TalkChess.com

[Joona Kiiski]

“Can we say that V is more accurate than H respectively?”

It depends on the domain. In chess when you search deep enough, sooner or later you end up in a silent position where all tactical complications are resolved and only material matters. That's why it's generally true to say that V is much more accurate than H. Of course, there are exceptions, for example fortress positions.

However for example in game of Go, where tactical complications can stay on the board over 100 moves and writing a good evaluation function is almost impossible, searching deeper with minimax doesn't necessarily help (or the gain is minimalistic).

I think I saw a paper long time ago which demonstrated games where searching deeper made a program weaker, but I'm not sure about this.

Btw, One funny result: Around 1-2 years ago, I tried to tune Stockfish's evaluation function to make delta(V-H) smaller on average, but this change made SF weaker!!
_________________
Joona Kiiski