Drifting vs. Maneuvering

A few months ago, in a post titled Learn from Your Engines?, I discovered that I lost a game because I had been slowly outplayed over the course of twenty moves. A comment to the post called this phenomenon 'drifting', where 'the evaluation will slowly drift from equal to lost for one side'. That is indeed what happened in my game.

A few years ago, I started a series of posts on the theme of Elastic Maneuvering. While I never succeeded in defining this concept, it's clear that there is a direct connection between drifting and maneuvering. One side maneuvers while the other drifts until a once equal position is no longer equal.

***

Returning to the theme of Learn from Your Engines?, I went back to my post Learn from Your Losses?, and selected another game for analysis by Houdini. This time the game was Weeks - Miettinen, a Closed Lopez, Worrall Attack, played in the same correspondence event as the game from the Engines? post.

The first diagram shows the position just after Black, who was rated about 100 points above me, offered a draw. I declined the offer, because it was the second (quarterfinal) stage of a multi-stage qualification event and I needed a win to have any chance to qualify for the semifinal stage. As things turned out, I ended up with my worst result ever in a correspondence tournament, finishing +1-4=1, where my only win was a forfeit.

Even if the game hadn't been important for qualification, I would have declined anyway. Most of the pieces are still on the board, and if I'm going to accept a draw in this sort of position, I might as well not play at all. Black's b-Pawn has become detached from rest of the Black Pawns and White has chances to win it. As compensation, Black has the Bishop pair, but the closed center renders them less effective.

The second diagram shows the position where Houdini says I went astray. After some maneuvering White has captured the b-Pawn, but the capturing Knight is caught in an awkward network of pins and potential pins. White's problem is to extricate the piece without exposing the King to attack. After considerable analysis and still unsure what to do, I continued 50.Bh4, underestimating the strength of 50...g5, which gave Black a winning game only five moves later.

Houdini indicates that 50.Kg2 is a better move. Defensively, the King strengthens the White squares on the exposed Kingside and prepares to counter any raids by the Black Queen. Offensively, it prepares a subsequent Qh3 as an offer to swap Queens, thereby removing the most dangerous pin on the b3-Knight. Houdini's move requires more analysis than I can give it for this blog post, but White is not going to end up with a lost position after five more moves.