Practicing the Method of Comparison with engines.

We look at an unusual piece up position.

The strange move h5 is suggested by the engine, so I show how it is I try to determine "why it is good" in a systematic fashion that I will refer to as Method of Comparison.

Here I first ask, what would white do, given an extra move? This can be done with either the threat toggle that appears when the engine is on/ with the shortcut (X) or by taking the FEN directly and switching the turn to move by changing the b that indicates that it is black to move to a w.

Turning on the engine in this position, we find that Be3 is evaluated at +1.3 and Bf4 is evaluated at +0.4. This means that somehow h5 works to make these moves bad. Now we can go to position A and see the response after h5 Be3. Here the engine gives only the move Nc4 for -0.8. To me it is not apparent how the inclusion of h5 has helped, so now I check position B after Be3 Nc4. We see Bd4 is +1.3. Back to position A. h5 Be3 Nc4 Bd4 Rh6 and it looks like h5 has helped. However, it seems odd to play Rh6, so we can again use the same method, giving white an extra move.

We see that b3 is given +3.3. Such a large evaluation often indicates that one side is losing a piece. We ask why is moving the knight away with Na5 or Nd6 not possible, and we see that Bb6 is a strong response to both of these. Returning to position A, we see after h5 Be3 Nc4 Bd4 Rh6 b3, Rh6 has enabled Nb6. It is possible to continue looking but at this point I judge that it's appropriate to examine white's other somewhat more human move of Bf4.

When doing the following, where you switch between your own positions A and B, it is helpful as separate chapters in a study, open in two different tabs with one of the tabs desynced, so that it is convenient to quickly switch between the positions.
You can access the study by going to one of the boards, using the menu option and selecting View on Lichess.

From position B, Bf4 is +0.4, so h5 is preventative. Position A, h5 Bf4 Rc8 -0.7. Position B Bf4 Rc8 a4 +0.2. Position A h5 Bf4 Rc8 a4 b4 -0.8. Position B Bf4 Rc8 a4 b4 Rc1 +1.5. Position A h5 Bf4 Rc8 a4 b4 Rc1 Rxc1 -0.7. B ... Qxc1 +2.7. A ... d6 -0.8. B ... Nc7+ +3.2 A ...Kd7 -1.4. B Ne5+ +3.3. A dxe5 -4.4. B Bxe5 +4.8 A Rh6! -4.6 We see that now 7 moves down the line, h5 has allowed black to play Rh6, enabling them to interpose Bd6 against Rd1+, where the material advantage should be decisive.

So we can somewhat conclude (as there may be other points that this analysis has not covered) that from the initial position (position A) h5 was a helpful move as it enables a defensive Rh6 manoeuvre in some critical lines.

This process in general does not need the engine, but I think this is a helpful example showing how it might be applied when using it.

For those curious, the wild position we were looking at came from the Smith-Morra Gambit.