Stockfish 16 released

[ImNotStockfish]

I don't think you understand how minority we "who use abrok compiles" are, among the millions of people using Stockfish only via official releases. People simply don't know they exist, and online chess platforms or mobile gui's are no help for that. Lots of people only know when there is an official release and their platforms use the new engine. So, actually, no need to wait or froze master for an official release, abrok users know how to download from there and official-to-official users don't know development versions exist.

And you would be surprised with the amount of abrok users that don't know you don't need abrok anymore to download the latest development build

[Eduard]

My short test between SF 16 and SF dev 010723 in tournament mode:
https://talkchess.com/forum3/viewtopic. ... 79#p949775

[Peter Berger]

This Stockfish program, though still quite young, has become a serious challenge to beat and overcome for most engines.

Here is a try against a recent Crafty version:

[Event "Lang 120min+10sek"]
[Site "Berlin"]
[Date "2023.07.05"]
[Round "?"]
[White "Crafty 25.6"]
[Black "Stockfish 16"]
[Result "1/2-1/2"]
[ECO "C67"]
[PlyCount "50"]
[TimeControl "7200+10"]

{4096MB, LAPTOP-NCDN8BTK} 1. e4 {[%emt 0:00:00]} e5 {[%eval 20,42] [%emt 0:03:
07]} 2. Nf3 {[%emt 0:00:07]} Nc6 {[%eval 15,42] [%emt 0:01:26]} 3. Bb5 {
[%emt 0:00:07] (Lc4)} Nf6 {[%eval 22,45] [%emt 0:01:59]} 4. O-O {[%emt 0:00:07]
} Nxe4 {[%eval 15,46] [%emt 0:01:34]} 5. d4 {[%emt 0:00:07] (Te1)} Nd6 {
[%eval 13,43] [%emt 0:01:58]} 6. Bxc6 {[%emt 0:00:06]} dxc6 {[%eval 14,43]
[%emt 0:02:12]} 7. dxe5 {[%emt 0:00:06]} Nf5 {[%eval 9,43] [%emt 0:01:45]} 8.
Qxd8+ {[%emt 0:00:07]} Kxd8 {[%eval 7,36] [%emt 0:00:05]} 9. Nc3 {[%emt 0:00:
06] (h3)} Be7 {[%eval 7,44] [%emt 0:01:59]} 10. Rd1+ {[%emt 0:00:08]} Ke8 {
[%eval 8,42] [%emt 0:01:59]} 11. Bg5 {[%emt 0:00:07] (h3)} Bxg5 {[%eval 10,39]
[%emt 0:01:54]} 12. Nxg5 {[%emt 0:00:07]} h6 {[%eval 6,42] [%emt 0:02:02]} 13.
Nge4 {[%emt 0:03:07] (Sf3)} Ke7 {[%eval 0,52] [%emt 0:02:06]} 14. f4 {[%emt 0:
07:10]} f6 {[%eval 0,53] [%emt 0:00:01]} 15. exf6+ {[%emt 0:02:48] (Td2)} gxf6
{[%eval 0,61] [%emt 0:01:55]} 16. Kf2 {[%emt 0:00:55] (Td2)} h5 {[%eval 0,61]
[%emt 0:02:06]} 17. b4 {[%emt 0:08:32] (Td2)} Kf7 {[%eval 0,50] [%emt 0:01:59]}
18. b5 {[%emt 0:02:58]} Re8 {[%eval 0,51] [%emt 0:00:01]} 19. bxc6 {[%emt 0:02:
23]} bxc6 {[%eval 0,52] [%emt 0:00:01]} 20. Rab1 {[%emt 0:02:29]} a5 {[%eval 0,
54] [%emt 0:00:01]} 21. a4 {[%emt 0:09:18] (Td2)} Kg6 {[%eval 0,53] [%emt 0:02:
01]} 22. Rd2 {[%emt 0:04:27]} h4 {[%eval 0,50] [%emt 0:00:01]} 23. Rbd1 {
[%emt 0:03:13]} Nh6 {[%eval 0,55] [%emt 0:02:09]} 24. h3 {[%emt 0:00:09]} Bf5 {
[%eval 0,57] [%emt 0:01:20]} 25. Nc5 {[%emt 0:01:38] (Td4)} Rab8 {[%eval 0,61]
[%emt 0:02:01]} 1/2-1/2

And yes, this is indeed the one and only game I played with these two ( also the only game by Stockfish 16 on my computers).

I have written about this phenomenon before, I have thought about it a little more in the meantime. The Stockfish testing procedure is amazing and effective - it delivers exactly what it is supposed to deliver. Optimizing things also means you ignore things of minor importance to you - so they are not taken into account.

Thus I'll try to simplify my message a little : Crafty will like never achieve a draw against current version of say Rebel: if in any rating list Stockfish had to play against Crafty more often, you'd see a serious decline in rating. Crafty can draw against Stockfish way more often than it is supposed to be able to do that.

[carldaman]

It's an inherent problem with NNUE risk aversion and lack of contempt settings.

[syzygy]

What is the idea to release stockfish16 in 29.6 when you already know that some tests at LTC passed SPRT and are not included?

Because a regression test has been done on SF16 is considered to meet all the requirements to be released as SF16.

Excuse me, but it is beyond ridiculous to make a point of criticism out of releasing SF16 when somehere down the line a further improvement might be looming.

[syzygy]

Crafty will like never achieve a draw against current version of say Rebel

Of course it will.

[dkappe]

Crafty will like never achieve a draw against current version of say Rebel

Of course it will.

We know that just 5 years after release an engine starts playing much worse. I’d estimate that Crafty is no better than an 1800 player.

[CornfedForever]

A line in the sand has to be drawn somewhere. I don't see any issues. If they've already made gains, all well and good - they'll be able to release v16.1 or v17 sooner than otherwise.

+1 As I said earlier...just means 17 will get released earlier.

Put 2 people in a room and 1 will always find something to complain about.

[Peter Berger]

Crafty will like never achieve a draw against current version of say Rebel

Of course it will.

Please show a game with some decent time control (so no super bullet, please,) I said "like never" as I am enough of a maths guy to know that this would happen at some point in time eventually. I haven't been able to produce such a game though, although I have probably tried more often than other ordinary earthlings

[syzygy]

Crafty will like never achieve a draw against current version of say Rebel

Of course it will.

Please show a game with some decent time control (so no super bullet, please,) I said "like never" as I am enough of a maths guy to know that this would happen at some point in time eventually. I haven't been able to produce such a game though, although I have probably tried more often than other ordinary earthlings

First show us a game where Crafty actually draws Stockfish:
[pgn][Event "Computer Chess Game"]
[Date "2023.07.07"]
[Round "-"]
[White "Crafty-25.2"]
[Black "Stockfish dev-20230706-e699fee5"]
[Result "0-1"]
[TimeControl "40/300"]
[Annotator "26. +0.00 26... +0.01"]

{4096MB, LAPTOP-NCDN8BTK}
1. e4
{[%emt 0:00:00]}
1... e5
{[%eval 20,42] [%emt 0:03:
07]}
2. Nf3
{[%emt 0:00:07]}
2... Nc6
{[%eval 15,42] [%emt 0:01:26]}
3. Bb5
{
[%emt 0:00:07] (Lc4)}
3... Nf6
{[%eval 22,45] [%emt 0:01:59]}
4. O-O
{[%emt 0:00:07]
}
4... Nxe4
{[%eval 15,46] [%emt 0:01:34]}
5. d4
{[%emt 0:00:07] (Te1)}
5... Nd6
{
[%eval 13,43] [%emt 0:01:58]}
6. Bxc6
{[%emt 0:00:06]}
6... dxc6
{[%eval 14,43]
[%emt 0:02:12]}
7. dxe5
{[%emt 0:00:06]}
7... Nf5
{[%eval 9,43] [%emt 0:01:45]}
8. Qxd8+
{[%emt 0:00:07]}
8... Kxd8
{[%eval 7,36] [%emt 0:00:05]}
9. Nc3
{[%emt 0:00:
06] (h3)}
9... Be7
{[%eval 7,44] [%emt 0:01:59]}
10. Rd1+
{[%emt 0:00:08]}
10... Ke8
{
[%eval 8,42] [%emt 0:01:59]}
11. Bg5
{[%emt 0:00:07] (h3)}
11... Bxg5
{[%eval 10,39]
[%emt 0:01:54]}
12. Nxg5
{[%emt 0:00:07]}
12... h6
{[%eval 6,42] [%emt 0:02:02]}
13. Nge4
{[%emt 0:03:07] (Sf3)}
13... Ke7
{[%eval 0,52] [%emt 0:02:06]}
14. f4
{[%emt 0:
07:10]}
14... f6
{[%eval 0,53] [%emt 0:00:01]}
15. exf6+
{[%emt 0:02:48] (Td2)}
15... gxf6
{[%eval 0,61] [%emt 0:01:55]}
16. Kf2
{[%emt 0:00:55] (Td2)}
16... h5
{[%eval 0,61]
[%emt 0:02:06]}
17. b4
{[%emt 0:08:32] (Td2)}
17... Kf7
{[%eval 0,50] [%emt 0:01:59]}
18. b5
{[%emt 0:02:58]}
18... Re8
{[%eval 0,51] [%emt 0:00:01]}
19. bxc6
{[%emt 0:02:
23]}
19... bxc6
{[%eval 0,52] [%emt 0:00:01]}
20. Rab1
{[%emt 0:02:29]}
20... a5
{[%eval 0,
54] [%emt 0:00:01]}
21. a4
{[%emt 0:09:18] (Td2)}
21... Kg6
{[%eval 0,53] [%emt 0:02:
01]}
22. Rd2
{[%emt 0:04:27]}
22... h4
{[%eval 0,50] [%emt 0:00:01]}
23. Rbd1
{
[%emt 0:03:13]}
23... Nh6
{[%eval 0,55] [%emt 0:02:09]}
24. h3
{[%emt 0:00:09]}
24... Bf5
{
[%eval 0,57] [%emt 0:01:20]}
25. Nc5
{[%emt 0:01:38] (Td4)}
25... Rab8 26. Ne2 {+0.00/32 9} Rb4 {+0.01/56 30} 27. Rd4 {+0.00/33 9} Rb2
{+0.00/60 15} 28. R1d2 {+0.00/33 9} Nf7 {+0.00/55 12} 29. Ng1 {+0.00/33 9}
Nd6 {+0.10/47 50} 30. Nf3 {+0.00/35 9} Kh5 {+0.08/45 8} 31. c3 {+0.00/34 9}
Rb1 {+0.08/47 14} 32. Rd1 {-0.26/33 37} Rbb8 {+0.06/49 15} 33. R1d2
{-0.26/31 9} Rg8 {+0.15/36 17} 34. Re2 {+0.00/33 9} Rg3 {+0.17/41 14} 35.
Na6 {+0.00/32 9} Rbg8 {+0.79/39 16} 36. Ne1 {-0.76/30 11} Bxh3
{+1.06/38 18} 37. Nxc7 {-1.01/30 12} Nf5 {+1.57/35 19} 38. gxh3
{-0.65/30 12} Nxd4 {+1.84/33 19} 39. cxd4 {-0.19/29 12} Rxh3 {+1.94/34 19}
40. Ng2 {-0.84/31 43} Ra3 {+2.10/30 15} 41. Ne8 Rh8 {+2.15/29 9} 42. Kg1
{-0.85/28 9} h3 {+2.17/30 8} 43. Ne1 {-0.79/28 9} Kg6 {+2.20/30 6} 44. Nd6
{-0.68/28 11} Rb8 {+2.46/31 6} 45. f5+ {-0.33/25 10} Kh6 {+2.59/31 6} 46.
Ne8 {-0.33/27 6} Kg5 {+2.66/28 7} 47. Nd6 {-0.39/28 6} Rbb3 {+2.98/33 20}
48. Nc4 {-1.88/27 51} Rxa4 {+2.98/28 5} 49. Nd2 {-1.75/26 22} Rb8
{+3.11/30 6} 50. Rf2 {-1.42/27 5} Kg4 {+4.26/29 6} 51. Ne4 {-2.28/28 21}
Ra1 {+5.68/35 17} 52. Nxf6+ {-3.67/28 30} Kh4 {+6.31/35 8} 53. Rf1
{-3.88/31 4} Rbb1 {+6.69/29 7} 54. Nf3+ {-4.15/32 4} Kg3 {+6.78/22 0.6} 55.
Ne4+ {-4.15/33 4} Kf4 {+7.19/27 7} 56. Rxb1 {-4.29/34 5} Rxb1+ {+7.58/27 8}
57. Kh2 {-4.84/34 8} Rb2+ {+7.76/25 7} 58. Ned2 {-4.94/34 9} Kxf5
{+8.12/25 7} 59. Kxh3 {-4.99/33 5} Kf4 {+8.43/25 10} 60. Nc4 {-6.58/36 7}
Rc2 {+9.75/27 10} 61. Nfe5 {-7.08/35 39} a4 {+34.42/25 16} 62. Ng6+
{-11.82/28 19} Ke4 {+34.63/26 11} 63. Nd6+ {-12.18/29 4} Kxd4
{+35.92/28 19} 64. Nf5+ {-12.22/29 6} Kc3 {+96.05/32 16}
{White resigns} 0-1[/pgn]
Please make sure you configure the engines correctly, because the thinking times look a bit strange (perhaps ponder mode)?