Draw by repetition - NOT found

I analyzed an over the board game I played on 05/04/2019. here is a link to the study I created:


I thought the knight sacrifice on g6 would win, but it is not the best move. When analyzing this in Stockfish, it gives equal value to both 5.Qh6+ and 5.Qxe6 (also at move 7 and move 9). If you keep playing Qh6+, there is a draw by threefold repetition. But this is not flagged as a draw by Stockfish.

I expect Stockfish to give a higher evaluation to the line of play with Qxe6, since that is a forced win.

Today I tried line using Chrome Browser (instead of Firefox) in Windows 7. This found a different line of play: 8.Rc3. This still does NOT give higher evaluation to line without threefold repetition.

It evaluates to +62.3 at move 18, then 0.0 at move 19, then +62.3 at move 20. Maybe this means that Stockfish recognized the draw potential at move 19?

5 Qh6+ still seems a fairly easy win, how is it a draw by threefold repetition? Don't you just go Rc6-c3-g3 and black is busted? I don't get it.

Agree that there is a simple win, but Stockfish doesn't see it on my computer. Even after Rc6-c3-g3, Stockfish prefers giving endless checks to Black's King instead of taking Black's Queen.

OK, I had a Linux computer handy, so I tried running the same study under Firefox. Now Stockfish plays MUCH more intelligently under Linux than it does under Windows.

Well, here Stockfish is basically a script that runs in your browser, so it may be derpy sometimes. Try giving it more time to think, it probably would figure out the win eventually.

Based upon the Universal Chess Interface (UCI) specification, Stockfish can only print a move/variation and a score.
"Claim Draw" is not a move.