# Relationship between Chess rating and I.Q?

• 3 years ago · Quote · #201
joeydvivre wrote:

"I really don't understand all the debate- there either is a correlation or there isn't."

Because if there is a relationship it has to be linear and it has to be best decsribed by a correlation?  It's not possible that there is any non-linear relationship between IQ and chess skill or that some other copula best describes the relationship?

If there is a correlation, I see no reason why it would have to be linear.  It is possible that there is a threshold level, and also that at some point the law of diminishing returns would set in.

• 3 years ago · Quote · #202

Einstein had an IQ of 160.  Kasparov one of 180-190. Bobby Fischer 187.  A dean of a college I once knew - 80.

Arnold Schwarzenneger 135.  George Bush and Ronald Reagan 96.

There's definately a correlation.   Here's a very close formula for your maximum rating -  10 x IQ plus 1000.   Of course this is depending on many years of hard, intense study and practice AND a strong desire and love for the game to reach this maximum, started at a younger age.

• 3 years ago · Quote · #203
joeydvivre wrote:
blake78613 wrote:
joeydvivre wrote:

"I really don't understand all the debate- there either is a correlation or there isn't."

Because if there is a relationship it has to be linear and it has to be best decsribed by a correlation?  It's not possible that there is any non-linear relationship between IQ and chess skill or that some other copula best describes the relationship?

If there is a correlation, I see no reason why it would have to be linear.  It is possible that there is a threshold level, and also that at some point the law of diminishing returns would set in.

Correlation is a measure of linear association.  There surely could be a curvilinear relationship which you wouldn't measure using correlation.

Correlation simply means that two variables are not statistically independent.

• 3 years ago · Quote · #204
joeydvivre wrote:

Nope.  In fact, 0 correlation implies independent only in the case that they are bivariate normal.

Words have definitions not only outside of, but more linguistically accurate than those inside of, narrow academic disciplines.

You may be a math guy, but blake is more correct than you.  Sorry, but it's true.

• 3 years ago · Quote · #205
[COMMENT DELETED]
• 3 years ago · Quote · #206

You mean PUSHING a door that says pull?

• 3 years ago · Quote · #207

Haha. Yes I saw and changed that. Perhaps that was even more stupid than the act itself..?

• 3 years ago · Quote · #208

Thought you made the mistake on purpose to prove your point further.  lol  Anyone understanding that, has an IQ high enought to become a 2400 player.

• 3 years ago · Quote · #209

As per the mini debate above. Blake is correct, a correlation doesn't have to be linear. I wasn't sure what bivariate normal meant, Joeydvivre, but it appears to be a term from probability theory [ http://mathworld.wolfram.com/BivariateNormalDistribution.html ] which is slightly irrelevent to the discussion at hand, but having the subject in your mind may have caused you confusion. The correlation of variables, x and y, in the equation y = x^2 is enough to disprove your argument.

• 3 years ago · Quote · #210

I don't know why you're getting so upset, joeydvivre. I really didn't mean to offend you.

I'm sure you know exactly what you're talking about, and I admit a great deal of ignorance concerning statistics. But I think you're attempting to apply some detail concerning your area of expertise to the whole of mathematics. 'Correlation' denotes relationship, and relationships by no means must be linear, that much is perfectly clear.

I suspect you are asking me for a linear correlation coefficient for a non-linear equation above, so I will decline to answer.

Edit: I can only presume I was correct that you are generally applying a very specific definition:  http://scholar.google.co.uk/scholar?hl=en&q=%22non-linear+correlation%22&btnG=&as_sdt=1%2C5&as_sdtp=

• 3 years ago · Quote · #211
DavyWilliams wrote:

There's definately a correlation.   Here's a very close formula for your maximum rating -  10 x IQ plus 1000.

Where on earth did you get this from? According to that, Carlsen and Aronian must have IQs well into the 180s, 190s even, i.e. mega-mega-genius. I've heard both of the talking and doing interviews, and they certainly sound like intelligent guys, who obviously excel in their field, but IQ of 180? I'm sure they would laugh at the suggestion.

And this is leaving aside all the points about IQ being subjective, etc etc.

I tried an IQ test and scored 125. According to your formula, I will never break 2250. I find this insulting. I am currently rated 2050 by my national federation and don't see any reason why I can't eventually get to master level if I put the hard work in.

• 3 years ago · Quote · #212

The strange thing is that joeydrive and I were saying pretty close to the same thing and then suddenly a battle of semantics broke out
Joeydrive defines correlation to mean the "Pearson product-moment correlation coefficient" even though I am sure that he knows that no else in this thread is using that definition.  This coefficient is a number between 0 and 1 showing how close observed data follows a straight line.  With 1 meaning that the data points form a perfectly straight line.  There are other ways to quantify correlation.  The quantification that comes closest to most people's intuitive idea of correlation is the "Spearman rank correlation".  If x increases as y increases then the spearman rank correlation is positive, although the relationship is not necessary linear.

• 3 years ago · Quote · #213
blake78613 wrote:
joeydvivre wrote:

"I really don't understand all the debate- there either is a correlation or there isn't."

Because if there is a relationship it has to be linear and it has to be best decsribed by a correlation?  It's not possible that there is any non-linear relationship between IQ and chess skill or that some other copula best describes the relationship?

If there is a correlation, I see no reason why it would have to be linear.  It is possible that there is a threshold level, and also that at some point the law of diminishing returns would set in.

Weird, it's like I said something about the correlation needing to be linear.

I take it neither of you has heard of a nonlinear correlation?

Here you go:

• 3 years ago · Quote · #214
DavyWilliams wrote:

There's definately a correlation.   Here's a very close formula for your maximum rating -  10 x IQ plus 1000.

I'd also like to know where you got this, although it hardly matters. If you let numbers tell you what you can and can't accomplish in life then you are a fool.

• 3 years ago · Quote · #215

The word "correlation" was introduced in the OP.   The standard procedure is to plot the data before trying a mathematical model.

• 3 years ago · Quote · #216

I am an average iq guy. But i worked myself to improve my game, and new ideas of gameplay eventually came to me. My calculations have now become more effective and efficient. Honestly, i would have to do a sample statistical test of pmcc to check this out myself, but intially i am of opinion: practice, practice, practice...

• 3 years ago · Quote · #217

Glad to see this mindless, four year old thread, can be continually brought back to life by people of any IQ.

Thank you @MrWizard, your legacy will outlive you.

There are about 50 other variations of this same thread (feels worse than KID theory) and all equally mindless.

Chess players are an eccentric lot.  Myself included.

"We're Chess Players, we must be smart!"

• 3 years ago · Quote · #218

Your posturing puts us at risk of vindicating post #11, joey, and you haven't commented on the numerous references to 'non-linear correlation' i posted from Google Scholar, so I think I'll just leave it at that.

• 3 years ago · Quote · #219

Here it is:

• 3 years ago · Quote · #220

http://onlinelibrary.wiley.com/doi/10.1113/jphysiol.2005.100271/full

I just thought some specific examples in actual scientific papers might help you along the path to admitting you made a (frankly, quite elementary) mistake. You don't seem like the kind of guy who would be able to do that kind of thing though :(

Not that there's any need to get insulting, but there's a great deal of irony in being called dumb by someone who can't grasp the concept of a non-linear correlation.