A Question on Board Position Repeats

Greetings folks, I have a question for the rules gurus:

The no-repetition rule of Go states: “One may not play a move that repeats a previous board position”.
If we examine the situation of a 1x1 board on OGS with New Zealand rules:

  1. Black plays the first move, it is captured because it has no liberties, board state is empty.
  2. White plays the second move, it is also captured for the same reason, board state is still empty.
  3. Black plays the third move, it works and is a suicide, board state empty
  4. White plays the fourth move, and this time it’s prohibited because it is an board position repeat.

I would have thought that the no-repetition rule would’ve prevented white from playing the suicidal move at 2, since that would clearly be a repeat of 1. How come three moves were allowed before white was prohibited to play at 4?

Is this a part of the game rule specification, or is it a quirk in the way OGS implements board position repeat?

Thanks in advance!

I dare not speak towards how it “should” be, but am pretty sure it was not conscious implementation from OGS part, thus more likely a quirk.

Our engine seems not to interpret kos at all until 4th move, which can sometimes mess up forked games if there was a ko present before the fork. In a “normal” game there can never be a ko that soon, but yeah in these special cases, it does not work that well…