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

This is exactly what our binary friend has been talking about with the "possible set of games". Given that 1300s are not truly random, could it be the case that some possible games are not in fact possible for a match between 1300 and 2700? Perhaps it could. So is it the case that the possible set of games does not include any wins for the 1300, i.e. is it impossible for the 1300 to win? Well, it could be, but you actually need to make some pretty big assumptions for that to be the case.

If it is in fact possible, I would expect it it happen much sooner than that.

But the probabilities are low enough and the games between 2700 and 1300 rare enough that it's safe to say it won't happen.

For me to answer this question correctly I've only had to say that there is a non-zero chance, even if it was one in a trillion or something, but truth be told, I don't think it's nearly that unlikely. I don't think it would be unreasonable for the 1300 to win after, perhaps, thousands of games. And certainly not millions.

"A lot of people keep approaching this as if it were a matter of chance. It simply is not."

Oh no, it very much is.

<TheronG12> no no no my friend - you will find that my estimate was quite accurate. 

Especially if you improve enough to understand and appreciate what is the actual meaning of levels such as 1500, 1600, 1700, 1800, 1900, 2000... each one progressively. 

What it is to be at that level, what it is to play somebody at that level...

For non-infinite very large numbers, we could have:

A always wins against B
B always wins against C
But C sometimes wins against A!

This is a cool observation for the way performance (in terms of win/loss/draw) against one opponent does not intrinsically predict performance against other opponents.

This undermines the whole enterprise which strives for an ideal rating system by using past results to predict future results... at least in the sense that when it interacts with reality that it produces ideal results.

Here is another question:

How much more, if at all, will a series of random moves beat a 1300 player vs beat a 2700 player?

It seems you can't reliably use the frequency of error vs good moves because the winner isn't necessarily the one who made fewer mistakes.

Or another way to ask perhaps is, is the winner more commonly than not, the player who made a fewer number of mistakes?

What if 1300 rated is sandbagger?! lOL

"This undermines the whole enterprise which strives for an ideal rating system by using past results to predict future results... at least in the sense that when it interacts with reality that it produces ideal results."

It's not necessarily much of a problem as long as you're taking results from a large amount of players. Like if you only played 4 players in your life at tournaments all the time, your rating might get a little weird. But as long as you were playing dozens of different players, there wouldn't be much to worry about. There can be some oddball players but it won't be the overall trend.

Sandy_McBaggins написал:

What the heck is a sandbagger?

Heheheh Sandy_McBaggins

Exactly, and that's why it'll take thousands of games for the 1300 to win -- because it's so hard to beat 2700 players.

I agree.

At that moment it was just sort of reaffirming to me our sometimes odd relationship with reality. Whether that's because the nature of our language, or the incompleteness of our information, or something else, there are these ideals (like a perfect rating system) that we can imagine, but not produce...

Even in a modest sense like how well Elo succeeds on its own terms. It seeks to give a reliable probability for a future event based on past events but it (seems) to fail for random players. Sure this may be obvious and have no (or little) practical meaning, but in a strict sense it still failed.

I play at a 2000 rating level now and there is ZERO chance I could beat a 2700. I even played to a 2300 strength at one time. 2700 is NOT a number, it is a rating. These guys are so very very strong that a 1300 would literally have no chance at all. I destroy 1800 players 29 games out of 30! It is virtually impossible it's like saying someone who barely knows how to play can beat someone who regularly beats extremely talented players. Please.

Now, can a 2300 player draw a 2700... THAT could happen. It's extremely unlikely but it definitely could happen. Maybe... ;)

lol wtf, a 2000 has a zero chance, but then suddenly a 2300 has a chance?

And how badly must a 2000 play to beat an 1800  96% of the time and still have a 2000 rating?

2300 draws against 2700s have probably happened a decent amount of times, in fact.

I know it's bait, but unsure of quality of the bait.

Maybe attention bait?

the question isn't if it's possible, but how likely it is. if the 2700 is drunk, sleepy, not paying attention or not trying to win, anyone could beat them. maybe there's a .000001 chance, but nonzero.

Geez, this is a hot topic. WIth 4 years and 204 pages. OMG 4069 comments including mine.