No opening book, 5-piece Nalimov TB, fixed depth of 12 ply.
True or false? Chess will never be solved! why?
Sort:
I hope it will never be solved
Mathemagics wrote:
Yes, given the openings deemed to be sufficiently strong enough to play (equal for both sides or only slight advantage)... it narrows the possibility of moves substantially from about 10^120 to something much much smaller.
For someone named Mathemagics, you don't seem to understand numbers very well. Even if we use your 10^120 number...if you were able to prune 99.9% of the tree with your "openings strong enough to play" (we'll ignore your logical failure in being able to definitively say what openings are strong enough to play without this self-same analysis for now) theory, the number would still be 10^117.
Your ignorance of computers in general and chess engines in particular is staggering. Engines do not always play themselves to a draw, and your aspersions about Stockfish winning against itself because it is open source is like saying that the earth revolves around the sun because Will Smith is the Prince of Bel-Air.
I'm going to therefore deduce, based on your ridiculous claims, your sudden arrival on the scene, and your brand new account, that you are just another sockpuppet/troll ;)...
What difference will it make if chess is ever solved or not ? Tic Tac Toe is solved and is a draw and very simple . People still play it though and lots of people still lose ...
Idolizing supreme forces while trampling down creation.
Unable to accept their unimportance, insignificance.
Yeah, its funny how even the non chess players don't understand their unimportance and insignificance until they are on their death bed.
If we could just move that up a few years, it would be such a better planet. Would probably produce more competitive chess games.
Thank you watcha for the time and effort you put into your very informative posts. Even though I'm a chemist, I find the discussions of mathematics here most interesting! Thanks also to thegrobe, btickler, and the other members who contributed.
Reb wrote:
What difference will it make if chess is ever solved or not ? Tic Tac Toe is solved and is a draw and very simple . People still play it though and lots of people still lose ...
There is nothing in the title of this thread that asks whether it matters or not. That's irrelevant to the discussion. Does trying to prove whether the Universe will collapse on itself or expand forever make a lick of actual difference/consequence to anyone working on it? For people that think the answer doesn't even matter to spend time arguing in this thread seems pretty counterintuitive ;)...
The Shannon number for chess played on a 20 x 20 board is on the order of 10^150:
http://www.wolframalpha.com/input/?i=400%21%2F%28320%21*%2820%21%5E2%29*%286%21%5E6%29%29
Since this game has more positions than the number of bits that can be contained in the entire observable universe ( 10^123 bits ), storing its strong solution in a tablebase is strictly impossible according to the known laws of physics.
If chess was played on a 20 x 20 board, this thread would indeed be meaningless as many here wish.
However ancient indians were good enough mathematicians and were vicious enough to limit chess to a 8 x 8 board, the solution of which is just on the borderline between possible and impossible, making the generations coming after them having vicious debates about solvability. In the meantime those responsible for inventing chess have reincarnated in new bodies and are having good laughs at us.
Mathemagics wrote:
watcha wrote:
This is a simple position:
Would you believe that not only it is a mate in 261, but there is only one winning move and all other moves draw.
These kind of secrets are unreachable through usual human reasoning ( and even engines fail to find the winning move ). Only exhaustive tree search can bring such hidden secrets to the surface.
Really... "engines fail to find the winning move"? How exactly do you propose it was found then mate? You literally contradict yourself...
"engines fail to find the winning move" and then " Only exhaustive tree search can bring such hidden secrets to the surface"
Unless you're unaware... engines use search-trees to find variations..
Unless you're unaware (sic)... No combination of computer+ software currently existing on Earth can tree-search 500+ plies.
Tablebases are generated backwards (from evaluated positions to the ones that lead to them).
Thats great= We 
can continue playing
Sep 12, 2014
0
#893
LongIslandMark wrote:
I was only an undergrad math major, so take it for what it's worth. But I think chess probably has a structure - problem is the topology (or whatever math concept you want to apply) changes every move.
That's what makes it an interesting game!
Mark,
do you see a correlation between those who are 'good' in Math (even general schooling) and chess 'talent'?
Peter (former Long Guylander who still roots for the Mets)
Mathemagics wrote:
TBentley wrote:
How many games did they play?
it wouldn't matter because given the choice they play the same moves every game
You have not even tried it once, did you?
Most chess software with an opening book choose the opening at random.
Ummm...21 times? Today? What time control are you using for the engines? Set them at 30+ ply.
Mathemagics wrote:
btickler wrote:
Ummm...21 times? Today? What time control are you using for the engines? Set them at 30+ ply.
The ply is irrelevant! And you're telling me I'm a laymen on engines? Seriously?
If they're both set to the same ply they will see the moves the engine is planning regardless of what number you set it to therefore making it a draw...
Or in other words... The plan the one engine is making at x depth the other will see if searching the same depth... and the depth was 10 by the way.
EDIT: NOW I SEE! YOUR EXPERIMENT WAS FUNDAMENTALLY FLAWED... 30+ depth with a given time control is a awful test... like are you serious!?!?! of course you are... There shouldn't be a time control! you should let the engine run through the entire 30+ depth calculation otherwise it's unequal for one side
All I am saying is that your running of engines against each other in "rapid' formats is not going to prove anything significant one way or the other, so why bother? Running Stockfish for 10 seconds a move is much like running Fritz from pre-Deep Blue days. Use the engine's full capacity. Using 10 ply is complete garbage. You will not see any significant characteristics of the engine's play at 10 ply.
Now, please stop pretending that you know anything about computers and coding, you're starting to embarrass yourself. Using the same ply depth obviously gives an extra half move creeping "horizon" to each side as each is playing its move. Using 10 plies obviously effects the engine's tendency to draw itself. Are you that dense?
You are reducing each side to maybe 1500-ish rating, where it only takes a few plies to find an adequate defense to pretty much every plan...hello...10 plies reduces offensive firepower far more than it reduces an engine's ability to defend against those offensive plans...
Yes, this could be solved by computers quite easily (would take some time to code the programs though)
"code the programs"?
Every time you open your mouth about coding, you make it more and more clear by your vernacular that you know nothing about coding, unless maybe it's COBOL on key-punched cards...
Mathemagics wrote:
LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL
That is all...have a nice day and maybe educate yourself while your at it.
Didn't think you had anything. Now LOL your way back to wherever you came from.
You keep on posting your 10-ply engine vs. engine PGNs, and we'll see who gets laughed out of town ;)...
The fact that you think chess rating would have much to do with understanding the infrastructure required for a chess engine just shows again how ignorant of software you are.
LongIslandMark wrote:
Well, to be fair, just about anything non-trivial takes time to make a good program. If you were making fun of the phrasing "code the programs", using "code" as a verb is not uncommon.
No...really? ;)
Developers do not say "it would take me some time to code the programs". In fact, even using the word "programs" at all is laughable. In the age of object oriented programming and event-driven architecture, monolithic "programs" are pretty much extinct. You compile objects and methods, which respond to events; it's all much more adaptive, more modular, and code is divied up at lot more than in the past. Describing how you are going to code something is therefore not something you impart as easily anymore, so referencing a program is unlikely. Rather, you'd just use an informal pronoun "it", or "this".
Seriously, if you were sitting in a room whiteboarding some new software release and your new hire said "it will take me some time to code the programs", your whole team would be looking at each other nervously wondering who had screwed up the interviews.
None of this thread really matters anyway. In a couple of hundred years, a giant asteroid will smash into the Earth and split the planet apart.
Or the twelve piece tablebases will cause an implosion.