S_P_A_R_T wrote:
Here's a really strange WD v7.02 self play test.

This was to mainly ensure that the new table-based move gen for the pawns worked (along with some other new stuff), but this game it just terrible!

The queen get's trapped twice (and neither time does the other side realise this) and after this, the rook vs rook ending was just terrible! I stopped this game short even though black was winning because it didn't look like progress was going to occur, and even if a decisive result happened, it wouldn't really be that “important”.

https://lichess.org/jN4nYBid#107

The queen trap needs a depth of 8 to find (14. e5! Qf4 15. Ne2 Qg4 16. h3 Qh5 17. Nf4 or Ng3 plus black’s next move makes 8 moves plus quiescence.

I replayed WD’s rook endgame blunders on Element :
44. Kg2?? , Element plays 44. Ke3 (Element reaches depth 7, the depth 6 move Ra6 is also acceptable).
44… g5?? , Element plays Kd6 (reaches depth 7, depth 6 move is Ke6??)
47. Ra6?? , Element plays Kf3?? Reaches depth 7)

I don’t recommend terminating games too often, especially in late endgames. For example, https://lichess.org/study/I02Nwcbd/Pdo2EVrU#329 (ignoring 50 move rule here)

White Dove 7.01 vs Element v1.4862 - Draw in 81 moves. Around the move 48… h5, Element was showing -2 and WD -3 (black is winning). I’m surprised Element picked this up. The B+2P v R exchange didn’t help WD either.

The stalemate situation is similar to Element’s 5+8 match against Bonsai in the SCF Cup (Element only has stalemate detection in the root nodes).