Hmm. No master would play 3..Bf5.
If you look a move earlier, you can see that is true for all practical purposes.
/ I think there are undoubtedly transposition errors in the database. (Where 3 e5 would have allowed 3..Bf5 and 4 Nc3.)
And the second game said it was from a blitz championship.
/ So yes, that last game should not be in the database. I don't know where they got the database from.
Hmm. No master would play 3..Bf5.
/ I think there are undoubtedly transposition errors in the database.
I would have said the same thing but if you look at those two games (particularly the 1st: Abasheev Denis (2419) vs Riazantsev Alexander (2596) - Sochi, 2004), the remaining moves are consistent with that move: e.g. 4...Bg6 (3...Bf5 is the only move that allows that). It looks like a bullet game and white didn't notice. I don't think that's a transposition error. I'm sure there are some but 1,700 for one move!!!?
And if they are transposition errors then the blunder on move 3 would be an indication of that and suggest masters games with blunders in the 1st 10 moves shouldn't be in the database (whether they're transposition errors, bullet games with mouseslips and premoves, masters throwing games deliberarely, or whatever)
Explorer is not move dependent but position dependent. So, when you reach a position and the options give a move, it does not mean the games there were reached by that move order but the piece being on that square, and all other pieces on their squares, is in the database.
Most database programs work on a similar premise. What matters is the reached position and transpositions of moves can result in reaching it.
I understand The two games where 3...Bf5 were [reported to have been played] are aberations (that should be purged in my opinion).
The 1700 moves are moves that would result in a transposed position (different move order where black never hung the bishop).
However, I think the way the explorer works is bad in that case: it's transposing a move too early. It should still only show the 2 games (and they shouldn't be in there) but if someone makes another move it will then be into the 1700 games from that position (with black to move). This position with white to move was only reached in 2 games and there were only 2 moves played (of course) from that position (both blunders that shouldn't be in an opening explorer).
I noticed this line in the Caro-Kann the other day:
https://www.chess.com/explorer?moveList=e4+c6+d4+d5+Nc3+Bf5&ply=6&origMoves=e4+c6+d4+d5+Nc3+Bf5
As you can see, apparently over 1,700 games with that blunder on move 3.
Looking at the 2 last examples, they are terrible games:
https://www.chess.com/analysis/game/master/1233451
Black resigned on move 8 for some reason (white never capitalised on any of black's blunders)
https://www.chess.com/analysis/game/master/13974947
Black resigned on move 21
So, my questions:
Shouldn't there be some criteria for a game qualifying for entry into the opening explorer? (Both players blundering on move 3 and 4 is not a great example of an opening.)
And if there's 1,700 games where 3. Bf5 was played why isn't there a single example of someone capturing the bishop (the obvious best move for white)?
Are these all bullet games where white premoved? If so, should such games be in the opening explorer?
What value the opening explorer (and especially its win/lose ratios) if there isn't some minimum quality to the games?