Solving chess? With no BS. (moderated)

Well here you want to switch to " x is impossible because you can't think of any way to achieve x". Just as invalid.

I mentioned EGTBs because these demonstrated the shortcomings of the general line of reasoning. For the counterexample you request the current limitation to 7 men probably means few or no examples in the current tablebases. In any case your assertion is irrelevant in terms of solving chess; relevant only in terms of weakly solving chess. There is no more guarantee that positions in tablebases would occur in games with perfect play than the positions you complain of. The tablebses so far produced are concerned mainly with a solution not a weak solution and many illegal positions are included (though admittedly not the majority).

There are, however, positions such as this. Black's most accurate defence requires the acquisition of four knights.

The fact remains however that you are making the assertion that such positions are not relevant, so it's really up to you to prove your case rather than me to disprove it. 

#181
'Solving chess' in this context means weakly solving chess. Also checkers is only weakly solved. Weakly solving is hard enough.

Modern chess engines are still improving, and therefore by necessity are still playing imperfectly.  So we shouldn't jump to any firm conclusions based on their results yet.  The most honest position is to admit that we simply don't know if chess is a forced draw with best play.  I personally think white can maintain their opening advantage by force all the way to checkmate. This is why computer chess tournaments still have some wins -- when an engine stumbles into a forced winning line.

I would prefer to say "solving chess" in this context means solving chess and "weakly solving chess" means weakly solving chess. It reduces confusion.

The two tasks are probably of very different complexity. 

This world changes incredibly fast and, I believe that relatively soon we'll be able to build a computer that will manage to beat this overrated but quite fun sort of spending time.

...humanity needs a lesson in humility (Richard Dawkins).

Just my .02...

I'm sure at some point in human history chess will be solved.  It certainly wont be in my lifetime, and hopefully before we destroy ourselves.

I don't think so.

People don't know how exponents work.

If I asked you what the difference in weight was between a shoelace and a mountain, the answer is the difference is a mountain In other words the shoelace is so small you can just count it as zero, so the whole difference is pretty much whatever the big thing is.

It's the same thing as the difference between 10^40 and 10^50. Since the exponent is only 10 digits away people might think it's basically the same, but in comparison 10^40 is so small it's basically zero.

Chess is really big... so big we need exponents. "Computers are pretty fast so we'll eventually do it in the future" is misunderstanding how big it is.

Like I was saying in my last post, the scale of things is hard to comprehend.

Whether it's something like solving chess or terraforming Mars, you have to start talking about things like "if we used every resource on the planet..."

You know, people tend to think of resources as unlimited, because most of the time we do small things like turn on a lightbulb, but for large things sometimes 100% of the resources of the entire planet wouldn't be enough. That's what a big difference in scale can mean.

That's like saying "the most honest position is we don't know whether the sun will rise tomorrow"

Well sure, that kind of deep philosophical doubt is an option, but it's an extremely impractical (useless) option. Day to day we live our lives making certain reasonable assumptions. It's extremely reasonable to assume chess is a draw with best play.

I just want to point out that solving chess means solving the basic game, not any tournament rules tacked on by FIDE or anyone else.  Those rules are not actually part of the game of chess at all, they are added constructs for competitive group play.  No 50 or 75 move rule.

...and appeal to dead authorities that never made any official claims is even sillier.  But I guess that is what one has to resort to when they cannot quote any living world champions stating categorically that chess is proven to be a draw to their satisfaction.

The use of Kasparov is the mistake that really blows the argument .  If Kasparov were willing to make such a statement, it would be posted all over these solving chess threads every time they came up.  But he isn't, and he hasn't.  He *thinks* chess is extremely likely to be a draw...just like most chess players.  Including me, and, I'm guessing, almost everyone here arguing that chess cannot be solved in our lifetimes.

The drawing margin of endgames is larger than the advantage given by moving first. This isn't a proof, but it's a super easy argument.

We can also consider the anatomy of a winning position using elements such as symmetry, contact between sides, and forcing moves... and then we can note that the opening position scores very low (in terms of winning chances) in all these metrics.

I've had these discussions years ago, I'm just repeating things I've said before. I haven't participated recently because Ponz is annoying, but assume tickler has blocked him so...

@btickler

The 50 move and triple repetition rules have been included in the basic rules for an awful lot of the history of chess. FIDE only removed them from the basic rules a few years ago (and I think they're still in USCF's version of basic rules).

That being said theoretical discussions of chess have mostly ignored them until (probably) this century. But more accurately the theoretical discussions have then not really been about chess.

KQKR? KBNK? 

KNNKP even? There are more winning positions for Black than White if Black has the move, but very much the reverse if White has the move.

This comparison isn't my favorite .  For the sun rising tomorrow, a vast majority of that problem space is known from 100s and 1000s of years of observations of the sun and other stars. 

For solving chess, it's completely the opposite, we've barely scratched the surface of the problem space, and new "discoveries" (i.e. engines continuing to leapfrog with every release) are constantly occurring.  Humans suck at chess, and have only played the tiniest sliver of games in their entire history.  So tiny as to be effectively nothing...your argument about exponents is dead on.

Engines are still learning by bootstrapping their own play at this point, but eventually their play horizon will leave the bounds of what any GM can pretend to understand.  The fact that A0 and Leela have just come out is actually the worst possible time to try and put a stake in the ground, being the first engines to really learn the game from scratch without human bias and valuations built into them.

Nope.  He just knows he won't fare well here, apparently.  This thread is about proving a solution for chess, not having an opinion that is based on conjecture and then claiming conjecture as proof...or in his case claiming a mountain of conjecture that has never even been produced .

Belief is irrelevant.  And the truth is we don’t know if with best play chess is a draw, win for white, or win for black (even this last possibility cannot be ruled out yet).  It’s fun to discuss, and we can make all the “reasonable” conjectures we want, but “knowing” in the math sense requires a rigorous proof.  I think this means achieving a 32-piece tablebase, which I think isn’t possible anytime soon, if ever.