Not logged inRybka Chess Community Forum
Forum RybkaChess Help Search Register Login
Up Topic Rybka Support & Discussion / Rybka Discussion / Polugaevsky variation 7...b5
1 2 Previous Next  
- - By jejega (**) [gb] Date 2012-04-18 08:14
Like many players, I have heard the variation was refuted but so far I haven't been able to find this refutation. The moves are 1 e4 c5 2 Nf3 d6 3 d4 cd 4 Nd4 Nf6 5 Nc3 a6 6 Bg5 e6 7 f4 b5 leading to highly complicated play
Parent - - By Eelco de Groot (***) Date 2012-04-18 22:40
rnbqkb1r/5ppp/p2ppn2/1p4B1/3NPP2/2N5/PPP3PP/R2QKB1R w KQkq b6 0 8


This is what the Vesely book says about the position:
Parent - - By jejega (**) [gb] Date 2012-04-20 16:28
I have downloaded the book and tried to open it with Chessbase but I can only see the board, no moves.
Parent - - By Eelco de Groot (***) Date 2012-04-21 02:30 Edited 2012-04-21 02:32
Hello Jejega,

When I first thought I had the complete book downloaded without the server timing out, it turned out I was still sort of a few kilobytes and the book would not work. Maybe the same has happened to you. Assuming that you tried to download the book as one file from Uly's link in this thread my copy was 1.318.645 KB  1.350.291.861 bytes (looking at the properties info of the file, still in 7Z compressed state). Could you check that? Ernest Bonnem had the same filesize, see the above thread where we discussed that I had probably a broken copy at first. There is also a site where you could download the book in parts but I don't know if that is better, the link was faster but I did no try to build a book from the separate files myself. It took me seven hours to download from Uly's link so that is not ideal but that was the original link from the book author. And posting the complete book is maybe too large to do here for an attachment on the forum.

Regards, Eelco
Parent - - By jejega (**) [gb] Date 2012-04-22 20:43
hi,

thanks for your tips but it seems I still have a problem for the download.
If you don't mind can you please attach a compressed file of the book after the 7...b5 move?
That would be great
Parent - By Eelco de Groot (***) Date 2012-04-23 00:59
Hello Jejega,

I only have the free Chessbase Light demo here and Hiarcs 11. I don't know if they support such a thing, Chessbase Light is out regarding book exporting I'm pretty sure. I don't even know if such a thing is possible, don't you need the full Chessbase program for that? I would not mind trying to send the book to you by e-mail if you PM me with an e-mail address. But as I said it is a pretty hefty file so you need a mail provider that accepts such large files. If anyone has better suggestions or knows how to export just a part of the full book?

Regards, Eelco
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-18 23:25
This opening should be fairly sound, though white will have a small advantage.
Parent - - By Uly (Gold) [mx] Date 2012-04-19 04:05
Hi turbo, have you seen the discussion about how distant we are from perfect chess I've had with Kappatoo? Your name has been mentioned a couple of times and I think your input would be interesting:

http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?pid=412361#pid412361
Parent - - By Barnard (Bronze) Date 2012-04-19 21:51
you found the 'shape of your shoe' [incomplete copy/paste] with Kappatoo,and you need [incomplete copy/paste] help to 'fight' against him? :lol:
Parent - - By Uly (Gold) [mx] Date 2012-04-20 04:51
No, but if turbo appeared and agreed with Kappatoo, it would greatly undermine my argument as it was based on concepts from turbo's postings.

So, if turbo himself says that he couldn't be able to draw his games against a player from the future, then I'd have to concede the argument.
Parent - - By Barnard (Bronze) Date 2012-04-20 17:32
we are in the present,and we cant be sure how will be the future,so all we can speak about it is theorizing

but even theorizing,if we agree that a player in the future plays perfect chess,we must agree we dont play perfect chess,so we wont be able to draw the game against a player from the future (of course,playing perfect chess;if the player from the future doesnt play perfect chess[perfect doesnt understand about 'grey scales',or is perfect,or it isnt,like the nalimov tablebases,a position is ''won'' or ''draw'',not with draw scales,so the chess is ''perfect'' or ''not perfect'',not with scales''],is the same than if he lives in the present,since we also play 'imperfect chess')

that in general terms
Parent - - By Uly (Gold) [mx] Date 2012-04-20 19:28
My claim is that with current resources, a player can find the perfect move in 48 hours. Actually, Kappatoo agrees to this point, but says it would require 200 hours.

This assumes the perfect path to draw is very wide, i.e. all first moves of white draw with perfect play, all black replies to 1.e4 draw with perfect play (including f5), etc. If the path to draw is narrow, as in, the only drawing moves of white from the opening position are e4, d4, c4, f4, e3, d3, c3 Nf3, Nc3, b3 and g3, all others losing to black's perfect play, and to e4 only drawing e5, d5, d5, e6, d6, g6, Nf6, Nc6, b6 and g6 (and so on, the moves of perfect moves of both sides are reduced continuously as the game advances) all others losing, then I'd agree it may be easy to a player of the future to beat an imperfect player.

However, I think the perfect path is wide, so the future player has a table that tells him how to mate an imperfect player ASAP that has blundered, but doesn't tell him anything about how to make the present player blunder or any way to predict the present player's moves. All he can do is hope that the present player will blunder.

I asked Kappatoo  for a position where a player of today can't find the only saving move in 48 hours, because I don't think there's evidence that such a position exists.
Parent - - By Kappatoo (****) [de] Date 2012-04-20 19:59 Edited 2012-04-20 20:03

> Actually, Kappatoo agrees to this point, but says it would require 200 hours.


No, I don't. Maybe you misunderstood my point that if you (YOU) say that correspondence players play perfect chess with 48 hours/move, then (assuming that correspondence players are roughly Elo points stronger than engines, this is again not my assumption), you should also say that an engine is capable of playing perfect chess with 200 hours/move.

Edit: By the way, I don't really understand the rest of you post, either. Of course, the player of the future would be able to predict the moves of a player of today - by relying on resources of our present, i.e. his past.
Also, today we can tell that 1.d4 offers better chances against less than optimal play than 1.f3. Why do you assume that this will be forgotten in the future?
Parent - - By Uly (Gold) [mx] Date 2012-04-21 06:10

> No, I don't. Maybe you misunderstood my point that if you (YOU) say that correspondence players play perfect chess with 48 hours/move, then (assuming that correspondence players are roughly Elo points stronger than engines, this is again not my assumption), you should also say that an engine is capable of playing perfect chess with 200 hours/move.


No, I claim the human component is critical and an unattended engine would eventually lose given enough games regardless of time control, due to flaws. The human, knowing about those flaws, wouldn't fall for them.

>Of course, the player of the future would be able to predict the moves of a player of today - by relying on resources of our present


Not true, players today fail to predict the moves of the opponent even if both use similar resources. Actually, as time goes on the move of the opponent gets harder to predict because the move possibilities increase as strength increases. Moves that seemed losing with resources of the past seem playable now.

Once chess becomes perfect you have no way to tell what move the past player will do, and no way to measure how complex a position is because you can't know how hard it would be for a player to avoid making a blunder.

Let's put an example of Tic Tac Toe, how can you predict if I will play center, corner or side?

Another one would be 6men tablebases, and this position I've posted before for a chess analogy:

3rk3/3p4/8/8/8/8/3P4/3RK3 w - - 0 1


Suppose a player of the present can't use 6men tablebases to pick his moves, and has to rely on engines without bitbases, etc.

Can a player from the future beat a player from the present from this position? No.

Can a player from the future predict the moves of a player from the present in this position? Not more so than players of the present can predict the moves of other players of the present.

The playable moves just go up exponentially with each move so even if you have the solution of the game you would lose any bet you'd make about what position will be reached after move 15.

>Also, today we can tell that 1.d4 offers better chances against less than optimal play than 1.f3. Why do you assume that this will be forgotten in the future?


No, the point still is that a future player with his resources would see those moves as "Draw", and nothing else. To regain the granularity loss he'd need to downgrade to resources that don't have yet f3 solved, so that 1.d4 looks like 0.00 while 1.f3 looks like -0.01, but with many moves looking like 0.00 he'd need to downgrade again to before those moves weren't solved, so that, say, 1.e4 is 0.02 and 1.d4 is 0.01. As you can see this player is approaching the resources of our present and having the 32men tablebases isn't helping him to choose a move.

You say that I underestimate the complexity of chess, but I don't, this complexity is what is not allowing the future player to predict the moves of the present player, even after solving the game. Since he can't know in what positions the present player is going to blunder, he can only have a plan against anything the present player tries, but no plan is guaranteed to win, as the present player can just find the perfect move and play it, this is the problem of the game being draw with perfect play.

That's why I think showing a position where current resources would miss the saving move is important, then you could point at it and say "a future player would just need to force his way into such a position to beat a present player".
Parent - - By Barnard (Bronze) Date 2012-04-21 17:59

>Can a player from the future beat a player from the present from this position? No.


yes,if the other side plays sub-optimal (or directly,wrong moves) moves,and continue playing those moves,until a draw position is converted to a win position
Parent - - By Uly (Gold) [mx] Date 2012-04-22 00:41
That's wishful thinking.

It's like claiming a perfect opponent can beat an imperfect one from this position:

4k3/8/7r/8/8/R7/8/4K3 b - - 0 1


It's trivially easy to avoid exchanging rooks. And the perfect player doesn't have a strategy to force a win.

The same carries to the other position, and I claim the same may carry up to the opening position.
Parent - - By Barnard (Bronze) Date 2012-04-22 00:48
being black,you can play Rh3,and you can bet i will win you

and an imperfect chess player make imperfect chess moves,and since you are claiming chess is draw,with 32 men tablebases,one move gives you 'draw' or 'losse',not measuring 'how wrong is the move'

so from that position yes,any side can win if the other side plays sub-optimal moves

the only position that any side can win is King against King,because any side can make suboptimal moves (and of course King against King+bishop,and King against King+knight)

but the position you showed me,can be won by any side
Parent - - By Uly (Gold) [mx] Date 2012-04-22 01:17 Edited 2012-05-15 14:01

> being black,you can play Rh3,and you can bet i will win you


No, I can't play Rh3, that's a thing that will never happen.

Let's make a bet, let's play the position one million times, and I will never play Rh3. If we played it infinite times, I'd never play it.

So I think your misconception comes forward from the idea that I may play a blunder like Rh3, but that blunder will NOT happen in reality, ever.

It's the same with the other position, and I claim it may carry to the opening position (after you up the resources and time control), because I haven't seen yet a position where 48 hours wouldn't suffice to avoid losing.

>so from that position yes,any side can win if the other side plays sub-optimal moves


In these positions with Rook and King against Rook and King even without tablebases, it's trivial for me to never lose against anything. You just assume that a player without the solution of the game is eventually going to make a mistake, but that will not happen.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-22 14:30
If you played the position infinitely many times, you would eventually play Rh3, or Ke7 and Ke6, or something like that.  Unless it's prevented by the laws of physics, it will eventually happen after infinitely many times.  Even in a million games, I bet that you'd accidentally play it--human fatigue would make you want to be creative, such as moving the rook to different places, and at some point, you'd accidentally drop it on Rh3, or maybe remembering a number of previous games, not keep the position in mind and accidentally play it.  But then again, that's probably not just you--a very strong player might do the same thing after a million games.
Parent - - By Uly (Gold) [mx] Date 2012-04-23 23:07

> If you played the position infinitely many times, you would eventually play Rh3, or Ke7 and Ke6, or something like that.


Then I claim it's impossible to play it infinite times, as all the times where the game would be played would be countable, and would never reach a number high enough in where Rh3 would be played.

In effect, if the games were numbered:

Game 1.
Game 2.
Game 3.
.
.
.

And we played them forever, at every instance in time a game would have a finite number, and none of those games with finite numbers would contain the losing Rh3. That is, you can't name the number of the game in where I played the blunder, the game in where I play it would always remain at infinity, and would be a moment that time never reaches.

If that doesn't convince you, I'm going to use the script argument. Assume I make a script that flips either coin or tails 50% of the time. If the script is run infinite times, the chance that it will play tails eventually reaches 100%.

But this case is different.

Suppose that I delete the code for "tails", so that the script only flips heads. Then I can run it infinite times, and since tails isn't in the code, this is the equivalent to flipping a coin with heads on both sides for infinite time. Tails can't come up. Ever. It's not even there!

Now, instead of a program that always flips heads, I make one that, when facing the position at hand, plays always Re6+. Actually, I can make a program that thinks the only legal moves in chess are checks, and thinks that the Rooks can only move side ways (but still give check as if they moved forwards), so that on this position it only considers Re6+ as the only legal move and can't play anything else, so that over infinite games it's impossible for it to play Rh3 (as impossible as flipping tails instead of making a move).

My claim is that I can behave like this script. If anything, suppose I make a program that blocks any move made to the h3 square so that it becomes an impossibility, and that I use an OS that disallows copy-pasting and a keyboard that has the ALT, H and 3 keys missing, so that I can't even type Rh3 by mistake when sending the move.

The claiming that over an infinite numbers of games I'm going to make Rh3 anyway is like claiming that I'm going to eventually make 1...Bh1+ (not possible because there's no Bishop), 1.Re3+ (The rook can't move there, or it's black's turn to move) or even 1...Tails which doesn't make sense. The infinity arguments only apply to events that are >0%.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-26 11:06
As the time played goes toward infinity (becomes unbounded), the number of times you play Rh3 also goes toward infinity.  (Of course, the game numbers will be known afterward as they occur, though cannot be predicted ahead of time.)  You will also play a bunch of illegal moves, too, but those won't be counted.  Unfortunately, Rh3 would be counted.  It cannot be "deleted" from the code in your head.  It's under the same logic that your phone number can be found in the digits of the number pi (though in this case, the probability is weighted...though doesn't decrease toward zero, making a 100% probability that you will eventually play the move).
Parent - - By Uly (Gold) [mx] Date 2012-04-26 15:35
So you say that if I make a program that flips Heads or Heads, and run it infinite times, Tails will eventually come up?

Turbo, I think your argument is flawed, as I can find infinite manners to avoid myself to play Rh3.

I think this argument can be converted so that it's not as specific (depending on this position and about whether Rh3 is played), for instance, suppose that we arrange a chess match in where I agree to send 1.e4 as white, every time, and that if I sent something else, I'd lose the match.

I claim that if we played the match Infinite times on the Corr Chess section, I'd never lose it, because I can simply, type 1.e4 in a NotePad file, and all the time copy the text to the forum, hit preview, check that it's 1.e4, and if it isn't, don't send the move, and only send it if it's 1.e4. I can repeat the process infinite times and never send anything else.

You argument that as infinity is approached the times I play something that isn't 1.e4 also approaches infinity doesn't make sense to me, that's like claiming that as the life of the universe reaches infinity, the times I appear on the moon the next second also approaches infinity. It's a 0% chance event, like me playing something else than 1.e4. 0+0+0+0+0... doesn't give something even if you do it infinite times.

I'm going to use another analogy. Suppose that an Alien race abducts you and makes you immortal, and creates a world that makes you very happy every day. Only thing, every night you have to come back to your room to sleep, and every day, you have to turn right in an intersection. It looks like this:

T

Your bed is at the bottom, and after waking up, you have to go to the intersection, turn right, walk 10 steps and drop in a hole in the floor, which leads to a paradisaical world of your choice. If you turn LEFT and walk 10 steps, you are faced with a door, protected by a keyboard, in where, if you press the keys 6, 6, 6, the door opens, and you see a pit of lava, and if you jump, you lose your immortality, and die.

Now, going to the left and pressing the buttons and dropping in the lava is a physical possibility, does that mean that if the scenario is repeated infinite times, one day you'll think "oh, today is the day in where I die, why? Because it was physically possible, and I've been here long enough that the chance I turn left, insert the code, and jump to my death reached 100%". I think that's nosense, if I was there, every time I wake up, the chance that I'd turn right would be 100% and the chance of committing suicide would be 0%, after all, if I get bored or something, telling the Aliens to bring me back to the time they abducted me, give me back my mortality and give me amnesia so I forget the whole thing would be more desirable than just jumping willfully to a pit of lava.
Parent - - By Chess_Rambo (***) [at] Date 2012-04-26 17:36
Both Turbo and you are right. From a nonsensical premise like "If I do this infinitely often..." you can deduce everything. :twisted:
Parent - - By Uly (Gold) [mx] Date 2012-04-26 23:48
While I don't think something can be done infinite times (mainly, because it would need to be unbounded to both sides, you can't "start" to do something infinite times, unless you've been doing it infinitely from the past up to today), the argument still applies to a huge amount of time. What turbojuice is claiming is that given enough times, eventually, an event that is physically possible will happen, while I hold that an event isn't dependent on any other event before it, so, an event doesn't care if it's the first time it could happen or if there were billions of billions of tries before. So, if there's no way that something will happen the first time around, the rest of the times it'll be just like the first time, even if they're infinite, and the event will never happen.
Parent - By Chess_Rambo (***) [at] Date 2012-04-27 06:46
Maybe it's easier to find an agreement, if one concentrates on a concrete example like the chess position from your post above. (KR vs. KR)
You claimed:

> Let's make a bet, let's play the position one million times, and I will never play Rh3.


I claim that this experiment will never happen. After you played this position a few hundred times you would become bored and frustrated and stop the nonsense. :smile:

More generally: If on thinks about a physical experiment that should be done very often, than all kind of unexpected influences may happen. This is different from a pure thought experiment. One can imagine that 1/n approaches zero when n goes to infinity, but one cannot observe it. :smile:
Parent - - By Kappatoo (****) [de] Date 2012-04-27 08:34
I'm with turbojuice here. I agree, of course, that the odds of you blundering your rook here are extremely low, but I find it hard to imagine that they are zero. Even the odds of you dematerializing and reappearing on Pluto are not zero.

> What turbojuice is claiming is that given enough times, eventually, an event that is physically possible will happen
> So, if there's no way that something will happen the first time around, the rest of the times it'll be just like the first time, even if they're infinite, and the event will never happen.


I am not sure what to make of these two sentences. The antecedent of the second statement suggests that you think it is impossible that you blunder the rook, while the first suggests that you concede that it is (physically) possible.
Parent - - By Uly (Gold) [mx] Date 2012-04-27 17:39

> the first suggests that you concede that it is (physically) possible.


Well, if the challenge started and a week after we keep playing the position someone appears at my door and tells me "Hi Uly, I'm with Kappatoo and turbojuice on this, and I'm going to offer you one million dollars if you blunder the rook, just to prove them right", then, well, I'd blunder the rook (erm, I'd probably do it for $200...).

Is that the kind of factors that you and turbo have in mind? Because, my claim is that I could keep playing the same move forever if I didn't have a reason to play something else.

Also, bridging this discussion with the previous one we were having, what you are basically saying is that a player of today could beat a player of the future that has 32men tablebases if they played games for infinite times? And again, what makes the game in where the player from today beats the perfect player from the future different from the first game? Are you claiming that a player from today can beat a perfect player from the future in the first game? Because, that's a really weird claim.

I claim that playing 1.e4 every time in the opening position or Re6+ every time on the endgame position is easier for me than it's easier from a player with access to 32men tablebases to play perfect chess (he has to check first that his move doesn't lose, and I don't have to check anything.)

Here, we're not taking about the equivalent of playing Tic Tac Toe in person on a sheet of paper infinite times in where I'd expect to hold all the time (maybe you'd defeat me), we're talking about the equivalent of doing that and I being able to play in the center of the Tic Tac Toe infinite times. That, I can do.

>Even the odds of you dematerializing and reappearing on Pluto are not zero.


It's the non zero chance that I'd do that due to some reason like an advanced alien race having the technology of teleporting, and for some reason choosing me to use it on and materializing me in Pluto? Because, there's no evidence that that's possible. Believing that would be like believing there's a terrestrial planet around Alpha Centaury inhabited by electric blue unicorns that can fly by using Psychokinesis. That you don't know if the chance of that is 0 doesn't mean over an infinite time it'll surely happen.

Now, I'm going to bring up Wikipedia:

http://en.wikipedia.org/wiki/Almost_surely#Tossing_a_coin

Suppose that an "ideal" (edgeless) fair coin is flipped again and again. A coin has two sides, head and tail, and therefore the event that "head or tail is flipped" is a sure event. There can be no other result from such a coin.

The infinite sequence of all heads (H-H-H-H-H-H-...), ad infinitum, is possible in some sense (it does not violate any physical or mathematical laws to suppose that tails never appear), but it is very, very improbable. In fact, the probability of tail never being flipped in an infinite series is zero. Thus, though we cannot definitely say tail will be flipped at least once, we can say there will almost surely be at least one tail in an infinite sequence of flips.

(emphasis mine)

See, the event of a Tail being flipped is >0, yet, the chance of always landing Heads is >0 too, so if one flipped the coin infinite times you can't be sure that tails will come up.

In the same sense, even if my chances of blundering the rook were 50%, over infinite times you can't be sure that I'll blunder it, just like you can't be sure that I won't blunder it every time.
Parent - - By Kappatoo (****) [de] Date 2012-04-27 18:13

> my claim is that I could keep playing the same move forever if I didn't have a reason to play something else.


So, the claim is 'if I intend not to blunder the rook, then there is no chance I will'. Is that correct?
If so, this is clearly not true. You could have some kind of muscle spasm, a stroke, suffer from some kind of illusion, be very absent-minded or whatever. All of this is very unlikely, but it seems weird to think that the odds of any of these events is zero.

> what you are basically saying is that a player of today could beat a player of the future that has 32men tablebases if they played games for infinite times?


Yes, he would.

> And again, what makes the game in where the player from today beats the perfect player from the future different from the first game? Are you claiming that a player from today can beat a perfect player from the future in the first > game? Because, that's a really weird claim.


He can. It is just that the chances are really really tiny. What makes the games different? Some strange random event. The 32men tablebases suddenly corrupt, there is a computer breakdown, some weird event in the player's brain, choose whatever you like. Do you really think that none of this is possible?

>> Even the odds of you dematerializing and reappearing on Pluto are not zero.
> It's the non zero chance that I'd do that due to some reason like an advanced alien race having the technology of teleporting,


No, just due to some very unlikely random quantum events.

Sorry, but I don't see the relevance of your last bit about coin flips to the discussion. You were claiming that the odds of you blundering a rook in the position at hand is zero. Do you want to claim something different now?
Parent - - By Uly (Gold) [mx] Date 2012-04-27 20:22

> The 32men tablebases suddenly corrupt, there is a computer breakdown, some weird event in the player's brain, choose whatever you like. Do you really think that none of this is possible?


I think those are possible. I also think that some future player would exist that would also never have any of these problems, do you really think that all future players would suffer one of those things by force? The laws of physics doesn't prohibit a player from playing perfectly after having access to 32men tablebases, since it's physically possible for a player to play perfectly all his games, I claim that after infinite time some player will manage to avoid all potential problems and avoid losing all his games. And there's no physical impossibility of that player being me, so that if I can do that, I also could be the player that, from an infinite number of players trying to avoid the rook blunder or trying to play 1.e4 every game, over an infinite time manages to avoid all problems, and keeps doing it successfully.

> Sorry, but I don't see the relevance of your last bit about coin flips to the discussion. You were claiming that the odds of you blundering a rook in the position at hand is zero. Do you want to claim something different now?


One is theoretical and the other is in reality. Theoretically after infinite time every >0 chance event happens. Realistically, infinity is never reached and some events never happen, I claim the rook blunder is such an event (e.g. if we started now I would avoid playing the blunder over my lifetime.)
Parent - - By Kappatoo (****) [de] Date 2012-04-27 20:48
You said that YOU could not blunder a rook in the position you gave, i.e. that the chances of you losing such a game are zero.

Now you are saying two very different things. Firstly, that it is physically possible that there is a being for which the odds of blundering a rook are zero and secondly, that you would not blunder a rook in this position within your lifetime. Both are considerably weaker claims than the one you started with (and the one I have been discussing).
Parent - - By Uly (Gold) [mx] Date 2012-04-27 21:17

> You said that YOU could not blunder a rook in the position you gave, i.e. that the chances of you losing such a game are zero.


Over finite times, I stand by this claim. Weird things happen when reaching infinity and it seems most of our disagreement it's about it (for instance, what's the chance that eventually the universe resets itself to the big band and time starts repeating? If that happens there are a lot of >0 chance events that will not happen.)

>Firstly, that it is physically possible that there is a being for which the odds of blundering a rook are zero


And that I could be that being.

That's what I've been saying all along, that I could do it, as discussion advances how I explain how I manage to do it becomes more convoluted.

>that you would not blunder a rook in this position within your lifetime.


And this should be a given since the start since I can't continue the match if I I'm dead, I just made the statement clearly.
Parent - - By Kappatoo (****) [de] Date 2012-04-27 21:27
The discussion was about chances from the beginning. 'The chances of me blundering a rook in this position are greater than zero, but if we play the position a finite number of times, the chances that I blunder a rook in one of those trials are zero' does not really make sense.
Parent - - By Uly (Gold) [mx] Date 2012-04-27 22:32

> The chances of me blundering a rook in this position are greater than zero


Theoretically (I even said that if someone paid me to blunder that I'd do it.)

>but if we play the position a finite number of times, the chances that I blunder a rook in one of those trials are zero


Realistically (I expect to be able to avoid all the theoretical factors of above if I wanted, and never blunder.)

In the end, I think that if I must avoid the blunder to save my life, or to save the universe, or something else, so that there was no conceivable event in where I would want to blunder, that I could avoid it every single time.

You can claim that you could theoretically hire someone so she comes to my house and forces me to make the blunder. And I'd agree. But, if the match started, would you realistically hire someone to do it? No.

So there's a difference between what could happen on paper and what would happen in reality on practice. My claim of 0 chance would be in reality, the ways to make me blunder would remain hypothetical.
Parent - - By Kappatoo (****) [de] Date 2012-04-27 22:41
Sorry, but I honestly think you misunderstand what probabilities are.
Parent - - By Uly (Gold) [mx] Date 2012-04-28 04:59
Try to look at it from outside the box:

I believe in chaos, which means, the more distance in the future one looks, the harder it is to predict. You can't predict what I'll do tomorrow, do you think you can predict what I'll do several years down the line?

You believe that actions can be predicted, such that, you can predict that if the match took place, down the line I'm going to blunder.

When you say I'll eventually blunder you're trying to act as a prophet, while I claim you can't predict the future, and so, you can't say that I'm guaranteed to blunder years down the line.
Parent - - By Kappatoo (****) [de] Date 2012-04-28 09:44
And that's exactly where the misunderstanding lies. I am not saying any of this, so you are arguing against a strawman.
What I am saying is that there is a nonzero chance that you would mess up. I am of course not saying that it is 100 % certain that you would mess up if you tried it some finite number of times. (This would mean that there is a 100 % chance that you would blunder the first time, and that's of course a silly claim.) My claim only implies that if we were to make this test often enough, it becomes highly likely that you will blunder the rook at least once.
Parent - By Uly (Gold) [mx] Date 2012-04-28 17:13

> My claim only implies that if we were to make this test often enough, it becomes highly likely that you will blunder the rook at least once.


I can agree to that (and I'm glad we come to an agreement, since, yesterday's night I thought about examples involving the Many-world interpretations of Quantum Mechanics and the Grand Hotel paradox to exemplify why you can't be certain of the blunder, but that would be for the strawman).

turbo, however, claims that as games approach infinity, I'm going to surely make the blunder, just like he'd surely turn left and die in the Alien example. Not only that, he also claims that if we continued playing games after the blunder indefinitely, he's certain that the times I blunder would also approach infinity.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-27 17:29

> I'd never lose it, because I can simply, type 1.e4 in a NotePad file, and all the time copy the text to the forum, hit preview, check that it's 1.e4, and if it isn't, don't send the move, and only send it if it's 1.e4. I can repeat the process infinite times and never send anything else.


You cannot do this infinitely many times.  There is a very small, nonzero probability that you will eventually "screw up".  Given an infinite number of tries, for example, eventually someone will hack into your account (given that this also has a small, nonzero probability, which it does) and transmit a different move.  The same is true for this situation.  Another possibility is a malfunction in your code--your computer crashes (which will happen an infinite number of times in this situation, given an infinite number of chances), you lose your code (this is also an infinite subset), and you rewrite it.  Occasionally, you will accidentally type a different move when you recode.  There are other ways, too.  In other words, there are uncountably many ways in which you can screw up.  The same goes for trying to avoid playing Rh3.

Of course, the number of times that it actually would occur is so small that it would never affect your Elo.

As for the alien analogy, the human brain is not perfect, and given the possibility of an infinite number of trials, you would (of course) eventually turn left, and among the (unbounded number of) times you do this, you will eventually manage to press that sequence of 3 numbers.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-27 17:30

> 0+0+0+0+0...


Okay, but are you going to tell me that 0*(infinity) = 0?
Parent - - By Uly (Gold) [mx] Date 2012-04-27 17:51

> Okay, but are you going to tell me that 0*(infinity) = 0?


What result would you give to that that wouldn't violate the rules of math? Because, as infinity is not a number I can just say that 0*(infinity) = &, and it may work fine until I get to attempt operations with infinity and & together where I'd need to keep adding more symbols.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-28 14:57
Well, 0*(infinity) can be any number, except zero or infinity.  (I claim that neither zero nor infinity is a number in the normal sense; zero is an identity, as is 1, and it is a member of many of the various sets of numbers, but in the normal intuitive sense of a number, it is a bit different.)  The point is that it is some number.  Another point is that in truth, the analogy to our discussion isn't 0+0+0+..., but is instead 0.0000000...? + 0.00....? + 0.00....? +...
Parent - - By Uly (Gold) [mx] Date 2012-04-28 17:32

> but is instead 0.0000000...? + 0.00....? + 0.00....? +...


I don't think there's any sums in this. You're trying to calculate the chances that I have of avoiding the blunder, and it never gets to 0 after finite trials.
Parent - By turbojuice1122 (Gold) [us] Date 2012-04-28 19:44
I thought that you were making an analogy with adding zero uncountably many times, and my point is that it's not really zero that you're trying to add--it's something that's very small, but not zero.  Of course, probabilities can be "added up" in whatever way makes the math seem easier, whether by adding parts that have already been calculated after separating out terms, or by tackling the hole thing at once and taking any exponentials, etc., depending on the actual problem and the necessary method of solution.
Parent - - By Uly (Gold) [mx] Date 2012-04-27 17:47
Well, see above, it seems you're based on the claim that the chances of something bad happening are >0, so they're going to eventually happen, while my claim is that the chance of nothing bad happening are also >0. We'd need to play an infinite number of games for your case to stand, but I claim the chance of playing infinite times is 0, and that there's no finite number of games we could play to approach infinity (i.e. playing 4 quadrillion games or 80 quintillion games wouldn't get us closer to infinity than our first game.)
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-28 14:59
No, you don't need to play an infinite number of games--one can, in fact, calculate a non-infinite expectation value for the number of games that would need to be played for it to happen N times.  You give the value of N, and someone can calculate, using Poisson statistics (the variance would also be N), the expectation value for how many games should be played for it to happen.
Parent - - By Uly (Gold) [mx] Date 2012-04-28 17:28

> No, you don't need to play an infinite number of games--one can, in fact, calculate a non-infinite expectation value for the number of games that would need to be played for it to happen N times.  You give the value of N, and someone can calculate, using Poisson statistics (the variance would also be N), the expectation value for how many games should be played for it to happen.


Can you do that for coin flips? Because, please tell me how many coin flips are needed to be sure that the next coin flip will be tails. I claim it remains at 50% all along, and you can't predict at what point a run of all-heads will end. It could go on forever (see "Almost Surely" in Wikipedia above).

This means that the chance of getting a tail over a finite number of tries never gets to 100%, when you reach the expectation value of tries, the coin can still land on heads the next time, and actually, that event will not be different than the first coin flip (e.g. the coin doesn't know how many times it has been flipped before.)

I agreed with Kappatoo that the chance of getting tails over a long run would eventually be highly likely, but at no point could you be certain of getting tails, just like you can't be certain I'll blunder the rook after finite tries.
Parent - - By keoki010 (Bronze) [us] Date 2012-04-28 17:49
or it could land on it's edge!!!:lol:
Parent - By Uly (Gold) [mx] Date 2012-04-28 18:46
See Wikipedia article of above, we're using their "ideal" coin that can only land on head or tails, and will always land. That is, after flipping it you can be certain that it will land either head or tails, it's this level of certainty I'd expect of the blunder happening, and I don't think it'll get nowhere near this level.
Parent - - By turbojuice1122 (Gold) [us] Date 2012-04-28 19:42
I can't be certain how long it will take for you to get "heads", and I cannot be certain which game number will be the first at which you'll blunder the rook.  However, I can say with 100% certainty that if you keep flipping the coin, you will eventually get heads, and that if you keep starting from that position, you will eventually blunder the rook.  The use of Poisson statistics is simply to give an expectation value for how many times it should (on average) take for you to get N heads or to blunder the rook N times.
Parent - - By Uly (Gold) [mx] Date 2012-04-28 19:57

> However, I can say with 100% certainty that if you keep flipping the coin, you will eventually get heads


Can you tell me an upper-bound of the number of tries in where you'd expect me to get heads? Because, I claim that for any finite number of tries the coin is flipped, all of them could be tails. That is, for any given number of tries, there's a non-zero chance that all of them are tails. It's physically possible that all of them are tails, no matter how large, and there's no magic number of tries that makes it impossible for the coin to land on tails the next time around.
Up Topic Rybka Support & Discussion / Rybka Discussion / Polugaevsky variation 7...b5
1 2 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill