Is this winning or a draw for white?

This should be winning for white. [maybe not, see edit]

in endgames, the principle of 2 weaknesses usually shines, which is the concept of having multiple weaknesses across the board. In this case, Black has a severely weak g7 pawn which will be picked up soon, and White will then have a far advanced outside passed pawn. Based on those factors alone White should be winning, and on top of that Blacks bishop is very passive and can’t defend.

EDIT: the outside passed pawn is indeed just one weakness and after looking at this position some more, it’s hard to capitalize on it. Black lacks a second weakness, notably, and meanwhile his bishop can defend his central pawns while his king and rook keep the outside pawn at bay. it’s hard to say whether or not zugzwang is possible but probably not given the range of moves available to the rook and king.

hope this helps!

Stockfish stacked with Static evaluation +1.94 at 1 billion nodes in my phone at depth 53 ( 3 mins search) with 5 men EGTB .

Evaluation not increasing!

Looks like 95-98% chance( very high chance of draw/ fortress)

I am pretty sure that Stockfish 8 did not have access to Tablebases against A0. A0 games were controversial anyway ( many suboptimal conditions of Stockfish 8).

@drmrboss If the lichess database doesn't work with this position, how have you found the database on droidchess? The endgame databases are the same...

Heh, after the preprint was released on arxiv, it became popular to whine about the conditions for SF.

People continued to whine even after the full paper was released, showing that those complaints no longer applied. I suppose this is because people didn't bother to read the paper and just repeated stuff they were told by others (which was in turn just stuff they heard "somewhere").

So much of the nonsense that's been repeated about the matches could have been prevented if people would bother to read :)

Direct quote from the Materials and Methods information in the supplement at

"To evaluate performance in chess, we used Stockfish version 8 (official Linux release) as a baseline program.

Stockfish was configured according to its 2016 TCEC world championship superfinal settings:

44 threads on 44 cores (two 2.2GHz Intel Xeon Broadwell CPUs with 22 cores), a hash size of 32GB, syzygy endgame tablebases, at 3 hour time controls with 15 additional seconds per move.

We also evaluated against the most recent version, Stockfish 9 (just released at time of writing), using the same configuration."