Bug Fix Release of Maverick (Version 0.51)

Maverick 0.5 has been in the wild for about a month. I’m pleased to say it seems to be playing at 2300 ELO or above.

However, occasionally it would crash. This really niggled me.

I do quite a lot of turbo-testing using CuteChess-Cli at fast time controls.  So I had convinced myself the occasional crash was probably an overload of the standard input and output pipes (I’m not even sure if this is possible). However, since I’ve tested Maverick quite thoroughly using perft positions, I was confident the basic chess engine routines were bug free. I was wrong.

This week I have been down in São Paulo Brazil on business. Graham Banks has been testing Maverick and broadcasting the games.  Since I had nothing better to do I logged in and watched a few games.  It’s amazing how nerve wracking it is to watch your engine play a game at 40 moves in 40 minutes. As I was watching, and to my surprise, Maverick suddenly crashed. I quickly copied the position and thankfully was able to reproduce the crash!

The bug was quite bizarre. Take a look at the position below:

discovered-check-bug

After black has played pawn to g4, white is in check from the queen on h6. However, the code which finds the source of the check is shown below:

discovered-check-bug-codeThis was a “copy and paste” error. The code thinks the knight on c2 is actually a king attacking the white king. This was the bug.  It was easy to fix.  Hopefully this will result in no more crashed.

You can downloaded the fixed version (Maverick 0.51) on the Download Page. There are really no strength improvements for this bug-fix release.  If you are testing Maverick you can keep the games already played.

Is Maverick a Clone?

Let me state unequivocally, Maverick is not a clone!

There have always been problems with clones and computer chess. The first one I recall was Quickstep.  This was a program competing in the 1989 World Computer Chess Championship in Portorož.  After a couple of rounds it was discovered to be a clone of Richard Lang’s Mephisto Almeria program. It was scandalous at the time.  Thankfully there was little cloning activity after this for about ten years. Then Crafty had a few clones (e.g. Bionic Impakt).  But when Fruit burst onto the scene it became a big problem. This then lead to the Rybka controversy (which I’m not doing to get into).

The computer chess community has developed a test to quickly identify clones.  This work was pioneered by Adam Hair.  You can read about it here – A Pairwise Comparison of Chess Engine Move Selections. I must say I was a little skeptical about the test. I wondered if a basic engine with piece-square-tables, mobility and passed-pawn code would choose similar moves even if they weren’t at all related.  Adam was kind enough to put Maverick through his test – here are the results:

------ Maverick 0.5 x64 (No Popcount) (time: 100 ms scale: 1.0) ------
45.50 Fruit 2.1 (time: 100 ms scale: 1.0)
45.24 Movei00_8_438 (time: 100 ms scale: 1.0)
44.74 SmarThink 1.20 (time: 100 ms scale: 1.0)
44.72 Strelka 2.0 B x64 (time: 100 ms scale: 1.0)
44.50 Fruit 2.2.1 (time: 100 ms scale: 1.0)
44.32 Doch64 09.980 JA (time: 100 ms scale: 1.0)
44.19 Loop 2007 32-Bit (time: 100 ms scale: 1.0)
44.17 Strelka 1.8 UCI (time: 100 ms scale: 1.0)
43.99 Toga II 1.0 (time: 100 ms scale: 1.0)
43.72 Nemo SP64n 1.0.1 Beta (time: 100 ms scale: 1.0)
43.38 Rybka 1.0 Beta 32-bit (time: 100 ms scale: 1.0)
43.26 Daydreamer 1.75 JA (time: 100 ms scale: 1.0)
43.17 Twisted Logic 20090922 (time: 100 ms scale: 1.0)
42.94 Nebula 2.0 (time: 100 ms scale: 1.0)
42.91 RedQueen 1.1.4 (time: 100 ms scale: 1.0)
42.66 Cyrano 0.6b17 JA (time: 100 ms scale: 1.0)
42.61 Hamsters 0.7.1 (time: 100 ms scale: 1.0)
42.60 Naum 4.2 (time: 100 ms scale: 1.0)
42.33 Murka 3 x64 UCI (time: 100 ms scale: 1.0)
42.32 spark-1.0 (time: 100 ms scale: 1.0)
42.26 Octochess revision 5190 (time: 100 ms scale: 1.0)
41.81 DiscoCheck 4.3 JA (time: 100 ms scale: 1.0)
41.77 Bobcat 3.25 (time: 100 ms scale: 1.0)
41.67 cheng3 1.07 (time: 100 ms scale: 1.0)
41.58 Gandalf 6.0 (time: 100 ms scale: 1.0)
41.37 Glass 2.0 PERSONALITY (time: 100 ms scale: 1.0)
41.28 RobboLite 0.085d3 x64 (time: 100 ms scale: 1.0)
41.22 Spike 1.4 (time: 100 ms scale: 1.0)
41.04 Houdini x64 1_CPU (time: 100 ms scale: 1.0)
40.93 Houdini 3 x64 (time: 100 ms scale: 1.0)
40.79 Gaviota v0.86 (time: 100 ms scale: 1.0)
40.75 Ruffian 2.1.0 (time: 100 ms scale: 1.0)
40.71 Stockfish 4 64 (time: 100 ms scale: 1.0)
40.70 Godel 2.3.7 (time: 100 ms scale: 1.0)
40.62 Pawny 1.0.x64 (time: 100 ms scale: 1.0)
40.59 Shredder 11 UCI (time: 100 ms scale: 1.0)
40.37 MinkoChess 1.3 x64 (time: 100 ms scale: 1.0)
40.36 Gaviota v0.87-a8 (time: 100 ms scale: 1.0)
40.24 Komodo CCT 64-bit (time: 100 ms scale: 1.0)
40.19 iCE 1.0 v1619 x64 (time: 100 ms scale: 1.0)
40.13 Arasan 16.1 (time: 100 ms scale: 1.0)
40.03 Alfil 13.1 x64 MT (time: 100 ms scale: 1.0)
40.01 GNU Chess 5.50-64 (time: 100 ms scale: 1.0)
39.92 Hannibal 1.3x64 (time: 100 ms scale: 1.0)
39.84 Tornado 4.88 x64 (time: 100 ms scale: 1.0)
39.78 Komodo 5.1 64-bit (time: 100 ms scale: 1.0)
39.66 Atlas 3.50 x64 (time: 100 ms scale: 1.0)
39.16 SlowChess 2.96 (time: 100 ms scale: 1.0)
38.84 Gull 2.2 x64 (time: 100 ms scale: 1.0)
38.54 Quazar 0.4 x64 (time: 100 ms scale: 1.0)
35.63 Texel 1.02 64-bit (time: 100 ms scale: 1.0)
34.33 Deep Sjeng WC2008 x64 (time: 100 ms scale: 1.0)

A score of >55 starts to get suspicious, while greater than 60 is highly likely to be a derivative work. As you can see all of Maverick’s scores are in the 40s! So it passed the test.

Some people seemed to think it was odd I asked about Maverick’s similarity to other engines. It was just simple curiosity. I know Maverick isn’t a clone. Having gone through the similarity test I now have more confidence in the similarity test.

Thanks again Adam for taking the time to test Maverick!

Maverick 0.5 Released!

You can download a new version of Maverick on the download page.

I’ve made quite a few changes since Maverick 0.5. 

  • Added more basic endgame knowledge
  • Added passed pawn evaluation (I think this added a lot of strength)
  • Beefed up the evaluation routine to include basic terms such as mobility
  • Fixed a nasty hash bug to do with mating positions
  • Added one-reply to check extensions in the first ply of the quiescent search (help with some tactics – little impact on playing strength)
  • Enabled hash cutoffs at PV nodes
  • Nodes per second have dropped considerably due to the slower evaluation (now between 3 million and 5 million on my 2.8 GHz i7 notebook)
  • A bazillion tweaks

How Strong is Maverick 0.5?

This is the first version of Maverick which I regards as reasonably strong.  In my tests it’s at least 200 ELO stronger than version 0.2.  If I had to guess I would say it’s about 2350 ELO on the CCRL rating scale.

Here are some test results:

[table align=”center” width=”800″ colwidth=”350|50|50|50″ colalign=”left|right|right|right”]

Match, Maverick, Opponent, Percentage(%)

Maverick 0.5 – Fruit 1.0,50.5,49.5,50.50%

Maverick 0.5 – Dragon 4.6,62.5,37.5,62.50%

Maverick 0.5 – Phalanx XXIII,57.5,42.5,57.50%

Maverick 0.5 – Tao 5.6,38.0,62.0,38.00%

Maverick 0.5 – TJchess 1.1U-x64,57.0,43.0,57.00%

Maverick 0.5 – Maverick 0.2 x64,75.5,24.5,75.50%

[/table]

All the games were played at 1 min per game on a 2.8 GHz Haswell i7 notebook.

Maverick only lost to Tao 5.6, a relatively old (2004) engine, but one I really like. Tao’s rating is about 2500 ELO so this results is expected.  I was surprised Maverick did so well against Dragon and Phalanx. I suspect their CCRL ratings of over 2350 ELO may be inflated.

Mate in 21:

I’ve used this simple position for many years to test hash tables. I first came across it in an advert for the Fidelity Mach III back in (circa.) 1988.  If a hash table is working as it should then an engine will have no problem finding c7, followed by an opposition king maneuver and the queening of the white pawn on b8.

Mate in 21

FEN: 2k5/8/1pP1K3/1P6/8/8/8/8 w – – 0 1

The previous version of Maverick could find c7 but when I added the hash cutoffs at PV nodes Maverick was able to find the mate in 21 in less than 10 seconds – I was pleased.

Next Developments:

I’m planning to release the next version when it can comfortably beat Tao 5.6, YACE and maybe Hermann 2.8. These engines are above 2500 ELO so it should take me a while.  My aim was to have a Maverick playing at 2500 ELO by Christmas – I think that’s still possible.  In the current version there is no king safety or selectivity apart from null move; so I think another 150 ELO by Christmas is possible.

Please met me know if you have any test results for Maverick or if you find a bug!


Fatal error: Uncaught Error: Call to a member function id() on array in /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_formatter.class.php:36 Stack trace: #0 /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_formatter.class.php(538): CrayonFormatter::format_code('', Array, Object(CrayonHighlighter)) #1 [internal function]: CrayonFormatter::delim_to_internal(Array) #2 /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_formatter.class.php(516): preg_replace_callback('#()#msi', 'CrayonFormatter...', '//-- Connected ...') #3 /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_highlighter.class.php(166): CrayonFormatter::format_mixed_code('//-- Connected ...', Object(CrayonLang), Object(CrayonHighlighter)) #4 /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_hig in /home/customer/www/chessprogramming.net/public_html/wp-content/plugins/crayon-syntax-highlighter/crayon_formatter.class.php on line 36