Just one annoying feature of the current solution algorithm is the possibility of failing with a move that wins just as easily as the solution, for example I failed this problem: http://en.lichess.org/training/3068 by playing Be1 on move 2, which is -8.71 compared to the solution's -10.29 according to HIARCs.
I'm not too bothered by failing by playing a mate in 2 rather than the mate in 1, but for non mating problems it's a different matter.
I'm not too bothered by failing by playing a mate in 2 rather than the mate in 1, but for non mating problems it's a different matter.