Interesting engine test

Earlier this year I played an interesting game where I sacrificed an exchange for a pawn with my pawns strong in the center but my white squares weak and opponent having two dangerous bishops aiming my king. Here it is:

I kept thinking about this position and finally I decided to have a small engine blitz tournament starting from it, Stockfish against Houdini (5 games with each colour, 4+1). The results surprised me. 

I started checking out the games to see if something was wrong and I found out that the engines each time chose completely different way of solving problems with black - Stockfish usually gave back the exchange eliminating strong white pawns in the center as compensation for his missing pawn obtaining initiative due to pair of bishops while Houdini wasn't so radical playing slowly sticking to the material allowing Stockfish (with white) to regroup and create pressure with his central pawns. Some games were quite nice! Hence, I wanted to share this.

Let's first check out Stockfish's approach with black.

And now see what happened with Houdini NOT returning the exchange.

Not bad! Another one:

I like the next one:

And the last one is my favourite - in a position without exchange Stockfish sacrifices another one!