Stockfish 12 has arrived!

[OliverBr]

If engine is not leading after 200 games it's not significantly better. BTW latest SF devs before NNUE were +25-30 ELO ahead of SF11 after 200 games always.

This depends what you mean with "significantly better". For an ELO difference of 25 200 games are not enough. By a far cry.
I had tests where after 900 games the difference was +36 ELO, but after 4000 games only +12 ELo.

I personally do believe anything from a test of under 2500 games. With 32-core this is good possible, just a couple of hours patience.

If you want I could test the 2 Stockfish Versions.

[Jouni]

[d]8/8/p2k1p2/1p1p3p/1P1P3p/P3NPP1/5K2/1b6 w - - 0 1

Analysis by Stockfish 12:

1.Ng2 hxg3+ 2.Kxg3 Bg6 3.Nf4 Be8 4.Kh4 Bf7
+- (3.61) Depth: 7/8 00:00:00 1kN
1.Ng2 hxg3+ 2.Kxg3 Bg6 3.Nf4 Be8 4.Kh4 Bf7
+- (3.61) Depth: 8/8 00:00:00 1kN
...
1.Ng2 hxg3+ 2.Kxg3 Ke7 3.Nf4 Bf5 4.Kh4 Bd7 5.Kxh5 Kf7 6.Kh6 Bc8 7.Ng2 Bf5 8.Ne3 Ke6 9.Kh5 Bh7 10.Kg4 Kd6 11.Nf5+ Kd7 12.Ng7 Bc2 13.Nf5 Bb3 14.Ne3 Ke7 15.Ng2 Bc2 16.Kg3 Bb3 17.Nf4 Bc4 18.Kg4 Kf7 19.Kh5 Bf1 20.Nxd5 Ke6 21.Ne3 Be2 22.Kg4 Bd3 23.Kf4 Kd6 24.Ng4 Ke6 25.Nf2 Bg6 26.Ne4 Kd5 27.Ke3 f5 28.Nc5
+- (3.84) Depth: 44/64 00:00:10 73460kN, tb=226854
1.Ng2 hxg3+
+- (3.76 --) Depth: 45/68 00:00:13 88559kN, tb=280576
1.Ng2
+- (3.84 ++) Depth: 45/68 00:00:13 90947kN, tb=296057
1.Ng2
+- (3.97 ++) Depth: 45/68 00:00:17 116mN, tb=391097
1.Ng2
+- (4.14 ++) Depth: 45/68 00:00:22 148mN, tb=474597

No other moves considered at all . Stunning.

[kasinp]

[d]8/8/p2k1p2/1p1p3p/1P1P3p/P3NPP1/5K2/1b6 w - - 0 1

Analysis by Stockfish 12:

1.Ng2 hxg3+ 2.Kxg3 Bg6 3.Nf4 Be8 4.Kh4 Bf7
+- (3.61) Depth: 7/8 00:00:00 1kN
1.Ng2 hxg3+ 2.Kxg3 Bg6 3.Nf4 Be8 4.Kh4 Bf7
+- (3.61) Depth: 8/8 00:00:00 1kN
...
1.Ng2 hxg3+ 2.Kxg3 Ke7 3.Nf4 Bf5 4.Kh4 Bd7 5.Kxh5 Kf7 6.Kh6 Bc8 7.Ng2 Bf5 8.Ne3 Ke6 9.Kh5 Bh7 10.Kg4 Kd6 11.Nf5+ Kd7 12.Ng7 Bc2 13.Nf5 Bb3 14.Ne3 Ke7 15.Ng2 Bc2 16.Kg3 Bb3 17.Nf4 Bc4 18.Kg4 Kf7 19.Kh5 Bf1 20.Nxd5 Ke6 21.Ne3 Be2 22.Kg4 Bd3 23.Kf4 Kd6 24.Ng4 Ke6 25.Nf2 Bg6 26.Ne4 Kd5 27.Ke3 f5 28.Nc5
+- (3.84) Depth: 44/64 00:00:10 73460kN, tb=226854
1.Ng2 hxg3+
+- (3.76 --) Depth: 45/68 00:00:13 88559kN, tb=280576
1.Ng2
+- (3.84 ++) Depth: 45/68 00:00:13 90947kN, tb=296057
1.Ng2
+- (3.97 ++) Depth: 45/68 00:00:17 116mN, tb=391097
1.Ng2
+- (4.14 ++) Depth: 45/68 00:00:22 148mN, tb=474597

No other moves considered at all . Stunning.

Especially considering that it reportedly took overnight analysis by Karpov's team to come up with it!