Taimanov Sicilian Trap OTB (it worked?!)

So I was playing an OTB tournament yesterday and on the board next to me, a 1300 sprung an awful trap on their 1200 opponent. According to the Lichess database, 2500 games were played where White fell for the trap, and at least a few hundred more via transpositions (in the OTB game, White played 5. c4 instead of 5. Nc3 and still fell into the trap).

Looking deeper into this - out of 2500 games in total, Black only took the knight 1400 times (!). Somehow, White still has a 25% (!) winrate! I'm pointing this out because there are cases when Black never capitalizes on his free piece! Here is a game where a 2200 rapid player misses it completely!

Note that the trap doesn't work in variations with an early Bb4, since Black's knight is not on c6 and cannot take on d4. Here is a game where that happened, where both players saw the potential trap and White wisely avoided it.

Now let me show you some notable games where the trap was sprung.

IM falls into the trap in Bullet

Bullet game with a very weird move order, trap still works

2200 walks into the trap in OTB

2500 blitz misses it first try, but his opponent doesn't see it either

An FM gets trapped, opponent misses it and loses

The same FM got trapped again, and his opponent saw it

In the Chess.com database, it says that black rarely won the queen. I don't understand.

In the chess.com master game database, it looks like white rarely played 8. f4.

/ Perhaps it is a transposition error.

Pretty sure this stuff is position-based (move order doesn't matter), so transpostions? Chess.com has a much smaller database than FICS or Lichess.

Also, I think you'll be hard-pressed to find a sober 2200+ that falls into this OTB. My condolences to anyone who does.