Solving chess? With no BS. (moderated)

Also true, although as a kid I distinctly remember that the kids that tried to copy white's moves as black (usually they tried this strategy after I beat them a couple times) were the same kids that like to trade all their material down to a drawn ending, so trading queens would probably be seen as validating their copy strategy (assuming a protected queen and a recapture, obviously ...).

Chess is definitely solvable, and it's been proven to be so by mathematicians. The only arguments which I've heard anyone say against this is something to the effect of "it's too complex to solve", but complexity doesn't make something unsolvable. It just makes it time-consuming to solve.

As far as the arguments go that "there aren't enough atoms in the universe to process these calculations", I believe that that argument is invalid, for 3 reasons:

1. The universe is constantly expanding.

2. The game tree can be heavily pruned down, since most possible moves are clearly inferior to others, and don't even need to be considered.

3. We can re-use the same "space" to solve a different problem. Like if we prove that certain groups of positions are wins and certain groups are draws, then instead of redoing the same calculations that we've just done, we can discard those calculations and instead only focus on the conclusions. Then it's just a matter of trying to reach one of those positions, rather than a matter of trying to finish the game from scratch.

Now, I think it's fair to say that we don't currently have the technology required to solve chess (unless you want to go full Descartes on me and say that I can't prove anything beyond my own existence), but unless humans go extinct or go back to the dark ages through some nuclear apocalypse, I believe that we'll solve chess eventually.

I also believe that chess will still be a fun game to play even when it's solved, just like tic-tac-toe is still played.

Uh oh. This guy, @Deranged, is on another level. Untracking .... Or not 

If you can prove the universe is expanding *and adding new matter at a significant rate* then you don't belong on this thread...you should be hobnobbing with colleagues at some supercollider or something.

As discussed, unless you can prune at least 999,999,999 positions out of every billion, you aren't making a significant enough dent in the problem for humans to have any shot at solving.  Most people have no conception of what 10^46 actually means.  I don't know how many times in these discussions somebody will say "well, checkers was solved with 10^14 positions, so chess is only a little more then 3 times that"... /facepalm.  10^14 is 100 billion positions.  10^46 is a billion x a billion x a billion positions with 10 friggin' orders of magnitude left over .

You can't "use the space" to solve another problem really...because inherent in proving chess to be solved is the notion that said solution must then be communicated, and must also resist challenges by showing the data.  If a computer scientist told you he had proved chess was a draw but he had to dump the storage that held the full proof, you would rightfully tell him/her to get bent.  Much like someone that claims chess is a forced draw proven by a mountain of evidence given in the past, but the mountain of evidence is only referred to indirectly and never shown/linked .

Chess is definitely "solvable" in theory, but there's a much better chance of God handing you answer personally and sharing his burrito with you than humanity solving it with technological advances within our lifetimes.

Why does the universe need to expand "at a significant rate"?

The question is whether chess can be solved eventually, not whether it can be solved within a set time frame.

If the universe even expands at 1 atom per triillion years, that's enough to prove that there'll eventually be enough space in the universe to hold a solution.

And that's assuming the universe isn't already big enough to hold a solution (it is).

As far as pruning goes: yes, we can prune down more chess positions than we did checkers positions. For example: white is up 20 points of material. There'll be very few things you can add or change to the position to make it so that white isn't winning.

And yes, you absolutely can recycle that space. That's how mathematical proofs are made. When someone writes a thesis, there are a lot of assumptions made that the author doesn't bother proving again, because they've already been proven by others historically. So we take their word for it, and we can always re-prove it later, if necessary.

Imagine if you were writing a mathematical proof and you had to prove every little detail along the way. Prove Pythagoras' Theorem again. Prove that 1 + 1 = 2. It would go on forever. So it's absolutely acceptable to continue your work from where previous conclusions were made.

That's not an analogy that works.  This isn't like Geometry building theorems on top of each other like building blocks.  It's more like saying you mapped every star and exoplanet in the galaxy, but you needed more room and had to delete 95% of the map as you went along.

As for pruning based on 20 point material advantages, that is a far, far cry from eliminating 999,999,999 out of every billion positions.  Which was already extremely generous given that it only drops you from 10^46 to 10^34.  You're not even in the same solar system .

It's not an analogy. This is literally a mathematical problem. Solving chess falls under mathematical proofs. And I'm just explaining how these proofs work.

It's not a requirement for us to build a computer which can play perfect chess. We can solve chess without having the solution compacted in one space.

By "grouping" together certain positions, and saying that all positions which share certain characteristics are winning/drawing/losing, we significantly cut down the workload.

I guess this is kind of just another way of pruning down the game tree.

In any case though, you said that there are only 10^46 possible chess positions, whereas there are 10^82 atoms in the known universe, so even if we ignore the pruning argument, and the expanding universe argument, and the recycling space argument, the answer is still the same: there's enough space in the universe for us to solve chess.

I don't disagree there's enough space in the universe to do the job, but not with our current technology, nor anything even posited for realistic future advancements.  If your argument is that chess can be a solved with the resources of our universe before the heat death of said universe, then yes, I agree.  But that's hardly of any value to anyone alive now...escaping our sun's expansion will become a problem before chess is solved, given current information .

That's precisely my argument.

I'm glad we agree on that.

Chess is not solvable. Lacking solid math to back that statement I'm going to say think of it this way:

Even if we did find a winning route within the billions of possibilities, that's just one game, and it would be said "that continuation is actually losing under certain conditions." That's what solving chess amounts to.

I've expressed this before. Chess is a draw. It's a blank slate upon which strategy is drawn. There's nothing special about a draw in chess, that's what chess is. Playing the game is to avoid the draw and push on until either contender cracks.

But let's not idolize the current iteration of this game. Even Fischer, a chess aficionado, explored beyond and created 960.

It's possible earlier forms of chess were indeed solvable. More squares and rules were thus added or perhaps removed. Whatever the case, we reached a point where it stood that both sides are equally winning.

Perhaps someday we'll solve this standard version too. It's not impossible.

https://www.chess.com/blog/WSama/the-first-ever-chess-match

I was gazing across the board, seemingly in another world, as Danny went 16. E4 and I threw up from sheer surprise. It cost me the game but I stood to gain my composure. 

To the last 3 posters...

I'm fine with people trying to inject a little humor, and I'm fine with people stopping by to read the topic to say *once* they find the content less than worthwhile or exciting.  I created this topic so people interested in discussing the possibility of solving chess without a bunch of dubious claims and all caps would have someplace to go.  I actually do not believe it is remotely possible that chess will be solved in any of our lifetimes, so it's merely a speculative question to ponder. 

I am also interested in more realistic future advancements, like asteroid mining, which will work a whole better with space elevators and monofilament wire to get the materials back down to earth without dealing with current more dangerous forms of re-entry.  Somebody should probably start a thread on that.

But...this is a chess site, so the topic here is whether chess can ever be solved, and how to possibly go about it.

Think about it this way, with all of our technology and hard work we have made 7-piece tablebases. To solve chess, you need a 32-piece tablebase.

(Opinion)

Chess can be solved. Just play 1.Ke8# 1-0

Does anyone know if an 8 piece tablebase is currently being worked on? 6 piece was completed in 2005, 7 piece in 2012 so 7 years to add 1 piece. With each additional piece the work to complete must grow exponentially. I guess it would be safe to assume that if work was started on an 8 piece tablebase in 2012 that at some point in the next decade or 2 it should be completed. And so on and so forth right up until 32 pieces in hundreds (perhaps thousands) of years time if the future Nalimov's et al have the time and inclination. Plus if supercomputers continue to become more powerful and are able to be used for the purpose. 

This is the set start point