Dragon 3.3 Released at KomodoChess.com

[Fritz 0]

KomodoChess.com has released Dragon 3.3, expected to be the final version of Dragon, primarily to fulfill obligations to subscribers who were still due for another upgrade. It is free for anyone who bought Dragon 3.2 or who received Dragon 3.2 via subscription. New purchasers of Dragon should receive 3.3 from now one.

Although Mark and I retired from running KomodoChess at the start of this year, turning it over to chess.com, we have continued to improve its search on a part-time basis, while Dietriche Kappe was able to upgrade the net to a 50% larger size with increased strength, still trained on Dragon moves. In addition, the Elo levels should be much more accurate now (aimed at 10' + 5" Rapid chess), due to a new methodology and also to correcting an issue that made the low Elo levels weaker than intended. We welcome feedback on the accuracy of the new levels for human opponents.

Dragon 3.3 is a normal size upgrade; with "UHO" openings we got about 12 elo plus (against Dragon 3.2) in bullet and 15 elo plus in our blitz testing, with the increased gains with longer time controls presumably due to the larger net. With normal openings, we got a +8 elo result in 40,000 single-thread games at 1' + 0.6", and have confirmed this with a +8 elo result after 11400 games at 5' + 2". Gains on 4 or 8 threads appear to be a bit better, at least not worse. MCTS and FRC play should also be a bit improved, although these were not specifically targeted since version 3.2. Considering how difficult it is to make further Elo progress from the current lofty levels, we feel that these gains fully justify this final release.

There may be a short technical delay in getting copies to testers. We hope it will be fully tested as the final Dragon version.

Larry, a few questions:

Why did you switch from 15'+10'' to 10'+5'' in the Elo levels?

Can you tell us more about that new methodology? Is it still based on node count and randomness?

Does Dragon 3.3 still has Regular Eval option? (just asking this because Stockfish removed it)

The switch in target time control was due to the fact that most online Rapid play is now at just 10' with no increment, so this is a compromise, and to the fact that the top level online events are now 15' + 3" instead of 15' + 10". Also it seems that most people who play against bots don't spend more than 15' per game, usually less (with the bot moving in one second or so).

The new methodology refers more to how we test rather than to how the levels are defined. However we did raise the minimum number of nodes substantially (with more randomness for the low levels to compensate), because the low elo levels in Dragon 3.2 and prior would often just repeat position even when up multiple queens, making test results nonsense. This is now fixed. For testing the levels were run against varying fixed depth levels of a chosen unrelated conventional engine with the relative ratings determined from those results, no direct level vs level testing which inflates the differences (compared to human results), with a few data points vs. humans at higher levels to set the overall level. This should be much more accurate.

Yes, we still have Regular Eval. Although Stockfish has reduced UCI options to a minimum, we have not gone that route.

I have just run a 15000 games match between Dragon 3 and Dragon 3.3, both set to 2000 Elo. Dragon 3.3 won by 32 points. Taking into account that Dragon 3 Elo levels were aimed at 15'+10'', this is very strange. Is it possible that Dragon 3 Elo levles were so much weaker than they should be?

Yes, it is quite possible. The relative strength of the different elo settings should be much more accurate now, so even if say 2600 remains the same, other levels will be different. I would suggest that the chess.com Rapid ratings for humans make for a better comparison with Komodo levels than FIDE ratings for amateur players, because the lower FIDE ratings seem to vary widely depending on the age of the players and the nationality. A 1600 young FIDE player today from a young country will be much stronger than a middle aged 1600 FIDE player from an older country who earned his rating 15 years ago. Online Rapid ratings are a much more consistent basis for comparison, and on average there is little difference between chess.com rapid ratings and FIDE ratings.

At 2600, or any Elo settings, Dragon 3 should be 100 points or so stronger than Dragon 3.3, given the fact that humans play is about 100 Elo weaker at 10'+5'' than at 15'+10''. I'll run a match at 2600 and report the results.

[Krzysztof Grzelak]

Hallo Mark and Larry,

I would like to ask if one of you can write why I do not have the update to version 3.3.

Krzysztof Grzelak

[Fritz 0]

KomodoChess.com has released Dragon 3.3, expected to be the final version of Dragon, primarily to fulfill obligations to subscribers who were still due for another upgrade. It is free for anyone who bought Dragon 3.2 or who received Dragon 3.2 via subscription. New purchasers of Dragon should receive 3.3 from now one.

Although Mark and I retired from running KomodoChess at the start of this year, turning it over to chess.com, we have continued to improve its search on a part-time basis, while Dietriche Kappe was able to upgrade the net to a 50% larger size with increased strength, still trained on Dragon moves. In addition, the Elo levels should be much more accurate now (aimed at 10' + 5" Rapid chess), due to a new methodology and also to correcting an issue that made the low Elo levels weaker than intended. We welcome feedback on the accuracy of the new levels for human opponents.

Dragon 3.3 is a normal size upgrade; with "UHO" openings we got about 12 elo plus (against Dragon 3.2) in bullet and 15 elo plus in our blitz testing, with the increased gains with longer time controls presumably due to the larger net. With normal openings, we got a +8 elo result in 40,000 single-thread games at 1' + 0.6", and have confirmed this with a +8 elo result after 11400 games at 5' + 2". Gains on 4 or 8 threads appear to be a bit better, at least not worse. MCTS and FRC play should also be a bit improved, although these were not specifically targeted since version 3.2. Considering how difficult it is to make further Elo progress from the current lofty levels, we feel that these gains fully justify this final release.

There may be a short technical delay in getting copies to testers. We hope it will be fully tested as the final Dragon version.

Larry, a few questions:

Why did you switch from 15'+10'' to 10'+5'' in the Elo levels?

Can you tell us more about that new methodology? Is it still based on node count and randomness?

Does Dragon 3.3 still has Regular Eval option? (just asking this because Stockfish removed it)

The switch in target time control was due to the fact that most online Rapid play is now at just 10' with no increment, so this is a compromise, and to the fact that the top level online events are now 15' + 3" instead of 15' + 10". Also it seems that most people who play against bots don't spend more than 15' per game, usually less (with the bot moving in one second or so).

The new methodology refers more to how we test rather than to how the levels are defined. However we did raise the minimum number of nodes substantially (with more randomness for the low levels to compensate), because the low elo levels in Dragon 3.2 and prior would often just repeat position even when up multiple queens, making test results nonsense. This is now fixed. For testing the levels were run against varying fixed depth levels of a chosen unrelated conventional engine with the relative ratings determined from those results, no direct level vs level testing which inflates the differences (compared to human results), with a few data points vs. humans at higher levels to set the overall level. This should be much more accurate.

Yes, we still have Regular Eval. Although Stockfish has reduced UCI options to a minimum, we have not gone that route.

I have just run a 15000 games match between Dragon 3 and Dragon 3.3, both set to 2000 Elo. Dragon 3.3 won by 32 points. Taking into account that Dragon 3 Elo levels were aimed at 15'+10'', this is very strange. Is it possible that Dragon 3 Elo levles were so much weaker than they should be?

Yes, it is quite possible. The relative strength of the different elo settings should be much more accurate now, so even if say 2600 remains the same, other levels will be different. I would suggest that the chess.com Rapid ratings for humans make for a better comparison with Komodo levels than FIDE ratings for amateur players, because the lower FIDE ratings seem to vary widely depending on the age of the players and the nationality. A 1600 young FIDE player today from a young country will be much stronger than a middle aged 1600 FIDE player from an older country who earned his rating 15 years ago. Online Rapid ratings are a much more consistent basis for comparison, and on average there is little difference between chess.com rapid ratings and FIDE ratings.

At 2600, or any Elo settings, Dragon 3 should be 100 points or so stronger than Dragon 3.3, given the fact that humans play is about 100 Elo weaker at 10'+5'' than at 15'+10''. I'll run a match at 2600 and report the results.

After 10000 and something games, Dragon 3 2600 leads by 7 Elo points vs. Dragon 3.3 2600. So basically no difference. Which means that Dragon 3 2600 is about 100 points weaker than in should be. Am I missing something, Larry?

[lkaufman]

KomodoChess.com has released Dragon 3.3, expected to be the final version of Dragon, primarily to fulfill obligations to subscribers who were still due for another upgrade. It is free for anyone who bought Dragon 3.2 or who received Dragon 3.2 via subscription. New purchasers of Dragon should receive 3.3 from now one.

Although Mark and I retired from running KomodoChess at the start of this year, turning it over to chess.com, we have continued to improve its search on a part-time basis, while Dietriche Kappe was able to upgrade the net to a 50% larger size with increased strength, still trained on Dragon moves. In addition, the Elo levels should be much more accurate now (aimed at 10' + 5" Rapid chess), due to a new methodology and also to correcting an issue that made the low Elo levels weaker than intended. We welcome feedback on the accuracy of the new levels for human opponents.

Dragon 3.3 is a normal size upgrade; with "UHO" openings we got about 12 elo plus (against Dragon 3.2) in bullet and 15 elo plus in our blitz testing, with the increased gains with longer time controls presumably due to the larger net. With normal openings, we got a +8 elo result in 40,000 single-thread games at 1' + 0.6", and have confirmed this with a +8 elo result after 11400 games at 5' + 2". Gains on 4 or 8 threads appear to be a bit better, at least not worse. MCTS and FRC play should also be a bit improved, although these were not specifically targeted since version 3.2. Considering how difficult it is to make further Elo progress from the current lofty levels, we feel that these gains fully justify this final release.

There may be a short technical delay in getting copies to testers. We hope it will be fully tested as the final Dragon version.

Larry, a few questions:

Why did you switch from 15'+10'' to 10'+5'' in the Elo levels?

Can you tell us more about that new methodology? Is it still based on node count and randomness?

Does Dragon 3.3 still has Regular Eval option? (just asking this because Stockfish removed it)

The switch in target time control was due to the fact that most online Rapid play is now at just 10' with no increment, so this is a compromise, and to the fact that the top level online events are now 15' + 3" instead of 15' + 10". Also it seems that most people who play against bots don't spend more than 15' per game, usually less (with the bot moving in one second or so).

The new methodology refers more to how we test rather than to how the levels are defined. However we did raise the minimum number of nodes substantially (with more randomness for the low levels to compensate), because the low elo levels in Dragon 3.2 and prior would often just repeat position even when up multiple queens, making test results nonsense. This is now fixed. For testing the levels were run against varying fixed depth levels of a chosen unrelated conventional engine with the relative ratings determined from those results, no direct level vs level testing which inflates the differences (compared to human results), with a few data points vs. humans at higher levels to set the overall level. This should be much more accurate.

Yes, we still have Regular Eval. Although Stockfish has reduced UCI options to a minimum, we have not gone that route.

I have just run a 15000 games match between Dragon 3 and Dragon 3.3, both set to 2000 Elo. Dragon 3.3 won by 32 points. Taking into account that Dragon 3 Elo levels were aimed at 15'+10'', this is very strange. Is it possible that Dragon 3 Elo levles were so much weaker than they should be?

Yes, it is quite possible. The relative strength of the different elo settings should be much more accurate now, so even if say 2600 remains the same, other levels will be different. I would suggest that the chess.com Rapid ratings for humans make for a better comparison with Komodo levels than FIDE ratings for amateur players, because the lower FIDE ratings seem to vary widely depending on the age of the players and the nationality. A 1600 young FIDE player today from a young country will be much stronger than a middle aged 1600 FIDE player from an older country who earned his rating 15 years ago. Online Rapid ratings are a much more consistent basis for comparison, and on average there is little difference between chess.com rapid ratings and FIDE ratings.

At 2600, or any Elo settings, Dragon 3 should be 100 points or so stronger than Dragon 3.3, given the fact that humans play is about 100 Elo weaker at 10'+5'' than at 15'+10''. I'll run a match at 2600 and report the results.

After 10000 and something games, Dragon 3 2600 leads by 7 Elo points vs. Dragon 3.3 2600. So basically no difference. Which means that Dragon 3 2600 is about 100 points weaker than in should be. Am I missing something, Larry?

Well, the difference between 15' + 10" and 10' + 5" is probably not 100 elo, maybe about 75 I think in serious play. But the main point is that in test games with humans given 15' + 10", they generally don't use most of their time when playing an engine that moves in one second. So I think we were a bit off in saying that the earlier results were right for the longer time limit, because even though we used it, humans generally played about as if 10 + 5". That seems to be the approximate speed that most players will play at if there is no time limit but also no big financial or rating incentive to get the best possible result. For example when top player/streamers like Nakamura play super strong bots without a time limit they will rarely exceed 10' + 5" inc.

[Fritz 0]

KomodoChess.com has released Dragon 3.3, expected to be the final version of Dragon, primarily to fulfill obligations to subscribers who were still due for another upgrade. It is free for anyone who bought Dragon 3.2 or who received Dragon 3.2 via subscription. New purchasers of Dragon should receive 3.3 from now one.

Although Mark and I retired from running KomodoChess at the start of this year, turning it over to chess.com, we have continued to improve its search on a part-time basis, while Dietriche Kappe was able to upgrade the net to a 50% larger size with increased strength, still trained on Dragon moves. In addition, the Elo levels should be much more accurate now (aimed at 10' + 5" Rapid chess), due to a new methodology and also to correcting an issue that made the low Elo levels weaker than intended. We welcome feedback on the accuracy of the new levels for human opponents.

Dragon 3.3 is a normal size upgrade; with "UHO" openings we got about 12 elo plus (against Dragon 3.2) in bullet and 15 elo plus in our blitz testing, with the increased gains with longer time controls presumably due to the larger net. With normal openings, we got a +8 elo result in 40,000 single-thread games at 1' + 0.6", and have confirmed this with a +8 elo result after 11400 games at 5' + 2". Gains on 4 or 8 threads appear to be a bit better, at least not worse. MCTS and FRC play should also be a bit improved, although these were not specifically targeted since version 3.2. Considering how difficult it is to make further Elo progress from the current lofty levels, we feel that these gains fully justify this final release.

There may be a short technical delay in getting copies to testers. We hope it will be fully tested as the final Dragon version.

Larry, a few questions:

Why did you switch from 15'+10'' to 10'+5'' in the Elo levels?

Can you tell us more about that new methodology? Is it still based on node count and randomness?

Does Dragon 3.3 still has Regular Eval option? (just asking this because Stockfish removed it)

The switch in target time control was due to the fact that most online Rapid play is now at just 10' with no increment, so this is a compromise, and to the fact that the top level online events are now 15' + 3" instead of 15' + 10". Also it seems that most people who play against bots don't spend more than 15' per game, usually less (with the bot moving in one second or so).

The new methodology refers more to how we test rather than to how the levels are defined. However we did raise the minimum number of nodes substantially (with more randomness for the low levels to compensate), because the low elo levels in Dragon 3.2 and prior would often just repeat position even when up multiple queens, making test results nonsense. This is now fixed. For testing the levels were run against varying fixed depth levels of a chosen unrelated conventional engine with the relative ratings determined from those results, no direct level vs level testing which inflates the differences (compared to human results), with a few data points vs. humans at higher levels to set the overall level. This should be much more accurate.

Yes, we still have Regular Eval. Although Stockfish has reduced UCI options to a minimum, we have not gone that route.

I have just run a 15000 games match between Dragon 3 and Dragon 3.3, both set to 2000 Elo. Dragon 3.3 won by 32 points. Taking into account that Dragon 3 Elo levels were aimed at 15'+10'', this is very strange. Is it possible that Dragon 3 Elo levles were so much weaker than they should be?

Yes, it is quite possible. The relative strength of the different elo settings should be much more accurate now, so even if say 2600 remains the same, other levels will be different. I would suggest that the chess.com Rapid ratings for humans make for a better comparison with Komodo levels than FIDE ratings for amateur players, because the lower FIDE ratings seem to vary widely depending on the age of the players and the nationality. A 1600 young FIDE player today from a young country will be much stronger than a middle aged 1600 FIDE player from an older country who earned his rating 15 years ago. Online Rapid ratings are a much more consistent basis for comparison, and on average there is little difference between chess.com rapid ratings and FIDE ratings.

At 2600, or any Elo settings, Dragon 3 should be 100 points or so stronger than Dragon 3.3, given the fact that humans play is about 100 Elo weaker at 10'+5'' than at 15'+10''. I'll run a match at 2600 and report the results.

After 10000 and something games, Dragon 3 2600 leads by 7 Elo points vs. Dragon 3.3 2600. So basically no difference. Which means that Dragon 3 2600 is about 100 points weaker than in should be. Am I missing something, Larry?

Well, the difference between 15' + 10" and 10' + 5" is probably not 100 elo, maybe about 75 I think in serious play. But the main point is that in test games with humans given 15' + 10", they generally don't use most of their time when playing an engine that moves in one second. So I think we were a bit off in saying that the earlier results were right for the longer time limit, because even though we used it, humans generally played about as if 10 + 5". That seems to be the approximate speed that most players will play at if there is no time limit but also no big financial or rating incentive to get the best possible result. For example when top player/streamers like Nakamura play super strong bots without a time limit they will rarely exceed 10' + 5" inc.

Yes, this makes sense. I saw that 100 points difference in a graficon of Ken Regan, but his numbers did seem pretty exaggerated to me. Also, it is true that people don't use all available time when playing against engines and generally (everyone who played online for a while can confirm that), so there is no significant drop in playing strength from 15'+10'' and 10'+5''.

[JohnW]

Hallo Mark and Larry,

I would like to ask if one of you can write why I do not have the update to version 3.3.

Krzysztof Grzelak

I don't have access to it either though I purchased 3.2

[Stonker]

I still have no access to D3.3 either. Has Team Komodo/Dragon disabled their much quoted help email address? As previously stated, I just get instant auto bounce-backs, though I haven't tried recently, as I'm getting a little tired of it all. I'm disinclined to send proof of purchase with order number etc on this public forum. Komodo Team: please read your first post in this stream, re purchasers of D3.2 being entitled to the Dragon 3.3 upgrade. A little unlike them not to respond. FWIW: I have always had efficient, helpful and friendly exchanges with them in the past...

[lkaufman]

Regarding the issue of free upgrades from 3.2 to 3.3, it seems that this was only supposed to be for recent purchasers, but that condition was left off of the info I got for the announcement. So KomodoChess will honor any requests to update from 3.2 to 3.3 until Oct. 15, after that only for purchases made in Sept or Oct 2023. If you have requested an upgrade and haven't yet received it, I suggest you wait until Wednesday and if you still haven't received it send another email to komodochesshelp@gmail.com explaining your situation.

[Werewolf]

Regarding the issue of free upgrades from 3.2 to 3.3, it seems that this was only supposed to be for recent purchasers, but that condition was left off of the info I got for the announcement. So KomodoChess will honor any requests to update from 3.2 to 3.3 until Oct. 15, after that only for purchases made in Sept or Oct 2023. If you have requested an upgrade and haven't yet received it, I suggest you wait until Wednesday and if you still haven't received it send another email to komodochesshelp@gmail.com explaining your situation.

End of an era I guess.

Good luck with Torch.

[Stonker]

Thank you Komodo team: that sounds as reasonable as possible.
Unfortunately I am still not able to put in my upgrade request to your email address as my messages still bounce right back. I have no problem with my
emailing to anyone else. My Dragon 3.2 order number was #14316: hopefully that's enough information for you to activate the download. If not then please let me know.
Cheers,
M.W.