d4rkm4tter: Your guess is essentially right. The alternative response mode doesn't distinguish between correct, wrong or new (blue) opponent move when on a correct path. So it needs an additional checking if the opponent move is in the tree.
rst: I believe I figured out the way to trigger the bug mentionned by Clé.
The steps to reproduce it are:
- solve the tsumego
- review it and try a variation that is not in the variation tree for the opponent
- click on reset
- try to solve the tsumego again
It will accept your answer but will ask "what about this move?" with the move you added manually in the variation. My guess it that whenever it is on a green/correct subtree, it tries exhaustively all the variations even if they are not green/correct or red/incorrect so it tries your personnal variation.
TwT: Dear rst,
Thank you very much for your efforts in debugging this issue. We truly appreciate your help. However, we are unable to reproduce the bug based on the steps you provided. In BeosGO, there is only one active feature: reducing recurring positions, which is marked with a (-) in the solution tree.
My hypothesis is that if you input the moves in the wrong order, you might inadvertently trigger the "Alternative Response Mode," even without having two solution branches (since you are feeding the buffer with a different order of moves beforehand). If this does not resolve the issue, I would need the following to better understand the problem:
(1) A screenshot of the step where you "try a variation that is not in the variation tree for the opponent."
(2) A screenshot of the step where "It will accept your answer but will ask 'what about this move?'"
(3) Please add as subject of mail: Please forward to TwT
I would be very grateful if you could provide these screenshots for the specific tsumego problem, as it seems to be of minimal complexity.
Clé: Ok i'll try to be more clear about the bug.
So first solve, the sequence goes Q1 - P1 - Q2, and it shows the problem solved.
But after going into the review mode, hitting reset, and solving it again with the same séquence, it triggers a "Correct but what about this answer" sequence where white answers Q1 with Q2. Then, R1 for black is marked as incorrect, though in this séquence it is the only answer possible ?
TwT: We still can not reproduce your observation.
.
Clé: I'll add that this seems to be a bug - the "correct but what if" second part of the tsumego is only triggered when I go into review, play the corresponding séquence, and then reset the tsumego and solve it again
Ivan Detkov: It looks very straightforward to me. I do not see complexities. What is your first move exactly?
I fixed this.