The Open File - Chaos, Chess, and the Weather

Submitted by NM Zug on Tue, 02/10/2009 at 8:39am.

The Open File

by Life Master Mike Petersen (Zug)

Chaos, Chess, and the Weather

In 1960, an MIT meteorologist by the name of Edward Lorenz was using his computer to simulate a simple weather pattern in the earth's atmosphere.  He was admittedly using a small number of initial variables because of the limitations of computer hardware at the time.  He decided to repeat one "simulation run" in order to check some details.  He plugged in the starting numbers from the previous run, but rounded them off to three decimal places
instead of the six he used initially in order to save time.  He cranked up the computer and went out for a cup of coffee.  When he came back he received quite a shock.  Instead of an approximation of his previous weather forecast, it was totally different!  The small three decimal place difference had been grossly magnified by the iterations of the computer in solving the equations.  He later said he knew immediately that if real weather behaved like his model, then long-range forecasting would be impossible.  So, it seems that in complicated systems small changes in initial values can cause huge changes in final results.  Of course weather is a complex system with an almost infinite number of initial conditions to take into account, so it is natural to conclude that long-term weather forecasting is going to be next to impossible.  If you ask me, even short-term forecasting is impossible if you judge by our local weather forecasters!  But what has this got to do with chess?

Well, chess is a complex system with a large number of initial values to consider.  Many people who are not familiar with chess ask me if I play the same moves all the time.  I try to explain that there are various systems with which one can start the game, but very quickly the astronomical number of possible moves cause the game to branch into unknown waters, with the resulting positions never having been played before.  This doesn't seem to sink in, though.  They seem to think that chess is a game, though complex, that will yield itself to a set of concrete rules.  Not so, of course.  The incredible variety of moves obliterates any notion of predictability.  The total number of chess positions has been calculated to be more than the total number of atoms in the observable Universe (reference my other column "Is Chess Infinite?").  If you had a computer equal in size to the known universe, with each atom capable of performing a calculation of one move every second, even the amount of time since the big bang wouldn't be enough to put a dent in the number of positions possible.  Next to this kind of immense complexity, weather seems like a parlor game.

Look, another way to go at this is to consider the following:  I have played the same opening for black against e2-e4 for over 30 years - a certain variation of the Sicilian.  I have kept all my games.  No two have been even remotely alike.  Oh, sure they start out looking the same.  You can recognize the pawn structure, the ideas, etc.  But before long, one player will vary, if only slightly, from a previous game.  This is now analagous to Lorenz feeding in his rounded off numbers.  We don't get an approximation of a previous game, but a completely different game.  If we did, then chess would rapidly die off as a serious game, because then it really would yield itself to a set of concrete rules.

Here is another fascinating thing to think about.  Edward Lorenz, by accident, had stumbled on to a new branch of mathematics called Chaos Theory.  One of the more amusing observations resulting from this was that "a butterfly beating its wings in Tokyo can cause a thunderstorm in New York City."  Just another way of saying that small initial changes cause big differences in results, but the analogy is apt.  Is it possible that chess might be looked at under the microscope of Chaos Theory?  Who knows?  But at least now it gives you the opportunity to have another excuse for losing.  You can now say that your loss was caused by a butterfly in Tokyo beating his wings too hard.

==========================

Click here for links to Mike's other work on Chess.com

» posted in Other
 

Comments:

by kenmack - 8 months ago
United States
Member Since: Feb 2009
Member Points: 310

Summum,

You are both correct in your analysis and you are wrong at the same time.  First of all, there seems to be some miscommunication as to where we are drawing the system boundaries for this analogy.  If the boundaries are the chess board, pieces, and rules only then we are talking about a theortical construct and it would seem Summum has a very strong argument.  However, if the system boundaries are going to be that restrictive then, in reality, the system is non-functioning and further analysis mute.  In order to have a game of chess we need players and once they are added to the equation the game becomes chaotic. 

Remember, chaos at its core is in opposition to determinism.  Chaos is a quality of real systems contributing to unpredictability.  Chaos is a fact and not open to general debate:  All real systems are chaotic and inherently unpredictable in their behavior.  Real systems are unpredictable because it is impossible measure with infinite precision.  Zug, you mentioned slight changes in inputs leading to vast changes in outputs.  Your example is observable in a theoretical system like the one used by Lorenz.  In a real system we don't even know what the inputs are excepting crude generalization. So whereas it is possible to state that mate is the winning outcome of the endgame, Summan, it is impossible for you to predict a checkmate will be the only outcome.  Even at grandmaster level it is impossible to predict whether this move or that will be played no matter how obvious a winning or losing move may seem.  Chaos is not arbitrary; the rules of chess are irrelevant.  Chaos is a fundamental principle and escaping it is impossible, even for chess. 

by Stanislaus - 8 months ago
Winchester, New Hampshire United States
Member Since: Jul 2008
Member Points: 16

what about the chaos that results when one is so bent on winning that a loss results in a total emotional breakdown?

by Phil_from_Blayney - 8 months ago
Blayney, NSW Australia
Member Since: Apr 2008
Member Points: 256

Seeing as Summun quickly, and correctly, dismissed your analogy, perhaps you need to alter it slightly :)

A ship leaves the same port bound for it's regular destination, the weather in between departure and arrival can make for many very different voyages over the years!

by NM Zug - 8 months ago
Central Florida United States
Member Since: Feb 2008
Member Points: 747

To Summum_Malum:

I write "The Open File" strictly for entertainment purposes, which means I sometimes take license with literal facts in order to support my fun topic of the day.  This means that I should expect counter-arguments such as yours.  However, I have not seen anyone make their point as effectively as you.  Very nicely done.

Regards, Zug

by Summum_Malum - 8 months ago
Vancouver Canada
Member Since: Jun 2008
Member Points: 115

A very nice article, but one of the defining trades of chaos is the strong dependence on initial conditions just as Lorenz discovered. In fact a tiny difference in starting conditions should induce entirely different evolutions of the system(s) at hand - and their paths may never cross each other! So if you define your first move as the starting condition, then no chess is not a chaotic system - since different 1. - moves can lead to the same positions.Of course you could define your starting condition another way, and then analyze the problem from there.

What is more is that a chaotic system must be rid of noisy and/or random inputs (the chaotic behavior should arise from the non-linearity of the system). Beginners often make random moves, thus chess can not at that specific level be characterized as chaotic. How things are at higher levels where such moves should not occur is a bit more abstruse. At higher levels random moves should not occur but there are some moves, that are more likely to be played, in a given situation, than others.

Before this becomes too long I would like to finish off with the end-game.
If we want to characterize chess as a chaotic system, then we must take the end-game into account.
Remembering that different initial conditions should result in completely different results, we must ask ourselves what we define as a result in this situation - is the endgame, the mating position or what!? .. If we mean the endgame, then again we must define how we characterize the endgame.
Assume that we have two endgames with the same pieces on the board, but at different places - are these two endgames different? We all know that in an endgame with f.x. a two bishops against a king, there is a right way of proceeding, so all endgames with two bishops against a king can (if played correctly) end in a finite number of mates. Thus if we define the mating position as the result, then again - no chess is not a chaotic system.

by h777 - 8 months ago
Vancouver Canada
Member Since: Nov 2008
Member Points: 7333

Nice!

by Warp - 8 months ago
Long Island United States
Member Since: Oct 2008
Member Points: 40

great read

by Mm40 - 8 months ago
Essex County, New Jersey United States
Member Since: Sep 2008
Member Points: 2097

Thank you, Zug. I've enjoyed all of your articles very much Smile

by skwirlguts - 8 months ago
Florida United States
Member Since: Oct 2007
Member Points: 292

There may be infinte games with slight differences but if you take into consideration that both sides are plaing well to win then the number of infinite possibilities reduces immensly. Its these decisions that lower the number of possible moves. The above article suggests that you can't have a computer complex enough. But the puzzle is to design one that looks for wins not at every computation.

by Yonatanof - 8 months ago
Ramat Hasharon Israel
Member Since: Sep 2008
Member Points: 77

Nice! great article!

by luis3141 - 8 months ago
Argentina
Member Since: Feb 2008
Member Points: 93

A nice read, thanks.

 

Add your comment:

Join Chess.com for free to add your comment! Already a member? Then login now to comment.