Prune 3.1.0 - new version chess engine

-

Prune - UCI chess engine
Author: Thomas Girolami

Prune 3.1.0 - what's new?

this new version have:
an HCE build, as it can be funny to do some trick on HCE, however:

will be less important than nnue version (we will be more focused on the nnue version), but HCE only patch are accepted
is a lot weaker than the nnue, most likely between 2700-2800
a new nnue, with 128 HL and 8 outputs buckets, now trained using bullet
some search patches as:
removing delta pruning (+50 in non-regr test)
reduce more the depth for NMP
history score based reduction for LMR
double and negative extension
some new features, as:
Late Move Pruning
some kind of history pruning
some speedup, as:
prefetching TT entry
faster index calculation for TT (was also surely better, as it is uniform now)
STC progress:

Results of prune v3.1.0 vs prune v3 (10+0.1, 1t, 16MB, UHO_Lichess_4852_v1.epd):
Elo: 269.85 +/- 9.25, nElo: 408.67 +/- 9.63
LOS: 100.00 %, DrawRatio: 14.44 %, PairsRatio: 48.74
Games: 5000, Wins: 3575, Losses: 321, Draws: 1104, Points: 4127.0 (82.54 %)
Ptnml(0-2): [4, 39, 361, 891, 1205], WL/DD Ratio: 3.15
LTC progress:

Results of prune v3.1.0 vs prune v3 (60+0.6, 1t, 128MB, UHO_Lichess_4852_v1.epd):
Elo: 238.84 +/- 10.31, nElo: 388.03 +/- 12.43
LOS: 100.00 %, DrawRatio: 15.13 %, PairsRatio: 46.15
Games: 3000, Wins: 1965, Losses: 176, Draws: 859, Points: 2394.5 (79.82 %)
Ptnml(0-2): [1, 26, 227, 675, 571], WL/DD Ratio: 1.87
against stash 36:

Results of prune v3.1.0 vs stash 36 (10+0.1, 1t, 16MB, UHO_Lichess_4852_v1.epd):
Elo: 3.96 +/- 7.50, nElo: 5.09 +/- 9.63
LOS: 84.98 %, DrawRatio: 38.40 %, PairsRatio: 1.10
Games: 5000, Wins: 1797, Losses: 1740, Draws: 1463, Points: 2528.5 (50.57 %)
Ptnml(0-2): [256, 478, 960, 565, 241], WL/DD Ratio: 3.57
Results of dev vs base (60+0.6, 1t, 128MB, UHO_Lichess_4852_v1.epd):
Elo: -6.95 +/- 10.84, nElo: -9.77 +/- 15.23
LOS: 10.44 %, DrawRatio: 42.10 %, PairsRatio: 0.94
Games: 2000, Wins: 654, Losses: 694, Draws: 652, Points: 980.0 (49.00 %)
Ptnml(0-2): [84, 214, 421, 220, 61], WL/DD Ratio: 2.86
so a small antiscaler detected against stash 36

test of pruneHCE, against stash 23:

Results of pruneHCE 3.1.0 vs stash 23 (10+0.1, 1t, 16MB, UHO_Lichess_4852_v1.epd):
Elo: -3.82 +/- 7.85, nElo: -4.69 +/- 9.63
LOS: 17.00 %, DrawRatio: 37.80 %, PairsRatio: 0.99
Games: 5000, Wins: 1845, Losses: 1900, Draws: 1255, Points: 2472.5 (49.45 %)
Ptnml(0-2): [317, 466, 945, 499, 273], WL/DD Ratio: 5.52
Results of prune HCE 3.1.0 vs stash 23 (60+0.6, 1t, 128MB, UHO_Lichess_4852_v1.epd):
Elo: 27.68 +/- 12.22, nElo: 34.64 +/- 15.23
LOS: 100.00 %, DrawRatio: 36.60 %, PairsRatio: 1.47
Games: 2000, Wins: 801, Losses: 642, Draws: 557, Points: 1079.5 (53.97 %)
Ptnml(0-2): [91, 166, 366, 247, 130], WL/DD Ratio: 4.08


Prune 2.0.1 vs other engines:

Eubos 4.2 JA2/2+22 Games
Colossus 2025b1/2+02 Games
Bread Engine 2.0.00.5/2-12 Games
Qapla 0.3.20.5/2-12 Games
Reyna 1.00.5/2-12 Games
Tcheran v6.00.5/2-12 Games
c4ke v1.00.5/2-12 Games
PZChessBot 5.00/2-22 Games
Publius 1.0 slow0/2-22 Games
Arcanum 2.70/2-22 Games
Bitbit 1.60/2-22 Games
Minke 3.0.00/2-22 Games
Cwtch 30/2-22 Games
Critter 1.6a 64-bit0/2-22 Games
 



Comments

Post a Comment

Popular posts from this blog

We already have the new Fritz 20 program - what does it look like and how does it work?

-
Image
The purchase process for the download version on the ChessBase website took only 5 minutes. Downloading the files from the cloud was even quicker. The installation itself, which involved five large files, took maybe 7 minutes, including restarting the system and entering the serial number I received upon purchase. The start image for Fritz 20 has a strange mosaic of colors; perhaps I'll get used to it. As you can see in the second screenshot, the program interface hasn't changed much; only the icons are different (we've been using Fritz 17 until now). It looks nice, but it's no major revolution. The next two screenshots show engine management and the installation of the UCI chess engine (here, Cataphract 1.1). As you can see, there are no changes. Screenshot no. 5 shows the analysis of some position by Fritz 20 engines. The changes are minimal - it is still very legible. Finally, on screen 6 we have the engine tournament table generated by Fritz 20 . I think it is sti...
Read more

Reckless 0.9.0 - new version super strong chess engine

-
Image
  Reckless , a Rust-based UCI chess engine. Author: Arseniy Surkov    Rating Chess Engines Diary CEDR=3743 Reckless is the new star among chess engines. On the new Chess Engines Diary CEDR ranking list, Reckless is ranked 2th with a rating of 3743 . Testing of the new engine version is coming soon. Reckless 0.9.0 what's new? Reckless 0.9.0 is ready! You can download pre-built binaries for Windows, Linux, and macOS from the GitHub Releases page. Thanks to everyone who contributed through code, testing, hardware support, or any other effort. This release came together because of your help. Update highlights NNUE The neuron network has been updated to a modern threat-based architecture. It models the interactions between pieces, where one piece attacks or defends another, and extends the existing king-bucketed piece-square-based inputs with additional features. Multithreading and NUMA support NUMA (Non-Uniform Memory Access) support has been added to Reckless. Work...
Read more