<15.Qg3?!> White continues along dubious tactical lines. He perceives a tactical softness on the a3-f8 diagonal and thus invests the c-pawn in hopes to collect usurous interest.

<19.Be7> This is leimotive of White recent play.

<22...c5!> This is small print of the tactics here. White would have had a better play with draw in hand after 22...Nxb2 23.Rb3 Na4 24.Rxb7 Be8 25.Nb3... Of course, White can not play 23.Rxd3?? Bc5 24.Ke2 Rd8 after the c5 move in the game.

In fact, I did look at the tactics in this game tad superficially. What I wanted to see was whether once again I could detect certain features of how Morphy built his games.

Since Morphy created so outright beautiful games with such a great frequency, one starts to suspect that Morphy played more than just strong moves in each position to win his games : Morphy's games are so good that they appear composed!

But how do you do that when there is a living oponent in front of you doing half of the moves? That guy is not predisposed to cooperate with artistic intentions where he comes up the short end. My hypothesis is that Morphy, to a large degree, anticipated his oponent's reactions ahead of time.

I suspect that Morphy, like Mozart in music, had in his head stored a large collection of tactical strikes and finishes, some fantastic, some just clever. Then, he liked to know his oponent. Marache for instance, was a decent tactician and Morphy probably knew that Marache would opt to follow forcing lines whenever given a chance. (In fact, that probably was a safe assumption about most Morphy's stock oponents; such was the era.) So, like in N Marache vs Morphy, 1857, Morphy again offers Marache a forcing play oportunity against a on-surface-weak diagonal.

At the end of the forcing lines, beyond Marache's computing horizont, there awaits the Morphy's c6-c5 sting. All in all, not as fabulous a game as the other Marache-Morphy encounter, but few are. Still a good day in the office.

Here is a situation, pawn and move down, against a decent player where normal strategy is bad.

1 e4 e5 2 Qh5+
...c5 2 Qh5+
...d5 2 Qh5+

Morphy's careful moves 1...e6 2...d5 alternate between offense and defense, just barely eluding any tactical blow by White.

By the fourth move, Shredder estimates that White's initial advantage has been reduced from +1.55 to +1.22

Morphy continues to lunge and parry, like the hypnotized character Danny Kaye plays in The Court Jester, until White becomes confused, and then excited and and moves in for the kill (Bxh7+).

Tamar is a great friend and he knows that I do not use computing engines, but that I am at times curious how our silicone brethern would evaluate this or that position. In this game, not only is Morphy playing Black pawn down, but the missing pawn is his f-pawn, which seems especially dangerous right from the open.

Shreder calls White initial advantage +1.55 'pawns', which means a fairly clear winning edge. Somewhere at 1.2 and above, I think, is the fairly clear win zone for the side with the advantage.

Now Tamar points out a remarkable thing : while Morphy plays the position that is effectively lost, he choses risky, 'knife-edge' lines courting a disaster. In the parlor of today game theory, one would say that Morphy initially plays a "high-variance strategy". The logic of it is impecable, however. If the game goes from the initial +1.55 evaluation to, say, +2.55 or +3.55 evaluation, well then the game just remains lost. But if the game goes in the other direction to, say, 0.55 or -0.55 evaluation, then the game is a suddenly more or less even. Since Morphy either explicitly or intuitively knew that, as soon as this game indeed became roughly even, Morphy switches from the high-variance to a low-variance style of play. From this point on, he methodically overplayed Marache, slowly increasing Black's advantage one maneuver at the time.

--

I should have let it run longer, but after 10 minutes, the least bad move it found for Black was 1...e6.

I for one would not have thought ...e6 promised much, but that is normal game thinking; when playing at big odds, very few moves do not just lose. 1...Nc6 is similarly evaluated with ...e6 a move or so later. The other moves veer dangerously close to minus 2.00

That tells me Morphy had delved deeply into what we now call game theory, and was able to steer calmly into complications that emphasized the open f file, the hidden advantage of this variant.

Let's try a small variation on the line I suggested, moving the h-♙ first: 38. h4. If 38...Ke6, now white plays 39. Kd3.

A) 39...g6 40. Kc4. Now black cannot break through. For example, 40...Ke5 41. Kd3 Kf4 42. Ke2 c4 43. Kd2 c3+ 44. Kd3. If black ever takes the f-♙, white plays h5 and promotes first.

B) 39...Ke5. Now white can again play 40. Kc4. 40...g6 transposes into the previous variation. If 40...Kf5, then the continuation might be 41. Kd3 Kf4 42. Ke2 c5 43. Kd2 Kxf3 44. h5 Ke4 45. h6 gxh6 46. gxh6, and white again queens first. 39...Kf5 leads to a similar result.

C) 39...Kd5 40. h5 Ke6 (black must retreat his ♔ to stop white from promoting the h-♙) 41. f4.

If black answers 38. h4 with 38...Kg6, then 39. f4 takes us back to the variation I gave in my previous post.

if Morphy is playing to win, then the rook exchange needs to be after King centralization. This seems drawish, 33...Rc2+ 34.Rxc2 Bxc2 35.Ne5 Be4 36.h4 Bd5 37.g4 c4. Instead, 33...Kf7 34.h4 Rc2+ 35.Rxc2 Bxc2 36.g4 Ke6 37.h4 Kd5 is attractive.

but until 36.Kd2, the game is a draw.
Since Marache is a strong player of the time, maybe the problem is not him, but the state-of-the-art at the time. I can rationalize 36.Kd2 by :

WHT needs to stop the connected passed pawns
WHT needs an active King
a B-N exchange would remove the doubled pawn structure.

To this amateur, this game shows alot of endgame theory. At this stage of the game, BLKs plan is counter-intuitive. Instead of advancing the two connected passed pawns, use them as a threat instead. BLKs plan is to promote one of the other pawns! (by using its K to gobble the remaing WHT pawns). This plan is facilitated if WHT has as few resources as possible, so a K+P endgame is desired.