Weaker player looks at this move, scratches head, and starts up computer then, after finding big advantage black, adds punctuation and tells of how foolish white was. Pseudo-master looks in big library starting with MCO bible or the likes, searching for line to prove what "any master surely knows" (when he does not find, no worry just pontificate or make something up, or threaten to crack using engine LOL).

Strong player understands character of position and motive behind move, realizing subtlety to 14.Ng5! as reflection of David Bronstein's genius. Computer can give lines and alternatives, with careful scrutiny one can make good assessment of risk versus potential reward, but so far no-one has created engine that gives "difficulty to survive in" score. Like small group of other attacking geniuses, Bronstein was comfortable in such super sharp positions especially if he created it, most GMs are not, especially in line altogether new, as 7.h4 gave. 14.Ng5! creates higher level accuracy test, like Naka game a few small inaccuracies can become disaster.

Sure enough, 15...Nbd7 instead of 15...Nc6 (or 15...Bh5, in best line they transpose) lost nearly all advantage, after 17...e5 slightly loosening and position was already equal. But then seemingly innocuous 18...Qc5 (instead of 18...Rac8 or 18...Ng4) was all it took for Bronstein to get big edge. 19.Be3! blocked check and attacked Q, tempo gained (e.g., 19...Qa5 20.Be2) in this very sharp position are worth more than pawn/point. 19...Qc6 played was probably best (allow computer few hours before arguing, sharper positions generally take longer) with advantage already nearly decisive.