Stockfish search stops here after 99 plies, why?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Jouni
Posts: 2121
Joined: Wed Mar 08, 2006 7:15 pm

Stockfish search stops here after 99 plies, why?

Post by Jouni » Thu Nov 14, 2019 7:52 am



after 99 plies no more output in next hour even if CPU usage is still 100% :!:

analysis by Stockfish 121119 64 BMI2:

...
1.Bf2 Ke6 2.Kg1 Kd6 3.Be1 Ke6 4.Kf2 Ke7 5.Kf1 Kf6 6.Ke2 Ke6 7.Bd2 Kd7 8.Kf2 Ke6 9.Kf1 Ke7 10.Kg2 Kf6 11.Kh2 Ke7 12.Kh1 Ke6 13.Bc1 Kd6 14.Kg2 Ke7 15.Kf2 Kd6 16.Ke1 Kd7 17.Kf1 Kd6 18.Kf2 Ke6 19.Ke2 Ke7 20.Kd1 Kd7 21.Kc2 Kd6 22.Bd2 Kd7 23.Be1 Kd6 24.Bf2 Kd7 25.Kc1 Kd6 26.Kd2 Kd7 27.Kc2 Kd6 28.Be1 Ke6 29.Kb2 Kd7 30.Kb1 Kd6 31.Ka2 Ke6 32.Bf2 Kd6 33.Bg1 Kd7 34.Kb1 Kd6 35.Kb2 Ke6 36.Kc2 Kd6 37.Kd2 Kc6 38.Ke2 Kd7 39.Kf2 Ke6 40.Bh2 Kf6 41.Kg2 Ke7 42.Bg1 Kd7 43.Bf2 Kd6 44.Be1 Ke6 45.Kg1 Kd6 46.Bd2 Ke7 47.Kh2 Ke6 48.Kh1 Ke7 49.Be1 Kd7
+- (2.58) Depth: 98/98 00:00:06 71739kN
nothing
nothing
...

Something is wrong, when MAX_PLY is 246.
Jouni

Raphexon
Posts: 166
Joined: Sun Mar 17, 2019 11:00 am
Full name: Henk Drost

Re: Stockfish search stops here after 99 plies, why?

Post by Raphexon » Thu Nov 14, 2019 12:06 pm

It hasn't stopped though.

It just can't detect the fortress and thinks it has an advantage.
After 99 plies it's desperately trying to find the combination that doesn't draw. Since this position is 100% locked up and Stockfish has no fortress detection it's brute forcing it.

It will take a very long time before Stockfish reaches 100 ply depth and properly evaluates the position as a draw.

Michel
Posts: 2087
Joined: Sun Sep 28, 2008 11:50 pm

Re: Stockfish search stops here after 99 plies, why?

Post by Michel » Thu Nov 14, 2019 1:51 pm

Jouni wrote:
Thu Nov 14, 2019 7:52 am


after 99 plies no more output in next hour even if CPU usage is still 100% :!:

analysis by Stockfish 121119 64 BMI2:

...
1.Bf2 Ke6 2.Kg1 Kd6 3.Be1 Ke6 4.Kf2 Ke7 5.Kf1 Kf6 6.Ke2 Ke6 7.Bd2 Kd7 8.Kf2 Ke6 9.Kf1 Ke7 10.Kg2 Kf6 11.Kh2 Ke7 12.Kh1 Ke6 13.Bc1 Kd6 14.Kg2 Ke7 15.Kf2 Kd6 16.Ke1 Kd7 17.Kf1 Kd6 18.Kf2 Ke6 19.Ke2 Ke7 20.Kd1 Kd7 21.Kc2 Kd6 22.Bd2 Kd7 23.Be1 Kd6 24.Bf2 Kd7 25.Kc1 Kd6 26.Kd2 Kd7 27.Kc2 Kd6 28.Be1 Ke6 29.Kb2 Kd7 30.Kb1 Kd6 31.Ka2 Ke6 32.Bf2 Kd6 33.Bg1 Kd7 34.Kb1 Kd6 35.Kb2 Ke6 36.Kc2 Kd6 37.Kd2 Kc6 38.Ke2 Kd7 39.Kf2 Ke6 40.Bh2 Kf6 41.Kg2 Ke7 42.Bg1 Kd7 43.Bf2 Kd6 44.Be1 Ke6 45.Kg1 Kd6 46.Bd2 Ke7 47.Kh2 Ke6 48.Kh1 Ke7 49.Be1 Kd7
+- (2.58) Depth: 98/98 00:00:06 71739kN
nothing
nothing
...

Something is wrong, when MAX_PLY is 246.
After 99 plies SF failed low (suddenly discovering the 50 move draw). Then it starts researching the position at lower depth. But in order not to confuse the user it does not print those lower depths. That's why it does not print anything.
Ideas=science. Simplification=engineering.
Without ideas there is nothing to simplify.

User avatar
Eelco de Groot
Posts: 4236
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Stockfish search stops here after 99 plies, why?

Post by Eelco de Groot » Thu Nov 14, 2019 2:48 pm

Assumimg that Stockfish really sees the 50 move rule here, it should but I can't remember Stockfish actually doing this myself, but then it will find that every value other than a draw is wrong. Because even with only three pieces moving there are a lot of transpositions all with the same value because of the material plus for White, all the values in hash are wrong. All moves have the same value still but now should all have VALUE_DRAW. This can take forever to correct. Stockfish simply 'never' returns to the root at least not in foreseeable time That would be my guess IID should help as Michel says but probably it's out of its depth here, does not change much in the time? I have not read the fortress discussion Jouni mentioned, probably they have better answers there? Basically this is a fortress so you should see it with any fortress, only not everywhere SF can reach 100 plies deep.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

Colin-G
Posts: 157
Joined: Mon Oct 31, 2016 5:30 pm
Location: England

Re: Stockfish search stops here after 99 plies, why?

Post by Colin-G » Thu Nov 14, 2019 4:14 pm

A different way of analysing positions such as this one is required.
Since the Black king could only be on 24 different squares in any game played from this position and the White bishop on only 8 different squares, leaving only 15 different squares for the White king, there would be only 24x8x15=2880 positions that could occur in the game.
Since none of these 2880 positions would be a checkmate, the game is drawn.

It should be fairly trivial for a chess program to detect this if it was looking for it.
e.g. find all the possible squares for any pieces that can be moved.
Create a set of all possible future positions.
Check each position for checkmate at depth 0
If no checkmate is found for any of the positions, the original position is a draw.

I'm not suggesting it is practical to use this method at all, but it would work in similar fortress positions which humans immediately see as drawn and computers do not.

Karlo Bala
Posts: 306
Joined: Wed Mar 22, 2006 9:17 am
Location: Novi Sad, Serbia

Re: Stockfish search stops here after 99 plies, why?

Post by Karlo Bala » Thu Nov 14, 2019 9:14 pm

Jouni wrote:
Thu Nov 14, 2019 7:52 am


after 99 plies no more output in next hour even if CPU usage is still 100% :!:

analysis by Stockfish 121119 64 BMI2:

...
1.Bf2 Ke6 2.Kg1 Kd6 3.Be1 Ke6 4.Kf2 Ke7 5.Kf1 Kf6 6.Ke2 Ke6 7.Bd2 Kd7 8.Kf2 Ke6 9.Kf1 Ke7 10.Kg2 Kf6 11.Kh2 Ke7 12.Kh1 Ke6 13.Bc1 Kd6 14.Kg2 Ke7 15.Kf2 Kd6 16.Ke1 Kd7 17.Kf1 Kd6 18.Kf2 Ke6 19.Ke2 Ke7 20.Kd1 Kd7 21.Kc2 Kd6 22.Bd2 Kd7 23.Be1 Kd6 24.Bf2 Kd7 25.Kc1 Kd6 26.Kd2 Kd7 27.Kc2 Kd6 28.Be1 Ke6 29.Kb2 Kd7 30.Kb1 Kd6 31.Ka2 Ke6 32.Bf2 Kd6 33.Bg1 Kd7 34.Kb1 Kd6 35.Kb2 Ke6 36.Kc2 Kd6 37.Kd2 Kc6 38.Ke2 Kd7 39.Kf2 Ke6 40.Bh2 Kf6 41.Kg2 Ke7 42.Bg1 Kd7 43.Bf2 Kd6 44.Be1 Ke6 45.Kg1 Kd6 46.Bd2 Ke7 47.Kh2 Ke6 48.Kh1 Ke7 49.Be1 Kd7
+- (2.58) Depth: 98/98 00:00:06 71739kN
nothing
nothing
...

Something is wrong, when MAX_PLY is 246.
The behavior of SF is perfectly sound.

At play 99 it starts to kick in the 50 move rule. At that moment SF realizes that the entire tree is wrong, so start to build it up from scratch.

The score of PV drops to zero. The next move is a potential PV with a score >2.0. This move also drops to zero, and so on. Every move is searched as PV, which means without reductions!!!
Best Regards,
Karlo Balla Jr.

User avatar
MikeB
Posts: 3933
Joined: Thu Mar 09, 2006 5:34 am
Location: Pen Argyl, Pennsylvania

Re: Stockfish search stops here after 99 plies, why?

Post by MikeB » Sat Nov 16, 2019 6:14 am

Jouni wrote:
Thu Nov 14, 2019 7:52 am


after 99 plies no more output in next hour even if CPU usage is still 100% :!:

analysis by Stockfish 121119 64 BMI2:

...
1.Bf2 Ke6 2.Kg1 Kd6 3.Be1 Ke6 4.Kf2 Ke7 5.Kf1 Kf6 6.Ke2 Ke6 7.Bd2 Kd7 8.Kf2 Ke6 9.Kf1 Ke7 10.Kg2 Kf6 11.Kh2 Ke7 12.Kh1 Ke6 13.Bc1 Kd6 14.Kg2 Ke7 15.Kf2 Kd6 16.Ke1 Kd7 17.Kf1 Kd6 18.Kf2 Ke6 19.Ke2 Ke7 20.Kd1 Kd7 21.Kc2 Kd6 22.Bd2 Kd7 23.Be1 Kd6 24.Bf2 Kd7 25.Kc1 Kd6 26.Kd2 Kd7 27.Kc2 Kd6 28.Be1 Ke6 29.Kb2 Kd7 30.Kb1 Kd6 31.Ka2 Ke6 32.Bf2 Kd6 33.Bg1 Kd7 34.Kb1 Kd6 35.Kb2 Ke6 36.Kc2 Kd6 37.Kd2 Kc6 38.Ke2 Kd7 39.Kf2 Ke6 40.Bh2 Kf6 41.Kg2 Ke7 42.Bg1 Kd7 43.Bf2 Kd6 44.Be1 Ke6 45.Kg1 Kd6 46.Bd2 Ke7 47.Kh2 Ke6 48.Kh1 Ke7 49.Be1 Kd7
+- (2.58) Depth: 98/98 00:00:06 71739kN
nothing
nothing
...

Something is wrong, when MAX_PLY is 246.
An engine with fortress detection doesn't need 100 plies to see the draw ( Scoring % output)
Honey-XR7-FD

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
101	+50.00 	8.04G	3:54.25	Bf2 Kd7 Be1 Kc6 Bf2 
100	+50.00 	7.87G	3:48.82	Bf2 Ke6 Be1 Kd6 Bf2 
 99	+50.00 	7.37G	3:33.63	Bf2 Kf6 Be1 Ke7 
 98	+50.00 	7.02G	3:23.07	Bf2 Kf6 Be1 Kg6 Bf2 
 97	+50.00 	6.72G	3:13.90	Bf2 Ke6 Kg2 Kd6 Kh1 
 96	+50.00 	6.56G	3:09.24	Bf2 Ke6 Kg2 Kd7 Kh1 
 95	+50.00 	6.38G	3:03.71	Bf2 Ke6 Kg2 Ke7 
 94	+50.00 	5.87G	2:48.20	Bf2 Kd7 Kg2 Ke7 
 93	+50.00 	5.71G	2:43.16	Bf2 Kd7 Kg2 Ke6 Bg1 Kf7 Kf2 Kg8 Kg2 
 92	+50.00 	5.55G	2:38.40	Bf2 Ke6 Kg2 Kd6 Bg1 Ke6 
 91	+50.00 	4.68G	2:12.57	Bf2 Kd7 Kg2 Ke6 Bg1 Kf6 Kf2 Ke6 
 90	+50.00 	4.43G	2:04.99	Bf2 Kd7 Kg2 Kc6 Bg1 Kd7 
 89	+50.00 	3.90G	1:49.47	Bf2 Kd6 Kg2 Kd7 Bg1 Kd6 
 88	+50.00 	2.98G	1:22.36	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke6 
 87	+50.00 	2.27G	1:01.85	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Bf2 
 86	+50.00 	2.04G	0:55.02	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke7 Be1 
 85	+50.00 	1.93G	0:51.80	Bf2 Kd7 Kg2 Kd6 Be1 Ke7 Kf2 Kd6 
 84	+50.00 	1.82G	0:48.91	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke6 
 83	+50.00 	1.78G	0:47.54	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Kc7 Bc1 Kc6 Bb2 Kb7 Kf2 Ka6 Ba1 Ka7 Bb2 
 82	+50.00 	1.68G	0:44.62	Bf2 Kd7 Kg2 Kc6 Be1 Kd6 Bd2 Ke6 Be1 
 81	+50.00 	1.59G	0:42.31	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Bf2 
 80	+50.00 	1.55G	0:41.20	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Kc6 Be1 
 79	+50.00 	1.52G	0:40.20	Bf2 Kd7 Kg2 Kc7 Bg1 Kb7 Bf2 
 78	+50.00 	1.49G	0:39.39	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 77	+50.00 	1.41G	0:37.15	Bf2 Kd7 Kg2 Kd6 Be1 Kc7 Kf2 Kc6 Ke2 Kd7 Kd2 Kc6 
 76	+50.00 	1.35G	0:35.40	Bf2 Kd6 Bg1 Ke6 Kg2 Kd7 Kf2 Kd6 Ke2 Ke7 Kd2 Kd7 Bf2 Ke7 
 75	+50.00 	1.30G	0:34.23	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 74	+50.00 	1.25G	0:32.73	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 73	+50.00 	1.22G	0:31.79	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 72	+50.00 	1.18G	0:30.82	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 71	+50.00 	1.04G	0:27.09	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Kh1 Kd6 
 70	+50.00 	982.9M	0:25.50	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Kh1 Kd7 Bd2 Ke7 Be1 
 69	+50.00 	940.1M	0:24.32	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 68	+50.00 	812.8M	0:20.92	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 67	+50.00 	753.4M	0:19.33	Bf2 Ke6 Kg2 Kd6 Be1 Kc7 Bf2 
 66	+50.00 	683.6M	0:17.51	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 65	+50.00 	598.7M	0:15.28	Bf2 Ke6 Kg2 Kd7 Bg1 Kd6 Kf2 Kd7 
 64	+50.00 	537.2M	0:13.66	Bf2 Ke6 Kg2 Kd6 Be1 Ke7 Bd2 Kd6 
 63	+50.00 	425.4M	0:10.76	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 62	+50.00 	364.0M	0:09.17	Bf2 Ke6 Kg2 Kd6 Be1 Kd7 Kf2 Kd6 
 61	+50.00 	331.3M	0:08.31	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 60	+50.00 	295.9M	0:07.41	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 59	+50.00 	253.0M	0:06.32	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Bd2 Kb7 Kf2 Kb6 Bc1 Kc7 Bb2 Kd6 Ke1 Ke6 Bc1 Kd6 
 58	+50.00 	215.4M	0:05.38	Bf2 Ke6 Kg2 Kd7 Bg1 Kc6 Bh2 Kb6 Bg1 
 57	+50.00 	191.6M	0:04.76	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 56	+50.00 	159.9M	0:03.96	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 55	+50.00 	139.7M	0:03.46	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 54	+50.00 	122.1M	0:03.03	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 53	+50.00 	105.2M	0:02.62	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 52	+50.00 	88.6M  	0:02.20	Bf2 Ke6 Kg2 Kd7 Bg1 Ke8 Kf2 Ke7 Ke2 Kd7 Kd2 Kc6 Bf2 Kc7 Be1 Kd7 Bf2 
 51	+50.00 	72.3M  	0:01.80	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 50	+50.00 	61.1M  	0:01.53	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 49	+50.00 	53.5M  	0:01.34	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 48	+50.00 	39.0M  	0:00.99	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 47	+50.00 	31.8M  	0:00.81	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 46	+50.00 	23.7M  	0:00.60	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 45	+50.00 	17.8M  	0:00.45	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 44	+50.00 	15.1M  	0:00.38	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 43	+50.00 	12.6M  	0:00.32	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 42	+50.00 	10.5M  	0:00.27	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 41	+50.00 	8.68M  	0:00.22	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 40	+50.00 	7.54M  	0:00.19	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 39	+50.00 	5.17M  	0:00.14	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 38	+50.00 	4.16M  	0:00.11	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 37	+50.00 	3.36M  	0:00.09	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 36	+50.00 	2.47M  	0:00.07	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 35	+50.00 	2.00M  	0:00.06	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 34	+50.00 	1.80M  	0:00.05	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 33	+50.00 	1.44M  	0:00.04	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 32	+50.00 	1.13M  	0:00.03	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 31	+50.00 	838638	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 30	+50.00 	641960	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 29	+50.00 	553002	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 28	+50.00 	474090	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 27	+50.00 	360224	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 26	+50.00 	309896	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 25	+50.00 	298094	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 24	+50.00 	270102	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 23	+50.00 	230474	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 22	+50.00 	189380	0:00.01	Bf2 Ke6 Kg2 Ke7 
 21	+50.00 	162240	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 20	+50.00 	154038	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 19	+50.00 	141144	0:00.01	Bf2 Ke6 Kg2 Ke7 
 18	+50.00 	126535	0:00.01	Bf2 Ke6 Kg2 Ke7 
 17	+50.00 	113448	0:00.01	Bf2 Ke6 Kg2 Ke7 
 16	+50.00 	100209	0:00.00	Bf2 Ke6 Kg2 Ke7 
 15	+50.00 	90995  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 14	+50.00 	86009  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 13	+50.00 	82581  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 12	+50.00 	78032  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 11	+50.00 	75022  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 10	+50.00 	71756  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 Ke2 Ke6 
  9	+50.00 	69751  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 Ke2 
  8	+50.00 	67865  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 
  7	+72.70 	17685  	0:00.00	Bf2 Kf6 Be1 Ke6 Kh2 Ke7 Kg2 
  6	+72.07 	12308  	0:00.00	Bf2 Kf6 Kh2 Kf7 Kg2 
  5	+70.93 	5884    	0:00.00	Bf2 Kf6 Kh2 Ke6 Kg2 
  4	+71.30 	4605    	0:00.00	Bf2 Kf6 Kh2 
  3	+71.27 	3097    	0:00.00	Kg2 Ke6 Bf2 
  2	+70.83 	1898    	0:00.00	Kg2 Kf6 
  1	+72.70 	338      	0:00.00	Kg2 
  1	Found 1511 tablebases 
  0	# 

Uri Blass
Posts: 8752
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Stockfish search stops here after 99 plies, why?

Post by Uri Blass » Sat Nov 16, 2019 6:56 am

MikeB wrote:
Sat Nov 16, 2019 6:14 am
An engine with fortress detection doesn't need 100 plies to see the draw
An engine that works correctly does not need 100 plies to see the draw even with no fortress detection.

It is enough to see that the number of possible positions after 26 plies is equal to the number of possible positions after 28 plies when no position is mate to get the conclusion that it is a draw.

Jouni
Posts: 2121
Joined: Wed Mar 08, 2006 7:15 pm

Re: Stockfish search stops here after 99 plies, why?

Post by Jouni » Mon Nov 18, 2019 4:58 pm

OK no bug here but search explosion. BTW Crystal is clever and scores this fast 0,00:

Analysis by Crystal 270819 64 BMI2:

...
1.Bf2 Kd6 2.Kg1 Ke6 3.Kg2 Ke7
= (0.00) Depth: 34/14 00:00:01 23616kN
1.Bf2 Kd6 2.Kg1 Ke6 3.Be1 Kd7 4.Kf2 Kd6 5.Bd2 Ke7 6.Kf1 Kd6
= (0.00) Depth: 35/14 00:00:03 55612kN
1.Bf2 Kd6 2.Kg1 Kd7 3.Kg2 Ke7
= (0.00) Depth: 36/16 00:00:06 94670kN
Jouni

Paloma
Posts: 916
Joined: Thu Dec 25, 2008 8:07 pm

Re: Stockfish search stops here after 99 plies, why?

Post by Paloma » Wed Nov 20, 2019 7:26 pm

MikeB wrote:
Sat Nov 16, 2019 6:14 am

An engine with fortress detection doesn't need 100 plies to see the draw ( Scoring % output)
Honey-XR7-FD

Code: Select all

dep	score	nodes	time	(not shown:  tbhits	knps	seldep)
101	+50.00 	8.04G	3:54.25	Bf2 Kd7 Be1 Kc6 Bf2 
100	+50.00 	7.87G	3:48.82	Bf2 Ke6 Be1 Kd6 Bf2 
 99	+50.00 	7.37G	3:33.63	Bf2 Kf6 Be1 Ke7 
 98	+50.00 	7.02G	3:23.07	Bf2 Kf6 Be1 Kg6 Bf2 
 97	+50.00 	6.72G	3:13.90	Bf2 Ke6 Kg2 Kd6 Kh1 
 96	+50.00 	6.56G	3:09.24	Bf2 Ke6 Kg2 Kd7 Kh1 
 95	+50.00 	6.38G	3:03.71	Bf2 Ke6 Kg2 Ke7 
 94	+50.00 	5.87G	2:48.20	Bf2 Kd7 Kg2 Ke7 
 93	+50.00 	5.71G	2:43.16	Bf2 Kd7 Kg2 Ke6 Bg1 Kf7 Kf2 Kg8 Kg2 
 92	+50.00 	5.55G	2:38.40	Bf2 Ke6 Kg2 Kd6 Bg1 Ke6 
 91	+50.00 	4.68G	2:12.57	Bf2 Kd7 Kg2 Ke6 Bg1 Kf6 Kf2 Ke6 
 90	+50.00 	4.43G	2:04.99	Bf2 Kd7 Kg2 Kc6 Bg1 Kd7 
 89	+50.00 	3.90G	1:49.47	Bf2 Kd6 Kg2 Kd7 Bg1 Kd6 
 88	+50.00 	2.98G	1:22.36	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke6 
 87	+50.00 	2.27G	1:01.85	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Bf2 
 86	+50.00 	2.04G	0:55.02	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke7 Be1 
 85	+50.00 	1.93G	0:51.80	Bf2 Kd7 Kg2 Kd6 Be1 Ke7 Kf2 Kd6 
 84	+50.00 	1.82G	0:48.91	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Ke6 
 83	+50.00 	1.78G	0:47.54	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Kc7 Bc1 Kc6 Bb2 Kb7 Kf2 Ka6 Ba1 Ka7 Bb2 
 82	+50.00 	1.68G	0:44.62	Bf2 Kd7 Kg2 Kc6 Be1 Kd6 Bd2 Ke6 Be1 
 81	+50.00 	1.59G	0:42.31	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Bf2 
 80	+50.00 	1.55G	0:41.20	Bf2 Kd7 Kg2 Ke6 Be1 Kd6 Bd2 Kc6 Be1 
 79	+50.00 	1.52G	0:40.20	Bf2 Kd7 Kg2 Kc7 Bg1 Kb7 Bf2 
 78	+50.00 	1.49G	0:39.39	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 77	+50.00 	1.41G	0:37.15	Bf2 Kd7 Kg2 Kd6 Be1 Kc7 Kf2 Kc6 Ke2 Kd7 Kd2 Kc6 
 76	+50.00 	1.35G	0:35.40	Bf2 Kd6 Bg1 Ke6 Kg2 Kd7 Kf2 Kd6 Ke2 Ke7 Kd2 Kd7 Bf2 Ke7 
 75	+50.00 	1.30G	0:34.23	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 74	+50.00 	1.25G	0:32.73	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 73	+50.00 	1.22G	0:31.79	Bf2 Kd7 Kg2 Kd6 Be1 Ke6 Kf2 Kd6 
 72	+50.00 	1.18G	0:30.82	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 71	+50.00 	1.04G	0:27.09	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Kh1 Kd6 
 70	+50.00 	982.9M	0:25.50	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Kh1 Kd7 Bd2 Ke7 Be1 
 69	+50.00 	940.1M	0:24.32	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 68	+50.00 	812.8M	0:20.92	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 67	+50.00 	753.4M	0:19.33	Bf2 Ke6 Kg2 Kd6 Be1 Kc7 Bf2 
 66	+50.00 	683.6M	0:17.51	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 65	+50.00 	598.7M	0:15.28	Bf2 Ke6 Kg2 Kd7 Bg1 Kd6 Kf2 Kd7 
 64	+50.00 	537.2M	0:13.66	Bf2 Ke6 Kg2 Kd6 Be1 Ke7 Bd2 Kd6 
 63	+50.00 	425.4M	0:10.76	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 62	+50.00 	364.0M	0:09.17	Bf2 Ke6 Kg2 Kd6 Be1 Kd7 Kf2 Kd6 
 61	+50.00 	331.3M	0:08.31	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 60	+50.00 	295.9M	0:07.41	Bf2 Ke6 Kg2 Kd6 Be1 Ke6 
 59	+50.00 	253.0M	0:06.32	Bf2 Ke6 Kg2 Kd6 Be1 Kc6 Bd2 Kb7 Kf2 Kb6 Bc1 Kc7 Bb2 Kd6 Ke1 Ke6 Bc1 Kd6 
 58	+50.00 	215.4M	0:05.38	Bf2 Ke6 Kg2 Kd7 Bg1 Kc6 Bh2 Kb6 Bg1 
 57	+50.00 	191.6M	0:04.76	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 56	+50.00 	159.9M	0:03.96	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 55	+50.00 	139.7M	0:03.46	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 54	+50.00 	122.1M	0:03.03	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 53	+50.00 	105.2M	0:02.62	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 52	+50.00 	88.6M  	0:02.20	Bf2 Ke6 Kg2 Kd7 Bg1 Ke8 Kf2 Ke7 Ke2 Kd7 Kd2 Kc6 Bf2 Kc7 Be1 Kd7 Bf2 
 51	+50.00 	72.3M  	0:01.80	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 50	+50.00 	61.1M  	0:01.53	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 49	+50.00 	53.5M  	0:01.34	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 48	+50.00 	39.0M  	0:00.99	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 47	+50.00 	31.8M  	0:00.81	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 46	+50.00 	23.7M  	0:00.60	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 45	+50.00 	17.8M  	0:00.45	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 44	+50.00 	15.1M  	0:00.38	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 43	+50.00 	12.6M  	0:00.32	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 42	+50.00 	10.5M  	0:00.27	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 41	+50.00 	8.68M  	0:00.22	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 40	+50.00 	7.54M  	0:00.19	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 39	+50.00 	5.17M  	0:00.14	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 38	+50.00 	4.16M  	0:00.11	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 37	+50.00 	3.36M  	0:00.09	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 36	+50.00 	2.47M  	0:00.07	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 35	+50.00 	2.00M  	0:00.06	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 34	+50.00 	1.80M  	0:00.05	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 33	+50.00 	1.44M  	0:00.04	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 32	+50.00 	1.13M  	0:00.03	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 31	+50.00 	838638	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 30	+50.00 	641960	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 29	+50.00 	553002	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 28	+50.00 	474090	0:00.02	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 27	+50.00 	360224	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 26	+50.00 	309896	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 25	+50.00 	298094	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 24	+50.00 	270102	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 23	+50.00 	230474	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 22	+50.00 	189380	0:00.01	Bf2 Ke6 Kg2 Ke7 
 21	+50.00 	162240	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 20	+50.00 	154038	0:00.01	Bf2 Ke6 Kg2 Kd7 Bg1 Ke6 
 19	+50.00 	141144	0:00.01	Bf2 Ke6 Kg2 Ke7 
 18	+50.00 	126535	0:00.01	Bf2 Ke6 Kg2 Ke7 
 17	+50.00 	113448	0:00.01	Bf2 Ke6 Kg2 Ke7 
 16	+50.00 	100209	0:00.00	Bf2 Ke6 Kg2 Ke7 
 15	+50.00 	90995  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 14	+50.00 	86009  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 13	+50.00 	82581  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 12	+50.00 	78032  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 11	+50.00 	75022  	0:00.00	Bf2 Ke6 Kg2 Ke7 
 10	+50.00 	71756  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 Ke2 Ke6 
  9	+50.00 	69751  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 Ke2 
  8	+50.00 	67865  	0:00.00	Bf2 Ke6 Kg2 Kd6 Bg1 Kd7 Kf2 Ke7 
  7	+72.70 	17685  	0:00.00	Bf2 Kf6 Be1 Ke6 Kh2 Ke7 Kg2 
  6	+72.07 	12308  	0:00.00	Bf2 Kf6 Kh2 Kf7 Kg2 
  5	+70.93 	5884    	0:00.00	Bf2 Kf6 Kh2 Ke6 Kg2 
  4	+71.30 	4605    	0:00.00	Bf2 Kf6 Kh2 
  3	+71.27 	3097    	0:00.00	Kg2 Ke6 Bf2 
  2	+70.83 	1898    	0:00.00	Kg2 Kf6 
  1	+72.70 	338      	0:00.00	Kg2 
  1	Found 1511 tablebases 
  0	# 
Hi Mike,
is Honey-XR7-FD out or released soon?

Post Reply