And also, I found this piece of info in google which is the reason that no one has invented such a chess engine like I have said. here it is:
CHESS IS NOTHING BUT A LARGER VERSION OF TIC-TAC-TOE!!!!!!
Sort:
teju17 wrote:
And also, I found this piece of info in google which is the reason that no one has invented such a chess engine like I have said. here it is:
Which is sooooooooo mind boggling!
Tell me what do you think!
so , to find all the possible positions after just 4 moves, i am sure it will take about 10 years
lol???!?!?!?!
harrisisawesome wrote:
lol???!?!?!?!
THis is crazy to think of
Search for: Is chess a finite game?
Not so. While the game tree out to four moves is trivial to evaluate for a modern chess engine, generalized Tic-Tac-Toe is PSPACE-complete, while generalized chess is EXPTIME-complete (no deterministic algorithm for generalized chess can take less than exponential time in the size of the board), the latter being believed to be strictly higher in complexity than the former. The current 7-piece tablebase takes 140 TB of space and an 8-piece one would need a petabyte. A nine-piece tablebase might not even fit in the storage of the most powerful supercomputer, Summit. As more pieces are added, the amount of combinations continues to grow exponentially; a 32-piece tablebase, with around 10^43 positions, would likely need far more space than is going to be feasible in the near future: each position takes about a fraction of a kilobyte to specify in FEN, each move another few bytes, so you're looking at 10^45 bytes or so. A petabyte is on the order of 10^15 bytes so the biggest data warehouses are on the order of 10^20 bytes at most. We are going to need Moore's Law to continue holding for computer storage for quite a few years for this to be feasible, not to mention the amount of compute time that would be needed to compile the tablebase (quantum computing will likely be needed).
WindowsEnthusiast wrote:
Not so. While the game tree out to four moves is trivial to evaluate for a modern chess engine, generalized Tic-Tac-Toe is PSPACE-complete, while generalized chess is EXPTIME-complete (no deterministic algorithm for generalized chess can take less than exponential time in the size of the board), the latter being believed to be strictly higher in complexity than the former. The current 7-piece tablebase takes 140 TB of space and an 8-piece one would need a petabyte. A nine-piece tablebase might not even fit in the storage of the most powerful supercomputer, Summit. As more pieces are added, the amount of combinations continues to grow exponentially; a 32-piece tablebase, with around 10^43 positions, would likely need far more space than is going to be feasible in the near future: each position takes about a fraction of a kilobyte to specify in FEN, each move another few bytes, so you're looking at 10^45 bytes or so. A petabyte is on the order of 10^15 bytes so the biggest data warehouses are on the order of 10^20 bytes at most. We are going to need Moore's Law to continue holding for computer storage for quite a few years for this to be feasible, not to mention the amount of compute time that would be needed to compile the tablebase (quantum computing will likely be needed).
but, whatever, chess is nothing but a very very very very very large version of tic tac toe, right?
teju17 wrote:
but, whatever, chess is nothing but a very very very very very large version of tic tac toe, right?
No, because chess' complexity is almost certainly higher, especially without the (artificial) 50-move rule.
come on ! think of it! if u analise evry combinations of moves, u will ultimately get to finish the game (if u live to finish all the possible combinations)
SO, what if WE make a chess engine that is NOT A.I, and JUST AN INTERACTIVE MEMORY IN WHICH WE INSERT THE MEMORY LIKE: IF E4, PUT E6. AND IT NEVER LOSES EVEN TO ALPHA ZERO,........................
should i start planning on how to make one?
@
i also use windows
wrote: (quantum computing will likely be needed).
TRue
teju17 wrote:
come on ! think of it! if u analise evry combinations of moves, u will ultimately get to finish the game (if u live to finish all the possible combinations)
This is a trivial fact of all deterministic, full-information games. By definition of EXPTIME-completeness, any game whose complexity is that can be used, with at most polynomial overhead, to solve any other game whose complexity is in EXPTIME, and because it is believed that PSPACE is not equal to EXPTIME, a generalized tic-tac-toe algorithm most likely cannot solve chess without superpolynomial overhead.
Many of us don't know that if you search google on how to not lose in tic tac toe,you get some vids where they teach you tricks on how to get a forced win and how to block forced wins. Who knows? some day there may be a chess engine that will have analised every possible position in chess and will declare that if both sides play at their best, it will ALWAYS be a draw.OR, it could get like white will always win if he learns every possible position in chess.And also, we observe a lot of draws at the top level of chess matches, so my predictions may be TRUE.IF it is TRUE, then what is the point of chess?
SO, what if WE make a chess engine that is NOT A.I, and JUST AN INTERACTIVE MEMORY IN WHICH WE INSERT THE MEMORY LIKE: IF E4, PUT E6. AND IT NEVER LOSES EVEN TO ALPHA ZERO,........................
WOULDN'T THAT BE GREAT?!!