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

Agreed. It wouldn't be pretty. 

not untracking this thread is not unlike taking a hammer and bashing your own skull with it until you die.

just gonna keep saying that

But you flunked English 101.

Well, you're getting very good at saying it, I'll give you that much.

not untracking this thread is not unlike taking a hammer and bashing your own skull with it until you die.

That's OK, I'll just assume you're off your meds again.

Will a 1300 player have a chance? Well, first you need that GM to be sitting in an ejector seat and activate it...

To be honest, I think assuming that stats "don't work" at a certain range is at least just as hasty as assuming that they do. Sure, you will get to a point where some "unlikely" result seems to contradict with reality, like a super GM missing mate in 1 in "a position where it would be really obvious to see it."

But... the problem is that, why can't I just say that it merely seems like it contradicts reality because it's one isolated situation (that we never have seen happen before) that is so different from all the other possibilities? It seems incredibly plausible that an imprecise human would confuse the two. I know I would confuse a line that is exactly 1 inch long with a line that is 1.00000001 inches long.

I'm not saying that you couldn't get to the point where something is literally impossible, I'm just saying it's very hard to know when this would happen because whatever seems impossible may very well just be a result of our inability to comprehend extremely big/small things/chances. As far as I know, people are human, and they make mistakes. For them to just suddenly get to inhuman status just because of a really large rating difference seems to be giving humans too much credit. That's just my intuition; but I don't think it's any sillier to say that than to just insist that you know the cutoff point for when statistics don't work, just based on an arbitrary, touchy-feely guess.

Although, one reason to say the odds of a win for the 1300 are less than 1/3200 (or whatever it was) that actually seems reasonable and not arbitrary, is that, GMs are good at drawing if they need to, evidenced by all the draws at the top level, and the fact that even Carlsen struggles to win against a lower rated GM who is determined to play drawish stuff as white. So while the 1300 may score 1 point out of 3200, it is much more likely that this point will come from two draws rather than one win, especially because a GM would probably view a loss against a 1300 as an outright nightmare. A GM could generally bail out even if things were somehow getting tough in one game.

what a stupid thread

not untracking this thread is not unlike taking a hammer and bashing your own skull with it until you die.

Remove the double negatives and I may try to understand what you're talking about :)

Not many, but no one said many.

Where's that spam report thread? Jimmykay probably has been reported several times now.

"With normal playing conditions a 1300 will never best a 2700."

Well surely 200-300 point upsets happen under "normal playing conditions." Sometimes things are just chess related. You of course could always complain that, even with 100-200 point upsets, it had to be because the other guy wasn't playing as well as he usually is (almost by definition that would be true  ). But that's how humans play chess. We use imprecise heuristics and such and manage our emotions, physicality, etc. That's what human chess is. Ratings are able to measure the overall effect of all of these things because there is no real bias in them -- if you lose a game, it will be recorded in the books. You can't argue with results.

Yeah, that's true.

We can observe though how it's not so great >500 points out? What was it? I know you've seen those chessbase articles though. I don't think there's any reason to suspect it would be fine at the 1400 gap, but you're right that there's no proof it wouldn't be completely accurate.

Although "expected score" still wouldn't help us know the chance of winning.

As for the Kramnik (and others) example, blunders don't happen in isolation. In most cases you have to put a player under pressure (at some point during the game, at least to create some fatigue) to elicit a blunder. I think like most large mismatches, the stronger player wouldn't have a chance to even start playing before the eval is clearly winning for the GM.

But yes, that's no proof of anything. Philosophy isn't very good for this problem. We don't have any stats either so... yeah, this is how 3000 post topics are born I suppose.

"As for the Kramnik (and others) example, blunders don't happen in isolation. In most cases you have to put a player under pressure (at some point during the game, at least to create some fatigue) to elicit a blunder. I think like most large mismatches, the stronger player wouldn't have a chance to even start playing before the eval is clearly winning for the GM."

Oh yeah I definitely agree with that. Certainly you have to take these examples with a huge grain of salt. But it does kind of show that, at the end of the day, GMs rely on the same stuff we do. They can still, like Karpov, say, "oh crap, I thought this move was really good but I'm actually just allowing a fork" even if it happens to him super rarely. It seems to kind of remind me that you're still playing a human.

So, there are a lot of good and bad ways to react to these examples. It can give false hope to people who don't realize just how unlikely these events are, and, as you said, how even these huge blunders often happen after being put under pressure. But at the same time, people want to glorify and immortalize the great masters, and they want to feel like they are immune to doing something stupid... so much so, that even when they do see a blunder by a super GM, they explain it away by saying "well the gods intervened with that one, x 2700 GM is still perfect, that one doesn't count." Well isn't that convenient :) When you see a counterexample to your theory... just... ignore it? :)

Haha, ok :)

The 1300 better hope it's a la Kramnik, and the blunder allows an instant win. I don't think a knight or rook in the opening is going to cut it!

IMO, for what it's worth, there will be some individual combinations of 1300 vs 2700 where it is literally impossible for the 1300 to win. The number of possible games is limited both by the rules of chess, and even further by the heuristics each player will use to play each game. I think it's EASILY possible many of these sets of games are 100% winning for the 2700.

But ok, you (I) can't say that there is no possible combination of 1300 with 2700 where every result is a non-win.

When you're dealing with possible 1 in quadrillion chances, or any small chance down to infinity, I'm not sure I'd ever say zero :) I mean, no, no one will ever win king and knight vs king, but that's probably not what you were getting at :) Naturally things can be known to be possible/impossible by definition.

I'm still basically going with what I said before, that (at least in general), we simply can't distinguish between a really, really small chance, and something that's impossible. The reason why we try to do this is because A, it's more fun that way, and B, we're never really disproven (unless we got really unlucky and a serious game between a 1300 and a 2700 happened and the 1300 won ). The claim that a 1300 can't win is, practically speaking, an unfalsifiable claim, so there is not a lot of motivation for people to stop immortalizing the masters :)