John: Another variation also seems to be wrong: G1F1E1H1C1E3H2F1A6A3. This is double ko (A4 and G1), defender is dead. A3 should be at E4, leading to a direct ko at G1.
caranthir: Looks like it's been fixed according to your proposal.
John: I agree with heron, there is something wrong with variations here. As he mentioned, E3 cannot be played - attacker plays A4, finishing the ko, and defender is dead. Hence counting E1 as an ignored ko threat, this yields A+3KO, isn't it?
The variation G1F1D1H1A6A3A4 also seems to be wrong, due to the same reason. Moreover, it seems that defender should play C1 instead of H1, as in the D1C1 variation.
John: Finally I got the solution. I changed the status of the branch after A3A4 to A+2KO+2, since defender needs one more ko threat. The variation in my next comment has also been corrected. Please, recheck.
heron: Yea, the sequence I said below seems like it should end a move earlier? Because when the refutation plays E17 it should instead play a ko threat, or else they die.
John: Defender responses were corrected. The sequence you provided is not the optimal one.
heron: Something seems wrong with this one to me:
I played G19F19E19H19H18F19C19G19A14A17A16E17A15 and it says "incorrect", but defender looks totally dead, not even ko. I'm not sure my sequence is very good but there must be some error in the solution.
YuriyStepanovich: Bg1-Wf1-Bd1-Wh1-Ba6-Wa3-Ba4-We3-Ba5 kills the White, but message is "Incorrect". Why?
Ivan Detkov: This is incoherent. Could you please use capital letters and avoid colors (B and W). Check the answer of posetcay down below for an example.
kovarex: This problem is now the prototype as a status-based problem.
This means, that the branches are not marked as correct or incorrect by + comments, but instead, there is specification of the status, and the applet automaticallly assigns the correct branch based on the best possible status.
This is to show that the correct variation is A+3KO (3 approach move ko), while some other, similar variations are not good enough, as they might be only A+2KO (2 approach move ko)
There are still tweaks to be done (mainly to show the full-text status explanation in the review mode), but it seems to work nicely as a prototype.
d4rkm4tter: Very good. I want to make a special topic collection with this.
posetcay: The solution is correct, but opponent's refutations were wrong in many places. For example, R1-Q1-N1 should be answered by M1 and R1-Q1-T2 should be answered by P3 etc. Made the necessary corrections.
Looks like it's been fixed according to your proposal.