about making derivatives of engines with a different move choice

[Uri Blass]

My idea is to use the original engine and after half of the time search for second best move and making a different move only if the analysis suggest that second best move is not bad.

For example you can decide to change your mind if you get the same score for the second best move or not worse than 1% worse than it and it means that for evaluation of 0 you change your mind only if you find a different move with evaluation of 0 or more and for a move with evaluation of 2 pawns advantage you change your mind only if you find a different move with at least 1.98 pawns advantage.

I wonder how many rating points you lose if you do it for stockfish and what is going to be the similiarity relative to original stockfish.