Platform for testing engines

[MOBMAT]

I currently use Arena to test engines, but it lacks the feature I need.
It can detect when an engine finds the "best" move, but I want to also qualify the move when it detects mate.
Can I get a recommendation of a app that could handle that type of condition?

V

[phhnguyen]

Look like you need an analysis or a coach feature. Both features will use engines to run to evaluate move by move and give some advice. Banksia GUI has both features.

Below is the coach feature of Banksia GUI. It could run auto when a move is made. Users could re-coach for existing games too. When the gap between the score of a move and the score of the coach's move is over a threshold, the coach will show a warning and both PVs. It could show a graph of those gaps too (blunder graph).

[chrisw]

Look like you need an analysis or a coach feature. Both features will use engines to run to evaluate move by move and give some advice. Banksia GUI has both features.

Below is the coach feature of Banksia GUI. It could run auto when a move is made. Users could re-coach for existing games too. When the gap between the score of a move and the score of the coach's move is over a threshold, the coach will show a warning and both PVs. It could show a graph of those gaps too (blunder graph).

Can your platform host and run online UCI chess engine tournaments?

[MOBMAT]

That isn't my intent.
When determining if an engine solves a position, it sometimes finds the correct move but then changes its mind.
When I know the solution is a mate, then I don't want the server to end the test just because the engine found the "bm" near the beginning of the test. Finding the solution requires more evidence, hence the requirement to detecting the "dm" value as well.

I think I saw a solution using Python, but I'm still searching for that code.

V

[phhnguyen]

Can your platform host and run online UCI chess engine tournaments?

No, not yet!

But it is on my to-do list! Perhaps after the first official release of 1.0 (within a month).

[phhnguyen]

That isn't my intent.
When determining if an engine solves a position, it sometimes finds the correct move but then changes its mind.
When I know the solution is a mate, then I don't want the server to end the test just because the engine found the "bm" near the beginning of the test. Finding the solution requires more evidence, hence the requirement to detecting the "dm" value as well.

I think I saw a solution using Python, but I'm still searching for that code.

V

OK, so you mentioned about test suite.

I just write here how BSG works with the test suite for finding bestmoves (bm):

1) whenever BSG finds the engine prints out the best move matched (to the set of best moves for each test), it just flags it and goes to verify by letting the engine continues running for an extra ply (set by users as the below image)
2) if within/after that extra ply, the engine still prints out the same best move, BSG termites that test and the engine is counted as a success
3) if the engine prints out any different best move, the flag is cleared and the test is continued until (1) or timeout