Alvin_Cruz wrote:
Geez, this is a hot topic. WIth 4 years and 204 pages. OMG 4069 comments including mine.
A really long running dead horse.
Is there any chance that a 1300 rated player can beat a 2700 rated player?
Sort:
No!!
Is there anything we can do?
DJsStudio wrote:
Now, can a 2300 player draw a 2700... THAT could happen. It's extremely unlikely but it definitely could happen. Maybe... ;)
Can a 14 year old 2097 draw two GMs (2534 and 2705) and win against the rest to finish second in a 5 game tournament,at the same score as the winner and outperform the 2705 in the tiebreak? If that 14 year old kid is named Andreas Garberg Tryggestad, it happens.
http://turneringsservice.sjakklubb.no/standings.aspx?TID=ArendalGrandPrix2016-ArendalSjakkselskab
If someone underrated by 500 points has a tournament performance 500 points over their current rating...
Abivasu, either way you have to watch out for the people who have more time delay than others in a game. try too remember names and avoid them to up your stat and i'm not sure if you get stuck in a game and you try to abort that might lower your stat.
I think the red witch got to our horse.
Come on people, it's a yes or no question. The answer is yes.
The answer is "yes*" in the same sense that one can affirm that pigs fly.
*The odds are stacked immeasurably** high, but there is a chance.
**Ordinary people cannot comprehend the measurement, but those who excelled at advanced mathematics can measure the odds. It's harder than using a micrometer on a molecule, but some mathematicians excel at this sort of measurement.
We're gonna be tossing pigs at my uncle's farm later this summer. Come along. Whoever makes the pig fly furthest gets a free beer.
solskytz wrote:
<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...
I know the difference between me and a player rated 300 points higher. I've played them often enough. I've played a few games against people higher-rated than that. The difference between me and a Super-GM - yeah, I can't really understand that very well. But then I think that you don't actually understand the size of numbers you're talking about here. You're estimating a probability of 1 in 1 trillion times the age of the universe in seconds... oh, now that's a loaded question! If you put the age of the universe at about 6000 years, you're talking roughly 200 billion seconds. Some people put the age of the universe at 100 times that in years. That means in seconds more like 6*(10^20). And then you're multiplying that by 10^(10^(10^(10^10))). This is getting to be more than slightly ridiculous numbers. I mean, even 10^(10^10) will break an ordinary calculator. That's 10^10,000,000,000. For comparison, the number of atoms in the universe is estimated at 10^78 to 10^82.
So yeah, if it is actually possible, then I would say that the odds are much better than that. Certainly a computer making completely random moves has much better chances than that. The number of possible chess games doesn't come close to the numbers you're talking about, even if you're considering all possible moves for both sides.
By the way, even if we assume for the sake of argument that the odds are one in a trillion, that doesn't mean it will take a trillion games to happen. It may happen in the first game (but it probably won't). The odds are pretty good that it happens in the first half-trillion. But then you might also go two trillion games before the first time.
Edit: Sorry, I was mixed up about the upper estimates of the age of the universe. Oh well, I covered both extremes pretty well :-) That part of the calculation is fairly trivial anyway, in comparison with the 10^(10^(10^(10^10))).
TheronG12 wrote:
solskytz wrote:
<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...
I know the difference between me and a player rated 300 points higher. I've played them often enough. I've played a few games against people higher-rated than that. The difference between me and a Super-GM - yeah, I can't really understand that very well.
This is the point I'm making :-)
But then I think that you don't actually understand the size of numbers you're talking about here.
I actually kind of do.
You're estimating a probability of 1 in 1 trillion times the age of the universe in seconds...
Actually the probability I estimate is immeasurably tinier than what you write here...
oh, now that's a loaded question! If you put the age of the universe at about 6000 years,
No - to the best of my knowledge it's more in the order of 1.4 X 10 to the power of 10 years.
you're talking roughly 200 billion seconds.
Multiply by two million odd.
Some people put the age of the universe at 100 times that in years.
Doesn't come close
That means in seconds more like 6*(10^20).
No - it doesn't mean that. Your calculation is faulty.
And then you're multiplying that by 10^(10^(10^(10^10))). This is getting to be more than slightly ridiculous numbers. I mean, even 10^(10^10) will break an ordinary calculator.
Of course it will. It will break many other things as well...
That's 10^10,000,000,000.
Here you're right.
For comparison, the number of atoms in the universe is estimated at 10^78 to 10^82.
So yeah, if it is actually possible, then I would say that the odds are much better than that.
Don't get your hopes too high on that...
Certainly a computer making completely random moves has much better chances than that.
Hmmm - you may have a point there. Maybe the chances ARE better after all.
A great approach to actually calculate the ratio, would be -
all of the move sequences by which a 2700 player can lose, divided by all of the possible move sequences in chess.
The problem is, that the chances for each individual move sequence actually happening are not equal - so it's rather impossible to calculate.
However, the result would certainly still be much greater than the number of atoms in the universe. Let's be a bit more conservative and put it at 10 to the power of 140.
If on the average the 1300 has 31 possible moves for his move, that would give 10^3 possibilities for his move, plus the 2700's answer.
So in 60 moves (a reasonable length game, but by no means the longest one possible - and longer games reduce your odds) it's already 10^180.
Now let's be generous (arbitrarily generous of course) and say that a 2700 can realistically play losing games in 10^40 ways (which is still way way way more than he could physically ever play, if he plays a trillion games per second until the Universe evaporates) -
So against the random move generator (or 1300, as we fondly call it here) it will be a one in 10^140 chance to lose a game.
And that's small. So it will not happen.
A trillion games per second would make 3*10^19 games per year, or 3*10^34 games in total in case the universe survives for a thousand trillion more years - and we only wish it does.
The number of possible chess games doesn't come close to the numbers you're talking about, even if you're considering all possible moves for both sides.
Yeah - I exaggerated...
By the way, even if we assume for the sake of argument that the odds are one in a trillion, that doesn't mean it will take a trillion games to happen. It may happen in the first game (but it probably won't). The odds are pretty good that it happens in the first half-trillion. But then you might also go two trillion games before the first time.
And how many games do you have to play if the chances are one in 10^140? And I think that I'm still being real generous.
Note before reading-- I didn't read much of this forum, so what I'm about to say has already likely been mentioned before:
It's impossible to PROVE that a 1300 can never beat a 2700. Sure, the odds are very, VERY close to zero but it's not truly impossible. If in fact the odds were 0%, that would suggest that in an infinite amount of games between the two players the 1300 wouldn't win a single one. This whole topic kind of reminds me of the "Infinite Monkey Theorem" (read the whole page on it here). It states that, given an infinite amount of time, a monkey would be able to type the entire works of Shakespeare just by hitting keys randomly. This is obviously a ridiculous concept, but is it possible to prove that the monkey would NOT be able to perform this task? Of course not. A 1300 beating a 2700 is very unlikely as well, but I really can't imagine one could actually prove that this is impossible.
Just my two cents 
It's not impossible. It just will never happen, so long as -
1) the ratings reflect the actual playing strength of the player, and -
2) Throughout the game both players remain in good health, in good form, have sufficient time for the game and want to play the best moves according to their understanding.
@solskytz; I'm conceding your first point.
Re. calculation of the age of the universe, maybe my comment wasn't entirely clear; for my second calculation I used an estimate of 20 trillion years (100*200 billion, not 100*6000). The age you proposed is 14 billion years; that results in an age in seconds of 4.4*10^17. But like I said above, this part of the calculation is actually trivial.
I suspect you're still underestimating the 1300's chances here. Let's talk about openings. A 1300 likely knows a few decent openings. If they're playing to beat a 2700, they aren't going to open with a3, a4, h3, or h4, and you can probably rule out some more moves as well (1. f3 2. g4 
). There go a lot of your theoretically possible games, and a somewhat smaller proportion of your theoretically possible wins for the 1300. Of course this doesn't apply to the random computer, but at least it tells us that we can significantly improve the computer's chances by programming in a 1300's opening knowledge. The really interesting question though is how the 1300 compares to the computer later in the game.
Of course when you start ruling out some theoretically possible games then we get back to Binary Guy's argument. If we can say some games are in fact impossible, because due to their level of knowledge some bad moves are impossible, does that potentially mean that some good moves are impossible? And if so, do all potential wins for the 1300 end up in the impossible category? Which is why I say if it's possible I'd say the odds are better than you think. But still low enough that it will never happen.
In Youtube Is not impossible... It's Hard
I suspect you're still underestimating the 1300's chances here.
No no no I'm not... with all due respect to your no doubt impressive reasoning :-)
p.s. you get confused with your own arguments when you speak about "programming" a computer - as at that moment it stops playing random moves.
This link is to a Youtube video posted by 'kingchrusher' of a blitz game between GM Max Dlugy (blitz rating 3131, at the time) and a player rated at 1400. So if a 1400 player can beat a GM I suppose a 1300 layer could bear a player ranked at 2700 player. The time control was not specified.https://www.youtube.com/watch?v=H5l4S3-ClbQ
Alvin_Cruz написал:
Geez, this is a hot topic. WIth 4 years and 204 pages. OMG 4069 comments including mine.
Including me