An evaluation of +0.9 from an engine just means almost the value of a pawn at the start of a game, which doesn't mean "completely won". If anything it means still a draw with perfect play from both sides, like gambits often tend to tragically end up as. While it might be impractical for humans to respond to all the threats properly, perfect play probably tends to a draw.  

Despite the saying "a beginner can win with an extra queen, an amateur can win with an extra piece, a master with an extra pawn".... my guess is at GM level with no mistakes it's still probably a draw.

RdC wrote on 11/03/16 at 15:59:52:


I have to agree, Black's b-pawns look very weak.

Jupp53 wrote on 11/03/16 at 14:06:06:



I think it was between two postal players, Robert J Wakefield and Brian Thompson in what is described as Reserves A of the British Championship 2006/07. 

I doubt that game is the final word, as the idea of making e5 work looks like the way for White to make progress. 

In the ending 27. Rd7, rather than 27. Ne2 looks an obvious alternative. The Knight can stay on c1 for the time being. It shelters the King and if it stays on a defended square avoids ceding initiative  by being hit.

This endgame is important lol. If this is a draw then Qa5 is not busted.
Seems like Black is under pressure, but I agree with you, it is not completely won.
Seems like Im gonna have to find another line against Qa5.

This endgame arises after the supposedly refuted Qa5 dragon. It seems like White is obviously better and the engines give him a pretty good evaluation (about 0.90+) however the only endgame I could find was drawn which makes me really skeptical about the engine's evaluation.
However I am also skeptical about the way White played, he weakened c3 and allowed Black to exchange dark square bishops after which his problems seem to be gone.
Here is the position.



And here is how White played.