:14 am I am posting this for Larry Kaufma, who is traveling now:
" KomodoChess.com has released Dragon 3.1, which today won the World Computer Chess Championship after a tiebreaking match with
Lc0. It is essentially the same version that won the WCCC, except for adjustments to limited Elo levels. It fixes some issues in
Dragon 3 (occasional time forfeits in standard mode and crashes in MCTS mode), and so is free for anyone who bought Dragon 3 (or got
it by subscription). It has a newer net and search improvements compared to Dragon 3, which make it about ten elo stronger with
normal openings (at CCRL blitz tc, one thread), twelve elo stronger at FRC (chess960), and twenty elo stronger with "unbalanced human
openings". It also has a user-settable minimum time per move called "Skill Time" when using the LimitStrength option, and now allows for Elo setting up
to 3499 (3500 means full strength). As usual, it is free for subscribers, and 20% discounted for purchasers of Dragon 2 thru 2.6. No
price changes from Dragon 3 "
Note I will be sending out copies to rating list testers starting tomorrow.
- 
				peter
- Posts: 3419
- Joined: Sat Feb 16, 2008 7:38 am
- Full name: Peter Martan
- 
				mjlef
- Posts: 1494
- Joined: Thu Mar 30, 2006 2:08 pm
Re: Dragon 3.1 Released at KomodoChess.com
Fritz 0 wrote: ↑Sat Jul 30, 2022 10:46 amIn which way are the Elo levels adjusted? Are they made stronger or weaker? And in which range?mjlef wrote: ↑Sat Jul 30, 2022 4:14 am I am posting this for Larry Kaufma, who is traveling now:
" KomodoChess.com has released Dragon 3.1, which today won the World Computer Chess Championship after a tiebreaking match with
Lc0. It is essentially the same version that won the WCCC, except for adjustments to limited Elo levels. It fixes some issues in
Dragon 3 (occasional time forfeits in standard mode and crashes in MCTS mode), and so is free for anyone who bought Dragon 3 (or got
it by subscription). It has a newer net and search improvements compared to Dragon 3, which make it about ten elo stronger with
normal openings (at CCRL blitz tc, one thread), twelve elo stronger at FRC (chess960), and twenty elo stronger with "unbalanced human
openings". It also has a user-settable minimum time per move called "Skill Time" when using the LimitStrength option, and now allows for Elo setting up
to 3499 (3500 means full strength). As usual, it is free for subscribers, and 20% discounted for purchasers of Dragon 2 thru 2.6. No
price changes from Dragon 3 "
Note I will be sending out copies to rating list testers starting tomorrow.
- 
				lkaufman
- Posts: 6259
- Joined: Sun Jan 10, 2010 6:15 am
- Location: Maryland USA
- Full name: Larry Kaufman
Re: Dragon 3.1 Released at KomodoChess.com
The number of nodes looked at was reduced for elos starting with a "2" because Dragon 3.1 is stronger than Dragon 3 with equal nodes in this elo range. The actual strength of Dragon 3.1 with Elo set to 2300 (or any other number) should be roughly equal to Dragon 3 with same Elo setting.Fritz 0 wrote: ↑Sat Jul 30, 2022 10:46 amIn which way are the Elo levels adjusted? Are they made stronger or weaker? And in which range?mjlef wrote: ↑Sat Jul 30, 2022 4:14 am I am posting this for Larry Kaufma, who is traveling now:
" KomodoChess.com has released Dragon 3.1, which today won the World Computer Chess Championship after a tiebreaking match with
Lc0. It is essentially the same version that won the WCCC, except for adjustments to limited Elo levels. It fixes some issues in
Dragon 3 (occasional time forfeits in standard mode and crashes in MCTS mode), and so is free for anyone who bought Dragon 3 (or got
it by subscription). It has a newer net and search improvements compared to Dragon 3, which make it about ten elo stronger with
normal openings (at CCRL blitz tc, one thread), twelve elo stronger at FRC (chess960), and twenty elo stronger with "unbalanced human
openings". It also has a user-settable minimum time per move called "Skill Time" when using the LimitStrength option, and now allows for Elo setting up
to 3499 (3500 means full strength). As usual, it is free for subscribers, and 20% discounted for purchasers of Dragon 2 thru 2.6. No
price changes from Dragon 3 "
Note I will be sending out copies to rating list testers starting tomorrow.
Komodo rules!
			
						- 
				cc2150dx
- Posts: 408
- Joined: Sat Nov 30, 2013 9:51 am
- Full name: Jason Coombs
Re: Dragon 3.1 Released at KomodoChess.com
Congratulations on the release of Dragon 3.1   
			
			
									
						
										
						
- 
				Frank Quisinsky
- Posts: 7105
- Joined: Wed Nov 18, 2009 7:16 pm
- Location: Gutweiler, Germany
- Full name: Frank Quisinsky
Re: Dragon 3.1 Released at KomodoChess.com
Hi Larry & Mark,
thank you for the new version!
Note for Larry:
I am testing with contempt=0 and move average is great.
No longer all the 300 moves games without resign.
Example:
Run-6 from FCP-Tourney-KI without resign mode (40 in 20).
White and Black draws = HORROR
Nobody will like that ... very high quantity of games ... waste of time and electricity and so many boring draw endgames.
Results from a test tourney without contempt (40 in 8):
Dragon 3 NN (Komodo) purrs like a cat.
Average for white draw = 37 moves lesser without the bad contempt standard settings.
Average for black draw = 41 moves lesser without the bad contempt standard settings.
Sure that anyone like to see this bad standard settings in Komodo, absolutey sure?
For exactly 3 Elo more an average of 37/41 moves for draw games higher?
I am very sure, that nobody will like that for 3 Elo ... or better 2.7 Elo (test that).
Sure the Dragon team have testers that can confirm.
Or what do the testers?
Best
Frank
Please do not artificially increase the average move-number of draws in times when there are too many draws.
Produces an unnecessary additional problem!!
Standards settings should be changed!
			
			
													thank you for the new version!
Note for Larry:
I am testing with contempt=0 and move average is great.
No longer all the 300 moves games without resign.
Example:
Run-6 from FCP-Tourney-KI without resign mode (40 in 20).
Code: Select all
10. Dragon 3 NN (Komodo)           (695+,   502=,   3-)  78.8%
          CATEGORIES   MOVE_AVE    GAMES    TOTAL  GAME_RATE
           all games         98     1200
           white won         76      465
           black won         82      230
          white draw        132      135
          black draw        122      367
          white loss          0        0
          black loss        125        3
            all wins         78      695     1200      57.9%
           all draws        125      502     1200      41.8%
           all loses        125        3     1200       0.2%
Nobody will like that ... very high quantity of games ... waste of time and electricity and so many boring draw endgames.
Results from a test tourney without contempt (40 in 8):
Code: Select all
 4. Dragon 3 NN (Komodo)           (511+,   438=,   0-)  76.9%
          CATEGORIES   MOVE_AVE    GAMES    TOTAL  GAME_RATE
           all games         82      949
           white won         79      343
           black won         83      168
          white draw         95      132
          black draw         81      306
          white loss          0        0
          black loss          0        0
            all wins         80      511      949      53.8%
           all draws         85      438      949      46.2%
           all loses          0        0      949       0.0%
Average for white draw = 37 moves lesser without the bad contempt standard settings.
Average for black draw = 41 moves lesser without the bad contempt standard settings.
Sure that anyone like to see this bad standard settings in Komodo, absolutey sure?
For exactly 3 Elo more an average of 37/41 moves for draw games higher?
I am very sure, that nobody will like that for 3 Elo ... or better 2.7 Elo (test that).
Sure the Dragon team have testers that can confirm.
Or what do the testers?
Best
Frank
Please do not artificially increase the average move-number of draws in times when there are too many draws.
Produces an unnecessary additional problem!!
Standards settings should be changed!
					Last edited by Frank Quisinsky on Sat Jul 30, 2022 11:35 pm, edited 2 times in total.