Whiskers wrote: ↑Sun Mar 31, 2024 9:43 pm
When Patricia 2.0 releases (should be within a week or two) it should be the most aggressive engine ever, however it's much weaker than the other engines listed here (at only 3000-3200 depending on how many search patches I can get to work without the aggression level tanking).
Patricia was my grandmother's name, so I'm struggling to take this seriously as she was barely over 5 feet tall and rarely ill-tempered, let alone aggressive.
But seriously, what about it makes it so aggressive? Do say when it's released.
What is aggressive chess? Counting sacrifices is just one metric, somewhat meaningless in my opinion. Here is some data on sacrifices and exchange sacrifices that I extracted from the 40H database of Norman Pollock. (Disclaimer: The data may not be perfect, and the methods could be improved. For example, there may be a small fraction of games in which a player shown in the table may be at the receiving end of a sacrifice, while I have done only 'grep' for their name in a databases of games with sacrifices or exchange sacrifices. Also, I just started experimenting with the filters provided by pgn-extract and some of the filters written by Stefan Pohl. So the numbers could be plus/minus some delta.) You may get some surprises.
Number of sacrifices/100 games (in decreasing order)
It is nice to see one list ending with Fischer and the other ending with Carlsen.
Caruana ahead of Tal, Kasparov, Anand in the first list. Tal and Kasparov somewhat lower in both lists. On the other hand, Topalov, Polgar, Shirov seem to confirm the image.
Werewolf wrote: ↑Mon Apr 01, 2024 6:24 pm
So just so I'm clear here, out of all the engines you've tested, Velvet 4.1.0 is the most aggressive of them all?
Hard to say. Velvet 4.1.0, Komodo 14.1 aggressive, Revenge 1.0, Pedone 3.0 (not 3.1!) are all playing very aggressive. From these engines, Revenge 1.0 is clearly the strongest. Velvet 4.1.0, Komodo 14.1 aggressive, Pedone 3.0 (not 3.1!) are on the same Elo-level, but clearly weaker than Revenge 1.0.
If you look at high numbers of sacrifices, Uralochka 3.37c and Arasan 23.2 are interesting, too...
Dont forget about Stockfish... Stockfish 16.1 is playing aggressive...
chesskobra wrote: ↑Tue Apr 02, 2024 3:45 am
What is aggressive chess? Counting sacrifices is just one metric, somewhat meaningless in my opinion.
Yes I agree. This may be the reason, perhaps, I'm getting very different results to Mr. Pohl.
I'm defining it as a propensity to attack the enemy king.
Where are these different results? What was measured? And how?
(And by the way: My EAS-Tool does much, much more, than just counting sacs...)
OK noted about your EAS tool. I'm not criticising it, but I do get different results.
It'll take me another month to build up the data but I'm testing engines with around 1000 Elo difference in strength, where resignation is not allowed, based on this logic:
1) Where massive Elo difference exists the stronger side can usually exert its will without much resistance
2) Checkmate terminates a game immediately
3) Therefore, an engine with a tendency to attack the enemy king by flinging pieces at it will likely have shorter games on average than a grinder.
@Stefan, I came across this game labelled as queen sacrifice, 24.Re1 Qxe2 25.Bxf6 but it really isn't, see 29.Nxe2
The reason is, the EAS Tool looks for less material, a missing queen, but later the win for this color. But the 2 conditions (less material, a missing queen) must be true for 8 plies in a row. In this rare case here, the queen is recaptured after 9 plies...
In most cases, the 8-ply interval works fine. If I would increase it to 10 or 12, a lot of sacs would be missed by the EAS Tool.
So 8 plies work best, but there is no perfect world... If you parse games only by their moves with pgn-extract, without any evaluation by an engine (that makes the EAS Tool so brutally fast), there can be mismatches finding sacs. I am not able to avoid this completely, sorry.
chesskobra wrote: ↑Tue Apr 02, 2024 3:45 am
What is aggressive chess? Counting sacrifices is just one metric, somewhat meaningless in my opinion.
Yes I agree. This may be the reason, perhaps, I'm getting very different results to Mr. Pohl.
I'm defining it as a propensity to attack the enemy king.
Where are these different results? What was measured? And how?
(And by the way: My EAS-Tool does much, much more, than just counting sacs...)
OK noted about your EAS tool. I'm not criticising it, but I do get different results.
It'll take me another month to build up the data but I'm testing engines with around 1000 Elo difference in strength, where resignation is not allowed, based on this logic:
1) Where massive Elo difference exists the stronger side can usually exert its will without much resistance
2) Checkmate terminates a game immediately
3) Therefore, an engine with a tendency to attack the enemy king by flinging pieces at it will likely have shorter games on average than a grinder.
@Stefan, I came across this game labelled as queen sacrifice, 24.Re1 Qxe2 25.Bxf6 but it really isn't, see 29.Nxe2
The reason is, the EAS Tool looks for less material, a missing queen, but later the win for this color. But the 2 conditions (less material, a missing queen) must be true for 8 plies in a row. In this rare case here, the queen is recaptured after 9 plies...
In most cases, the 8-ply interval works fine. If I would increase it to 10 or 12, a lot of sacs would be missed by the EAS Tool.
So 8 plies work best, but there is no perfect world... If you parse games only by their moves with pgn-extract, without any evaluation by an engine (that makes the EAS Tool so brutally fast), there can be mismatches finding sacs. I am not able to avoid this completely, sorry.
Yep.
If have to see the perfection definition is a sac, anyone?
90% of coding is debugging, the other 10% is writing bugs.
