My point is that huge game tree/huge number of configurations means nothing by itself, if the game contains some structure, which can be exploited. There are problems where it is possible to considerably reduce search space. Hardest Rubic cube problem from 10²⁰ to 10^7. Minimum hints of Sudoku, Edge matching problems, real world SAT problems, even stupid TSP has solved instances of 10 of thousands vertices, which is, given naive estimates, many orders of magnitude harder than chess. Only thing that matters, if (and how) would it be possible to use chess rules to build some structure and reduce search tree rapidly. 10⁵⁰ to 10²⁰ or 10²⁵ or similar isn't that far fetched.