I was black, and I will list where I blundered. How difficult do you find these positions? What rating ELO would you expect someone to play accurately here?
- 37 ...b4 and I should have played Bd5+.
- 38 ...h6 and I should have played Bd5+ still.
- 39 ...b3 and I should have played Bxa6.
The post-game analysis can sometime frame a specific move as the best, most accurate, move while in reality there are multiple playable moves.
In this position it doesn't really matter what you play as long as you are aware that Nb6+ could happen. b4 was a blunder because now black can play Nb6+. If you take the Knight (axb6), the white pawn can promote. If you don't and just move the King, you will lose the Bishop.
Because you already played b4, white can play Nb6+ and after 38. g6 is also threatening to play the g-pawn. Multiple moves can prevent the pawn from promoting while saving the Bishop at the same time.
In summary, it's less important to play Bd5+ than it is to prevent fork Nxb6+.
Now, regarding your question. It's not easy to answer, but here is my guess. I would expect any 1000 rated player to see the threat more than half of the time. A 1500 rated player should see it, but can still make the same blunder some of the time.
I was black, and I will list where I blundered. How difficult do you find these positions? What rating ELO would you expect someone to play accurately here?