Is there any chance that a 1300 rated player can beat a 2700 rated player?

That's not a good analogy: Tyson (in his prime) would have physical strength and speed far, far superior to the average person. These factors are overwhelming (there's a good reason if every combat sport has weight classes).

On the other hand, in a chess game both players have the same pieces so the way a master beats a beginner is with tecnique alone. If I play against Kasparov, my Queen is just as powerful as his Queen. If I fight against Tyson, my punch is nowhere near as fast or as powerful as his.

That's why the average person would survive about 1.5 seconds against Tyson, but the average chess beginner would survive much longer against Kasparov and maybe even put up a decent fight (even if he will inevitably lose at the end).

The skills differential in sports are much, much greater than in chess.

Thanks for commenting on my analogy, but I strongly question your theory.

I mean, I hope their understanding of chess is so much deeper than ours.

1400 ELO difference is like us playing someone 200. Even without a queen we would win, no?

we may forget one thing some old chess players that were rated around 2000+ are now rated 1300. on a good day they can beat nearly everybody but on a bad day they can lose to even the weakest club member.
they might have a shot with old openings on a very good day.

Well, though I think chess ability decreases for most people with age, I wouldn't expect a former 2000 to fall as low as 1300, unless he suffers from severe health problems...

Very interesting, I am a 1700+ playing a few 1300 players and I win most and draw a couple just like you suggest.

This isn't how the question was constructed.  If the 1300 gets better 20 years later and wins, who cares.  He's not 1300 anymore so it's not within the parameters of the question.  Besides that happens every day.  Every 2700 player was 1300 at some point, and now they beat 2700 players.  So what.

As long as the monkey's keystrokes were completely random, at some point they'd type anything.

As for 1300 vs 2700, there is enough human error even in GM play to lose them the game I believe.  GMs hanging pieces and missing mates in 1 or 2.  As you said massive material is often not enough for them to win, but after a few of these errors (or a mate oversight) the 1300 would win.

In a tourney game I believe Judith Polgar lost to a B player because she overlooked a 3 move mate.

And of course this isn't a 2700 on his best day, that's not how ratings work, or how players play.  A 2700 at 2700 performance would never lose to 1300 player at 1300.  But because these "random" blunder occur where mates are given away, infinitely many games is enough that it will happen.

Now for a 1300 to beat a computer in infinitely many games, I think he'd have to play totally random moves (i.e. not 1300).

maybe pherhaps the 1300 hasn't played a lot in tournaments.

"If a 1300 player and an 2700 player infinitely many games, the 1300 would win as many times as he lost lol."

I see where you're coming from, but I'm not sure that's correct. Although there will obviously be infinite wins and infinite losses, some infinities are bigger than others. For instance, there are infinite real numbers and infinite natural numbers - but the set of real numbers is infinitely bigger!

http://en.wikipedia.org/wiki/Aleph-null#Aleph-naught 

Yeah, countable infinity and not countable infinity and all that... sometimes that makes sense to me, and other times it seems like a contradiction to say one is bigger than the other just because we can't list it in set notation or something.  Who cares?  Make a different kind of set notation and we'll list it that way

Anyway, I have to default to mathematicians, because they've been thinking about it a few hundred years longer than I have heh.

Quite true, but this doesn't apply here - both the GM and the patzer win the same number of games, as can be shown by putting them into one-to-one correspondence... just as the set of integers has exactly as many members as the set of even integers (1 maps to 2, 2 maps to 4, 3 maps to 6, and so infinitely on). However, we can say that in the limit as the number of games tends to infinity, the GM's win/loss ratio is, oh, say, 10,000:1. 

GG (math graduate)

Yep. Boggles the mind a bit, best just to leave them to it.

Anway my respose to the OP (like most others, I think) is simply that it is possible, but unlikely. However I don't think the question is analogous to me beating Usain Bolt in a 100m sprint, because there are solid physical reasons why that can't be done - with the presumption that we are expecting the participants in both cases to be performing to their ability, i.e. Usain doesn't trip, and the GM doesn't do something like move his knight back and forth endlessly.

It's more analogous to tossing 100 heads in a row. This is of course certainly possible, though undoubtedly many orders of magnitude less likely [ http://www.wolframalpha.com/input/?i=chance+of+tossing+100+heads ]

Even if you pitched a computer playing random legal moves against a 2700 GM 7.8 x 10^31 times, I would expect the computer would find many wins amongst the endless strings of losses. 

Ah. Thanks for clearing that up for me Gill-Gandel.

Your calculations are actually wrong.

The thing with random legal moves is that some are genius, and the comp., playing them randomly, will eventually play the sequence of moves to do that the whole game, yet this will be extremely unlikely.

For example, lets say that there are an average of 60 legal moves for one side in a position. To beat a Super-GM you would have to, say, in a game of 45 moves, play the genius move in 45 turns in a row. There are probably 2 moves per position that will do this (on average). Therefore, your chances of doing this are (1/30)^45= 1/(30^45)= 1/(2.954312707x10^66)= (2.954312707x10^-64)% chance of a comp. doing that beating a super GM, which, taking luck out of the picture (ie. if I tossed a coin 100 times I would always get 50 heads and 50 tails) would mean that out of aprox. 3.4x10^63 games the comp. would win one, so out of 7.8x10^31 games, the comp. would probably win 1 or 0 games, which is definetely not "many wins"

BTW tossing 100 heads in a row is much more likely (the chance is aprox (7.8x10^-29)% chance, aprox 2.5x10^35 times more likely than beating a super gm using a comp. w. the way described)

That's assuming the 1300 is playing completely randomly (if he were, then he wouldn't be a 1300). You can expect the 1300 to either pick a good move most of the time, and in positions where he can't pick a good move, at least he could narrow it down to two or three candidate moves where a good move is one of the candidate moves.

So maybe there are ten crucial moves out of the 60 where the 1300 is not capable of picking a good move with certainty, but can narrow it down to 1 out of 3 possibilities (where one of the 3 is actually a good move and the other 2 are bad moves).  Then, the chance of the 1300 getting all moves correct, would be 1 out of 59049.  Unlikely but possible.

There are many positions where a 1300 won't consider the 'good move' as a plausible candidate. Actually, you need to be very strong to consistently have the 'good move' among your candidates

There is no chance that a 1300 player can beat a 2700 player.

I never think the chance would be 0%, because any player is capable of blindness -- mate in 1s, hanging pieces -- the most basic blunders do occur at GM level, just extremely rarely. Almost everything would need to fit in place for the 1300 to win -- the GM makes the blunder (somehow before the 1300 makes one), despite the fact that he is probably not under a lot of pressure; the 1300 sees it; then the 1300 is somehow able to not blunder it all away to at least ensure the GM to draw... etc. It's very unlikely that all of that stuff would happen in the course of one game, but if we look at each individual component, none of them are impossible (an assumption, but hopefully an agreeable one), so there isn't some law in the universe that eliminates the upset (which is just all the components alligned together) as a possibility.

And to address Estragon's example: I'm sure there would be a few times where the 1300 would actually win that final position, even if he just suddenly came up with the right idea. Clearly, your story shows that it's possible for you to allow your opponent an overwhelming position; I am going to assume that, although the 1300 wasn't able to win in that case, that him somehow finding the right moves, is not impossible, and would eventually happen if that same situation occurred a few more times.  Although you will find a lot more "almost upsets" (like this one you have shown) than "actual upsets," that doesn't mean the latter can't ever exist.