Positionally, Black is unreasonably fearful of having a two-pawn center and as a result does not achieve an optimal set-up in the middlegame. I had worried that the pawn center could be too easily undermined by White, but analysis shows this was not the case. This judgment error is a good example of how an uncritical preference for a type of position (or against one) can lead to less effective play.
Around move 30 my calculation and judgment began declining rapidly when Black was faced with a menacing White pawn mass on the queenside. At least I had a bailout into a drawn position, rather than fully collapsing into a loss. "Turk" had played a typical computer handicap move on move 20 to give me the advantage, so I guess I just ended up returning the favor.
The training game was useful for highlighting the individual calculation and judgment errors mentioned above in the opening, middlegame and endgame phases, so was successful from that perspective.