Queen and pawn endgame

[Vinvin]

Did not check if b4 is the only winning move

[d]8/pp3p1k/6pp/7P/6P1/3K4/PPP5/8 w - - 0 38 bm b4

Seems so too, Uri.

Oops:

At least 39.a4 works too.

...

Thanks for this interesting position. SF is very bad after a4.

For 1 billion nodes, SF think that black is better but black is losing (because of f5 ? a5!!), it takes 1.5 billion nodes to see black is losing.

[d]8/pp3p1k/6pp/7P/P5P1/3K4/1PP5/8 b - a3 0 35

Code: Select all

Stockfish_x64_bmi2-2017-07-15:
Found 145 tablebases
...
 20/26    00:00     1.073k    4.048k    -0,05    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Rf4 Rf6 39.a5 h4 40.Rg4 h3 41.Rxh3 Rxf5 42.Rh4 Rf4 43.Rh5 Re3 44.Rxh6 Rd2 45.c4 Rd3 46.c5
 21/26    00:00     1.433k    4.059k    -0,06    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Rf4 Rf6 39.a5 h4 40.Rg4 h3 41.Rxh3 Rxf5 42.Rh4 Rf4 43.Rh5 Re3 44.Rxh6 Rd2 45.c4 Rd3 46.c5 a6 47.Rg5 Rc4 48.Rf5
 22/34    00:00     3.996k    4.238k    -0,70    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Re5 h4 39.Rf4 h5 40.Rf3 Rf6 41.Rg2 Rxf5 42.Rh3 Re4 43.Rxh4 Re3 44.a5 a6 45.Rxh5 Rd2 46.c4 Rd3 47.c5 Rc4 48.Rg4 Rxc5 49.Rg3
 23/35    00:02     9.552k    4.310k    -0,94    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Rf3 Rf6 39.Rf4 a5 40.Re4 h4 41.Rf4 h5 42.Rf3 Rxf5 43.Rg2 Rg4 44.Rh2 h3 45.c3 Rf4 46.Rxh3 Re3 47.b4 Rd3
 24/34-    00:03     13.136k    4.364k    -1,16    35. ... f5
 24/37    00:03     14.824k    4.385k    -1,27    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Re5 h4 39.Rf4 h5 40.b4 Rf6 41.b5 b6 42.c4 h3 43.Rg3 Rxf5 44.c5 bxc5 45.a5 Re5 46.b6 axb6 47.axb6 Rd6 48.b7 Rc7 49.b8D+ Rxb8 50.Rxh3
 25/33-    00:03     15.417k    4.384k    -1,35    35. ... f5
 25/35-    00:03     15.664k    4.392k    -1,42    35. ... f5
 25/35-    00:03     15.987k    4.401k    -1,53    35. ... f5
 25/37    00:03     17.091k    4.406k    -1,52    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.Rg3 Rxf5 42.Rh4 Rg6 43.Rg3 Rg5 44.Rh3 a4 45.Rg3 h4+ 46.Rh3 h5 47.Rg2 Rg4 48.c4 a6 49.c5
 26/33-    00:04     18.072k    4.407k    -1,59    35. ... f5
 26/37-    00:04     18.844k    4.429k    -1,66    35. ... f5
 26/37+    00:04     20.070k    4.437k    -1,59    35. ... f5 36.gxf5
 26/37+    00:04     20.725k    4.428k    -1,23    35. ... f5 36.gxf5
 26/37-    00:04     21.442k    4.427k    -1,39    35. ... f5
 26/37-    00:05     23.400k    4.446k    -1,77    35. ... f5
 26/37    00:05     25.245k    4.463k    -1,79    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.Re5 h4 39.Rf4 Rf6 40.Rg4 h3 41.Rxh3 Rxf5 42.Rg3 Re4 43.Rf2 a5 44.Re2 h5 45.Rf2 Rf4 46.Re2 Rg3 47.Rf1 Rf3 48.Rg1 Re3 49.c3 Rd2 50.Rf1 Rc2 51.Re2 Rb3
 27/35    00:06     28.263k    4.472k    -1,82    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.Rg3 Rxf5 42.Rh4 Rg6 43.Rg3 Rg5 44.Rh3 a4 45.Rg3 h4+ 46.Rh3 h5 47.Rg2 Rg4 48.Rh2 h3 49.c4 Rf4 50.Rxh3 Re5
 28/37-    00:06     30.397k    4.479k    -1,89    35. ... f5
 28/41+    00:06     31.178k    4.458k    -1,82    35. ... f5 36.gxf5
 28/41    00:07     31.583k    4.449k    -1,82    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Re3 De1+ 47.Rf3 Dd1+ 48.Re3 Dd6 49.Re2 De5+ 50.Rd1 Dd5+ 51.Rc1 Rg5 52.Df8 h5 53.De7+ Rxf5
 29/38    00:07     34.873k    4.427k    -1,81    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Re3 De1+ 47.Rf3 Dd1+ 48.Re3 Dd6 49.Re2 De5+ 50.Rd1 Dd5+ 51.Rc1 Rg5 52.Dc7 Rxf5 53.Dxa7 Dc4+ 54.Rb1
 30/56-    00:08     38.229k    4.407k    -1,88    35. ... f5
 30/61+    00:09     39.633k    4.353k    -1,81    35. ... f5 36.gxf5
 30/61    00:09     40.047k    4.345k    -1,81    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De2+ 49.Rh1 Df1+ 50.Rh2 Df2+ 51.Rh1 Df3+ 52.Rg1 Dxf5 53.Dd8+ Rg6 54.Dd6+ Rh5 55.Dh2+ Rg4 56.Dxh6 Db1+ 57.Rf2 Dxb2+ 58.Rf1 Db5+ 59.Re1
 31/45    00:10     45.776k    4.275k    -1,81    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De2+ 49.Rh1 Df1+ 50.Rh2 Df2+ 51.Rh1 Df3+ 52.Rg1 Dxf5 53.Dd8+ Rg6 54.Dd6+ Rh5 55.Dh2+ Rg4 56.Dxh6 Db1+ 57.Rf2
 32/55+    00:11     48.133k    4.251k    -1,74    35. ... f5 36.gxf5
 32/55+    00:11     49.157k    4.235k    -1,66    35. ... f5 36.gxf5
 32/65-    00:12     51.317k    4.222k    -1,74    35. ... f5
 32/65    00:12     54.685k    4.227k    -1,81    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Dh2+ 46.Rf3 Dh5+ 47.Rg3 Dg5+ 48.Rh2 Dh4+ 49.Rg2 De4+ 50.Rh2 De5+ 51.Rg1 Dd4+ 52.Rf1 Dd3+ 53.Rg2 Dd5+ 54.Rh2 h5 55.Da6+ Rxf5 56.Dc8+ De6 57.Dc2+ De4 58.Dc8+ Rf4 59.Dc1+ De3
 33/66+    00:13     58.243k    4.193k    -1,74    35. ... f5 36.gxf5
 33/66+    00:14     59.207k    4.189k    -1,66    35. ... f5 36.gxf5
 33/66    00:14     61.472k    4.181k    -1,66    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Dh2+ 46.Rf3 Dh5+ 47.Rg3 Dg5+ 48.Rh2 Dh4+ 49.Rg2 Dg4+ 50.Rh2 Df4+ 51.Rg2 De4+ 52.Rh2 De5+ 53.Rg1 Dxf5 54.Dd8+ Rg6 55.Dd6+ Rh5 56.Dh2+ Rg4 57.Dg2+ Rh4 58.Dh2+ Dh3 59.Dd2 Dg4+ 60.Rh2 Dg3+ 61.Rh1 h5 62.Dd8+ Rh3 63.Dc8+ Dg4
 34/53    00:16     68.633k    4.200k    -1,66    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Dh2+ 46.Rf3 Dh5+ 47.Rg3 Dg5+ 48.Rh2 Dh4+ 49.Rg2 Dg4+ 50.Rh1 De4+ 51.Rh2 De5+ 52.Rg1 Dxf5 53.Dd8+ Rg6 54.Dd6+ Rh5 55.Dh2+ Rg4 56.Dg2+ Rh4 57.Dh2+ Dh3 58.Dd2 Dg4+ 59.Rh2 Dg3+ 60.Rh1 Df3+ 61.Rh2 h5
 35/66+    00:18     76.318k    4.169k    -1,59    35. ... f5 36.gxf5
 35/66+    00:19     80.122k    4.148k    -1,52    35. ... f5 36.gxf5
 35/66+    00:20     85.303k    4.116k    -1,41    35. ... f5 36.gxf5
 35/81-    00:22     92.054k    4.050k    -1,50    35. ... f5
 35/81    00:24     100.743k    4.031k    -1,50    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De5+ 49.Rg1 Dg3+ 50.Rh1 Dh3+ 51.Rg1 Dg4+ 52.Rh1 Dxf5 53.Dd8+ Rg6 54.Dd6+ Rg5 55.Dg3+ Rh5 56.Dh2+ Rg4 57.Dg1+ Rh4 58.Dd4+ Rg5 59.Dg1+ Rf6 60.Dd4+ Re6 61.Dxa7 Db1+ 62.Rg2
 36/70    00:33     130.317k    3.873k    -1,50    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De5+ 49.Rg1 Dg3+ 50.Rh1 Dh3+ 51.Rg1 Dg4+ 52.Rh2 Dh5+ 53.Rg3 Dg5+ 54.Rh2 Df4+ 55.Rg2 Dxf5 56.Dd8+ Rg6 57.De8+ Rg5 58.De3+ Df4 59.De7+ Df6 60.De3+ Rh5 61.De2+ Rg6 62.Dd3+ Rg7 63.Dd7+ Rh8 64.De8+ Rh7 65.Rh3 Df3+ 66.Rh2 Df2+ 67.Rh3 Df1+ 68.Rg3
 37/78-    00:37     145.510k    3.871k    -1,57    35. ... f5
 37/78    00:39     152.803k    3.873k    -1,57    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De5+ 49.Rg1 Dg3+ 50.Rh1 Dh3+ 51.Rg1 Dxf5 52.Dd8+ Rf7 53.Dc7+ Rg6 54.Dd6+ Df6 55.Dd3+ Rf7 56.Dd7+ Rg8 57.Rh1 Dxb2 58.Dd8+ Rf7 59.Dd7+ Rg6 60.De8+ Rf5 61.Dh5+ Re4 62.Dg6+ Rd5 63.Df5+ De5
 38/80-    00:47     181.419k    3.823k    -1,64    35. ... f5
 38/80+    00:49     188.954k    3.829k    -1,57    35. ... f5 36.gxf5
 38/80-    00:49     190.414k    3.836k    -1,64    35. ... f5
 38/80+    00:50     193.857k    3.839k    -1,55    35. ... f5 36.gxf5
 38/80    00:52     201.395k    3.843k    -1,51    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De5+ 49.Rg1 Dxf5 50.Dd8+ Rg6 51.Dd6+ Df6 52.Dd3+ Rf7 53.Dd7+ Rg8 54.De8+ Rh7 55.Rh1 Df2 56.Dd7+ Rg6 57.Dd6+ Rg5 58.De5+ Rg4 59.Dg7+ Rf3 60.Df8+ Re2 61.Dxh6 Dc5 62.Rg2 Rd3 63.Dh7+ Rd2 64.Rf3 Rc1 65.Dg7 Rc2
 39/63-    00:55     212.678k    3.863k    -1,58    35. ... f5
 39/68    00:57     221.990k    3.864k    -1,58    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rg2 De4+ 48.Rh2 De5+ 49.Rg1 Dxf5 50.Dd8+ Rg6 51.Dd6+ Df6 52.Dd3+ Rg7 53.Dd7+ Rg8 54.Rh1 Df2 55.Dg4+ Rf7 56.Dh5+ Rg7 57.De5+ Rg6 58.Dd6+ Rg5 59.De5+ Rg4 60.Dg7+ Rf3 61.Df8+ Re2 62.Dxh6 Dc5 63.Rg2 Rd1 64.Dg7 Rc2 65.Rf3 a4 66.Db7 Df5+ 67.Rg3 Dd3+ 68.Rh4 Dc4+ 69.Rg5
 40/91-    01:09     266.189k    3.825k    -1,66    35. ... f5
 40/91-    01:12     280.904k    3.849k    -1,73    35. ... f5
 40/91-    01:16     297.382k    3.876k    -1,84    35. ... f5
 40/91    01:22     321.000k    3.876k    -1,84    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rh2 Dd2+ 48.Rh1 Dd5+ 49.Rh2 De5+ 50.Rh1 Dxf5 51.Dd8+ Rg6 52.Dd6+ Rg5 53.Dg3+ Rh5 54.Dh2+ Rg4 55.Dg1+ Rh4 56.De1+ Rh5 57.Dd1+ Dg4 58.Dd5+ Dg5 59.Df3+ Rg6 60.Dd3+ Rf7 61.Dc4+ Re7 62.Dc7+ Rf6 63.Dc3+ Re6 64.Dc8+ Rd5 65.Dc7 Dh4+ 66.Rg2
 41/80-    01:49     414.228k    3.781k    -1,91    35. ... f5
 41/87    01:59     451.110k    3.770k    -1,91    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dd3+ 47.Rh2 Dd2+ 48.Rh3 De3+ 49.Rg2 De4+ 50.Rh2 De5+ 51.Rh1 Dxf5 52.Dd8+ Rg6 53.Dd6+ Rg5 54.Dg3+ Rh5 55.Dh2+ Rg4 56.Dg1+ Rh4 57.Dd4+ Rg5 58.Dc3 Df1+ 59.Rh2 Df4+ 60.Rh3 Df5+ 61.Rg2 Rh4 62.b3 Dg5+ 63.Rf2 Rh5 64.Re2 Dg4+ 65.Rf2 Df4+ 66.Rg2 Rg4 67.Dxa5 Df3+ 68.Rg1 Dg3+ 69.Rh1 Dh4+ 70.Rg2
 42/78+    02:10     489.246k    3.744k    -1,84    35. ... f5 36.gxf5
 42/78    02:13     501.450k    3.750k    -1,84    35. ... f5 36.gxf5 gxh5 37.Re4 Rg7 38.a5 Rf6 39.a6 bxa6 40.Rf4 a5 41.c4 h4 42.c5 h3 43.c6 h2 44.c7 h1D 45.c8D Df1+ 46.Rg3 Dg1+ 47.Rh3 De3+ 48.Rg2 De4+ 49.Rh2 De5+ 50.Rh1 Dxf5 51.Dd8+ Rg6 52.Dd6+ Rg5 53.Dg3+ Rh5 54.Dh2+ Rg4 55.Dg1+ Rh4 56.Dd4+ Rg5 57.Dc3 Df1+ 58.Rh2 Df4+ 59.Rh3 Df5+ 60.Rg2 Rh4 61.b3 Dg5+ 62.Rf2 Rh5 63.Re2 Dg2+ 64.Rd1 Df1+ 65.Rc2 Df2+ 66.Rc1 Df4+ 67.Rd1 Dg4+ 68.Rc1 Dg1+ 69.Rc2
 43/68+    02:36     586.363k    3.743k    -1,77    35. ... f5 36.a5
 43/68+    02:43     610.099k    3.741k    -1,69    35. ... f5 36.a5
 43/68+    05:53     1.307.497k    3.703k    +2,35    35. ... f5 36.a5

[Rodolfo Leoni]

P.S. Daniel Josè (Andscacs author) made a nice version of Andscacs with PHs, with the option to load/save the hashes if an analysys lasts several days. It works fine enough, but hash size should be 32-64 MB, at least. With little wash some positions and scores get lost quickly enough. Daniel looks interested to apply that system to a newer Stockfish version, if possible. And when possible, because he must think at his engine too....!

Done!

Source and executable:
www.andscacs.com/downloads/stockfish_x6 ... vehash.zip

It has the same added uci options than in Andscacs.

So I added to Stockfish the capability of saving the full hash to file, to allow the user to recover a previous analysis session and continue it.
The saved hash file will be of the same size of the hash memory, so if you defined 4 GB of hash, such will be the file size. Saving and loading such big files can take some time.

To be able to do it I have added 4 new uci parameters:

option name NeverClearHash type check default false
option name HashFile type string default hash.hsh
option name SaveHashtoFile type button
option name LoadHashfromFile type button

You can set the NeverClearHash option to avoid that the hash could be cleared by a Clear Hash or ucinewgame command.
The HashFile parameter is the full file name with path information. If you don't set the path, it will be saved in the current folder. It defaults to hash.hsh.
To save the hash, stop the analysis and press the SaveHashtoFile button in the uci options screen of the GUI.
To load the hash file, load the game you are interested in, load the engine withouth starting it, and press the LoadHashfromFile button in the uci options screen of the GUI. Now you can start the analysis.

I have not tested it much. Anything just tell me.

Thanks Daniel !!

It's clear the reason why so many people have a great consideration of you. You always are generous and user-oriented. You behave as a friend, before being a programmer. This is a real sense of sharing an interest.

Thanks again.

Rodolfo

[BeyondCritics]

Great, but without source this is of limited utility (and against the license btw).

Could you do a fork of offcicial-stockfish on github and incorporate this changes into a new branch, which you would then push back to github? This way, your branches could be continuously rebased onto the latest stockfish, ports to mac and linux could be made and more.

[cdani]

Great, but without source this is of limited utility (and against the license btw).

Could you do a fork of offcicial-stockfish on github and incorporate this changes into a new branch, which you would then push back to github? This way, your branches could be continuously rebased onto the latest stockfish, ports to mac and linux could be made and more.

The source is included in the same file!!

Source and executable:
www.andscacs.com/downloads/stockfish_x6 ... vehash.zip

I never used seriously Github. I have to spend some time on it to understand how it works.

[Rodolfo Leoni]

..............
Could you do a fork of offcicial-stockfish on github and incorporate this changes into a new branch, which you would then push back to github? This way, your branches could be continuously rebased onto the latest stockfish, ports to mac and linux could be made and more.

Could be interesting, if/when Daniel does it. In that case, are you willing to "upgrade" it with Jeremy Bernstein's patches? I mean you, and others skilled programmers too...

As a chess analyst, I like the idea of having Andascacs versions growing fast, so I'd not ask him to dedicate too much time to a Stockfish fork, even if greatly interesting.

Rodolfo

[Rodolfo Leoni]

finally, 26 draws, or 12 white wins and then 13 draws?

It's not matter of statistics. The "learning" engine needed to play some amount of games to give a good final output. We shouldn't forget Stockfish PA GTB is 3 years old.

[MikeGL]

finally, 26 draws, or 12 white wins and then 13 draws?

It's not matter of statistics. The "learning" engine needed to play some amount of games to give a good final output. We shouldn't forget Stockfish PA GTB is 3 years old.

You are spot on. That's the real purpose of those 1 minute shootout and
nothing else.

Perhaps he needs to review the subject on this wiki page and also this one.

If I remember correctly, it was Vasik Rajlich who did this Monte Carlo thing in testing Rybka. But I could be wrong.

regards

[Lyudmil Tsvetkov]

finally, 26 draws, or 12 white wins and then 13 draws?

It's not matter of statistics. The "learning" engine needed to play some amount of games to give a good final output. We shouldn't forget Stockfish PA GTB is 3 years old.

You are spot on. That's the real purpose of those 1 minute shootout and
nothing else.

Perhaps he needs to review the subject on this wiki page and also this one.

If I remember correctly, it was Vasik Rajlich who did this Monte Carlo thing in testing Rybka. But I could be wrong.

regards

hash contains no knowledge, it just saves time by avoiding repeated calculations, so nothing very promising to expect from that.

you don't need Monte Carlo to randomise games, just running in busy environment or using multiple cores.

[Rodolfo Leoni]

hash contains no knowledge, it just saves time by avoiding repeated calculations, so nothing very promising to expect from that.

you don't need Monte Carlo to randomise games, just running in busy environment or using multiple cores.

It's not matter of knowledge. As ordinary hashes, an hash key is made by a certain amount of bytes for position, score, depth and best move. When on search, engine saves valuable entries on the PH file, depending on search depth. It's subjected to some rules, as shallower depth score cannot overwrite a deeper one, or root score is more valuable than non-root one.

The fact of having faster search depends on forward propagation (as in Critter). The min-maxing process for real "learning" is backward propagation. In Critter, the option exists (resolve score drops); in SF PA GTB id doesn't, so it depends on hash hits and it's not always working.

Critter has another advantage, though. It forward-propagates with multipv too.

[Lyudmil Tsvetkov]

hash contains no knowledge, it just saves time by avoiding repeated calculations, so nothing very promising to expect from that.

you don't need Monte Carlo to randomise games, just running in busy environment or using multiple cores.

It's not matter of knowledge. As ordinary hashes, an hash key is made by a certain amount of bytes for position, score, depth and best move. When on search, engine saves valuable entries on the PH file, depending on search depth. It's subjected to some rules, as shallower depth score cannot overwrite a deeper one, or root score is more valuable than non-root one.

The fact of having faster search depends on forward propagation (as in Critter). The min-maxing process for real "learning" is backward propagation. In Critter, the option exists (resolve score drops); in SF PA GTB id doesn't, so it depends on hash hits and it's not always working.

Critter has another advantage, though. It forward-propagates with multipv too.

as far as I know, hash key is the description of the position, and is part of the same structure, where other members of the structure as score, depth, bound, etc., would belong.