This sounds like the precursor to the movie "War Games". The super computer randomly takes targets.....military,....financial.....medical......logistical......and tries to figure out who would win in a neclear attack.
Any way to determine game moves from ending position
Sort:
llama_l wrote:
But like I said, feel free to provide a small example. Pick some position, apply your method, and show us the output.
I have no programming ability so I can't do it myself.
llama_l wrote
To be clear, it "stops" in the sense that your terms become 5^10 + 5^10 + 5^10 etc. This is replacing, for example, 15^10. Finding a forcing move doesn't "stop" it in the sense of removing exponents completely.
And in any case, you can't assume that a human will play the only good move.
Yes,the formula would be
((x/5)(100^5)) where
1. X=the total number of moves in the game
2. We assume 100 possible moves per turn
3. We deal in sets of 5 moves each
So as I said in my post yesterday, "in a 40 move game, it would be around (8(100^5)) total positions to calculate."
If anyone is interested, here was the end position of the game that made me first start this post.
I actually just realized the photo I took of the board had an object obscuring several of the middle squares so I 'm not sure of the exact end position, which is a bit disappointing. but this position is very close, only the two white bishops and knight on e 3 could have been on adjacent squares.
I am not sure how realistically this particular ame can actually be solved, but this conversation has at least turned into an interesting theoretical proof of concept!
Sep 28, 2023
0
#26
Heathcliff37 wrote:
Martin_Stahl wrote:
After that, you can create a plausible game, but the odds of it being the game played are extremely unlikely.
Plausible is enough for a proof of concept.
But it's not going to be the game that was played. Your initial premise was "I played a game of chess that I very much want to analyze.."
Coming up with plausible doesn't give that. It wouldn't be that hard to build a plausible game from an endgame, as mentioned but it doesn't get you anything useful, other than being a thought experiment.
Martin_Stahl wrote:
But it's not going to be the game that was played. Your initial premise was "I played a game of chess that I very much want to analyze.."
Coming up with plausible doesn't give that. It wouldn't be that hard to build a plausible game from an endgame, as mentioned but it doesn't get you anything useful, other than being a thought experiment.
Let's put it into context, I made my first post thinking there already such a program developed and that all someone would do is send me a link to a page where I input my game and it solves my question.
I realized very early on that such a think didn't exist, so my initial inquiry was essentially moot.
But then people were suggesting that the entire idea is absolutely absurd, and the conversation developed into something valuable in and of itself. If it is possible to develop plausible games from an endgame, then my idea is far from absurd. If we can build plausible games, then perhaps we can work towards building useful games, but indeed that is all hypothetical at this point because, as I said, my initial point is already moot.
Furthermore, even after my initial point was obviously moot, I felt obligated to answer everybody's inquiries and it developed into a great intellectual conversation.
Wait a minute. 
It just dawned on me that high level tournament games are all annotated. It may be a good mental exercise to try and reverse engineer it. Get one of those super computers to work it backwards and see if it matches the annotated results. 
RonaldJosephCote wrote:
Wait a minute. It just dawned on me that high level tournament games are all annotated. It may be a good mental exercise to try and reverse engineer it. Get one of those super computers to work it backwards and see if it matches the annotated results.
Exactly, I think it's worth experimenting with.
But trying to do that with no information...
I would be open to inputting information, I said that early on. The idea is that a computer working backwards could help fill in the gaps and guide my memory.