Okay, make a chess computer that follows a strict algorithm but has no random variables. Whatever move I make, it works its algorithm and has exactly one choice to make. If two moves are exactly equally strong, it will take the first one in alphabetical order, if that’s what its algorithm says, or the last one, whatever its algorithm says. Now, the algorithm can be super-human good and the challenge is to beat it.
The user could be offered a random one of many algorithms or might select one and play it until she can beat it.
We may finally learn whether one side has a forced mate from start, or not. Like freecell 11982, there may be chess computer algorithm 11982 which is impossible to beat.
The GO community thought it would take at least a decade, or never, for computers to crack the game. Their world champion, like Kasparov, went down to defeat in a match, 4-1. I can’t imagine anyone in the chess world wanting to take on the DeepMind team if they turn their chess algorithms on again and expect a different result. As for finding the ultimate truth in chess, good luck. It will still take thousands of years to go through all the permutations. No one will commit that level of resources to create the ultimate tablebase. Besides, everyone knows that 1.d4 is the best first move. Or is it 1. Nc3. Hmm, I will have to ask the robot who lives next door which is better.
Note: unlike chess, Go doesn’t have a professional World Championship, only one for amateurs. The Go player who took on AlphaGo was a very strong player who had won 21 international tournaments, but not a World Champion.
