as long you don't have wise...your intelligence is useless!
Relationship between Chess rating and I.Q?
Sort:
Feb 11, 2025
0
#1683
i used to play at club before I can write... and I had to take an written test to go to a good trainer.... and my mother was not aloud to help me... I did the test orally...and join a GM trainer group...
Feb 11, 2025
0
#1684
Here is the main part of the answer to the question of this forum: there is a correlation coefficient of 0.35.
Here is some synthesised data which has a distribution precisely like IQ(with mean 100 and standard deviation 15, by design) and roughly like Elo (with mean 1200 and standard deviation 200 points by design) and and a correlation coefficient of exactly the correct 0.35. The red lines indicate the 1 and 2 standard deviation bands for IQ and Elo separately.
Note that while IQ is genuinely normally distributed by definition, chess Elo has in truth got long tails. This means is that there are more players at extreme Elos (particularly the upper end). This makes the graph below undersample these Elos, but the general intuition is valid.
It's a million points. so you can't really see that over 90% of the points lie within the outer rectangle made of the 2-sd lines (there are a lot of points on top of each other in the middle.
Interesting that with these parameters, we don't see any players over 2250, but note that the few players near this have IQs between 115 and 130, showing the effect of the significant but only moderately strong correlation.
There is no reason to have confidence in the number Nakamura gave, with interesting humility!
Feb 11, 2025
0
#1685
What is the data of the experiment? Is it just a graph? Is there data on the history, sequences and observations, conclusions,qualitative and quantitative data? Or some data of who specialist in the management and what method was used to conclude on the graph?
Feb 11, 2025
0
#1686
is not intention of detract,but is the properly way to present a result.We all can drawings a tain
Feb 11, 2025
0
#1687
HernanCacciatore1 wrote:
is not intention of detract,but is the properly way to present a result.We all can drawings a tain
I didn't draw anything: Python did. It drew a distribution of two variables, one of which has the defined distribution of IQ (normal distribution, mean 100, standard deviation 15), and the other one which had an mean of 1200 and a standard deviation of 200 (meant to be like Elo). But the most important thing is the the correlation between the two variables is 0.35, which is the value determined by a peer-reviewed study looking at IQ and chess rating.
Feb 11, 2025
0
#1688
Optimissed wrote:
E
If that data is real and reliable,
The correlation coefficient of 0.35 is real and reliable. There are multiple studies that have got quite similar results.
it strongly suggests that IQ is not the major factor in attaining a high grade/rating.
Well, it's more that IQ only explains a minority of the variance in Elo. The rest is due to some combination of unknown factors that are independent of IQ. This is proven by the size of the correlation coefficient.
That's due to the falling away of results (ratings) towards the higher IQ ranges. Fewer people have such ability, obviously. However the effect seems stronger than one would expect if IQ were the major factor.
See above.
Note that there are a million points plotted on the graph, and the occurrence of the extremes of Elo (on a rather compressed scale in my simulation, but this could be done better to have it reach GM levels) at moderate IQ levels is mainly because these are so much more common. Even with a positive correlation, if you have 100,000 people with a moderate IQ and 10 with a very high IQ, it's much more likely the strongest chess players are going to be in the moderate IQ band, because the sample is so much bigger (you get more examples multiple standard deviations away from the mean in a huge sample, and here this outweighs the mean being higher for the small, high IQ sample (because of the correlation).
Of course any two variables which are jointly normal have the sort of elliptical pattern seen, with the slope (in standard deviation measurements) being the correlation coefficient.
That's just an immediate, off the top of head assessment.
And that's a bit more detail.
Feb 11, 2025
0
#1689
Ok ...Although I do not understand the process of creating the graph, I suppose it is based on conventional statistics.
Feb 11, 2025
0
#1690
Optimissed wrote:
It suggests that there may be two more factors of equal importance, although that's only a possibility.
A few years ago we identified a few variables such as ability to learn, concentrate and memorise. I would suggest that those are of roughly equal weight and that possibly, IQ could be the major factor. However, I would prefer to go for ability to concentrate and focus.
Accurate, fast calculation requires excellent short term memory but long term is also important, as well as enough but not too much practice.
Can you think of anything else which is a relevant variable?
Yes: Visualization--the ability to form and hold an accurate picture in one's mind of what the position will look like following a contemplated series of moves. Clearly this will enable a player to make a better judgement of the value of any proposed line of play.
Visual memory--recognizing positions from games/positions played or studied in the past and applying analysis therefrom to similar positions in a game in progress.
c124875 wrote:
Yes, but I take the same test and get 121-125, and I'm just a 600-700
The main determinant of chess ability is how deeply and often you're willing to study. Simply playing and watching Gothamchess will get you farther than most other players, of course. But what really counts is the understanding that study provides.
Take me as an example: my IQ is 72 and I'm quite good. Study will get you farther than innate talent.
punchdrunkpatzer wrote:
The main determinant of chess ability is how deeply and often you're willing to study. ... what really counts is the understanding that study provides. ... Study will get you farther than innate talent.
+1
Feb 11, 2025
0
#1693
I think IQ is just how much stuff you know, but ELO is how much you’ve learned about chess. ELO is a much better way to test smarts and mental capacity
Rating depends on the playing skill of the person and not depend on IQ
punchdrunkpatzer wrote:
c124875 wrote:
Yes, but I take the same test and get 121-125, and I'm just a 600-700
The main determinant of chess ability is how deeply and often you're willing to study. Simply playing and watching Gothamchess will get you farther than most other players, of course. But what really counts is the understanding that study provides.
Take me as an example: my IQ is 72 and I'm quite good. Study will get you farther than innate talent.
Yes, what I was saying is that in my opinion IQ don't have a connection with chess elo
It's an interesting topic. Whenever the relationship of IQ and a given subject is brought up and discussed, it tends to get a pretty strong reaction from people. Some of it, I think, has to do with people interpreting the word correlation too strongly, as if it was implied that something cannot be achieved without a certain level of intelligence. In reality that rarely the suggestion (atleast when talking about above average levels), but the question of how useful is intelligence in this particular subject, or if there are other reasons that may lead to a correlation. I think in terms of chess it's both, intelligence is useful but intelligent people also tend to be more interested in problem solving. This isn't to say there aren't other factors that are more significant.
To me the relationship appears to have some similarities to one of intelligence and level/field of education. There is some correlation (I havent done any research, dont know how much) that may be explained by intelligence being a useful asset in studying, but also theres the motivational factors. A masters degree may allow one to be highly skilled and knowledgeable about what is going on in their field and apply their knowledge in pratice, where as a doctorate degree will force them to ask more fundamental questions on why and how, providing them more tools for generating new knowledge. A highly generalized statement, but what I'm aiming at is that an intelligent person might be more likely to be asking those fundamental questions and not be satisfied without that knowledge. Does that mean that a person of a slightly below average iq couldn't achieve a doctorate degree? No, it is certainly possible. More challenging and less likely, yes.
Both education and iq are influenced with (almost) unlimited other factors, external and internal, not the least of which is culture. If one would like to truly measure the role of iq in chess performance, the study would need control groups with standardized studying resources and maybe even backgrounds in more detail.
Some already got into the questions of what do the iq tests truly even capture. That's a whole another topic but they do capture something when the sample size is appropriately big. When an individual person says their iq is either 109 or 125, I'd say it's unclear what conclusions you can draw out of those numbers.
Feb 12, 2025
0
#1697
Optimissed wrote:
mpaetz wrote:
Optimissed wrote:
It suggests that there may be two more factors of equal importance, although that's only a possibility.
A few years ago we identified a few variables such as ability to learn, concentrate and memorise. I would suggest that those are of roughly equal weight and that possibly, IQ could be the major factor. However, I would prefer to go for ability to concentrate and focus.
Accurate, fast calculation requires excellent short term memory but long term is also important, as well as enough but not too much practice.
Can you think of anything else which is a relevant variable?
Yes: Visualization--the ability to form and hold an accurate picture in one's mind of what the position will look like following a contemplated series of moves. Clearly this will enable a player to make a better judgement of the value of any proposed line of play.
Visual memory--recognizing positions from games/positions played or studied in the past and applying analysis therefrom to similar positions in a game in progress.
I purposely left that out because there seems to be new evidence supporting the idea that visualisation is not such an important skill as was thought. If you recall, I supported visualisation a couple of years back but I think I'm reconsidering. What do you think?
When I calculate, I don't do it visually but rather by building up an abstract pattern. I think that may be more memory based but I'm not sure. Just been walking for 4 hours so am a bit too tired to play chess and notice my thought patterns at the same time.
A study that several of us have referenced (you and I argued about whether it found a "modest" or "strong" correlation) did note that, surprisingly, the correlation between chess skill and visualization as measured by the test was weak.
I think that most people assume it should be strong and the researchers indicated that they had done so.
Feb 12, 2025
0
#1698
Optimissed wrote:
Yes Ziryab, you've remembered where I couldn't. Well done!
I think I may have been arguing on the wrong side of that one but I don't remember the details.
I don't recall that we argued about visualization, but numbers where there was a clear correlation. I was of the opinion that the authors slightly overstated in language what there numbers show. You found the correlation strong, and they used a similar term.
I'm sure the study is linked in this thread a year or so ago.
Ziryab wrote:
Optimissed wrote:
mpaetz wrote:
Optimissed wrote:
It suggests that there may be two more factors of equal importance, although that's only a possibility.
A few years ago we identified a few variables such as ability to learn, concentrate and memorise. I would suggest that those are of roughly equal weight and that possibly, IQ could be the major factor. However, I would prefer to go for ability to concentrate and focus.
Accurate, fast calculation requires excellent short term memory but long term is also important, as well as enough but not too much practice.
Can you think of anything else which is a relevant variable?
Yes: Visualization--the ability to form and hold an accurate picture in one's mind of what the position will look like following a contemplated series of moves. Clearly this will enable a player to make a better judgement of the value of any proposed line of play.
Visual memory--recognizing positions from games/positions played or studied in the past and applying analysis therefrom to similar positions in a game in progress.
I purposely left that out because there seems to be new evidence supporting the idea that visualisation is not such an important skill as was thought. If you recall, I supported visualisation a couple of years back but I think I'm reconsidering. What do you think?
When I calculate, I don't do it visually but rather by building up an abstract pattern. I think that may be more memory based but I'm not sure. Just been walking for 4 hours so am a bit too tired to play chess and notice my thought patterns at the same time.
A study that several of us have referenced (you and I argued about whether it found a "modest" or "strong" correlation) did note that, surprisingly, the correlation between chess skill and visualization as measured by the test was weak.
I think that most people assume it should be strong and the researchers indicated that they had done so.
That's interesting. I suppose visualization is emphasized in calculating long lines and visualizing how the position ends up at the end of the line. Many top players, including Magnus, have stated that it is much more important to develop intuitive chess understanding, being able to evaluate a position and maybe see a couple moves ahead, rather than being able to do rigorous calculation of long lines. There are different styles though, of course.
Optimissed wrote:
Kotshmot wrote:
It's an interesting topic. Whenever the relationship of IQ and a given subject is brought up and discussed, it tends to get a pretty strong reaction from people. Some of it, I think, has to do with people interpreting the word correlation too strongly, as if it was implied that something cannot be achieved without a certain level of intelligence. In reality that rarely the suggestion (atleast when talking about above average levels), but the question of how useful is intelligence in this particular subject, or if there are other reasons that may lead to a correlation. I think in terms of chess it's both, intelligence is useful but intelligent people also tend to be more interested in problem solving. This isn't to say there aren't other factors that are more significant.
To me the relationship appears to have some similarities to one of intelligence and level/field of education. There is some correlation (I havent done any research, dont know how much) that may be explained by intelligence being a useful asset in studying, but also theres the motivational factors. A masters degree may allow one to be highly skilled and knowledgeable about what is going on in their field and apply their knowledge in pratice, where as a doctorate degree will force them to ask more fundamental questions on why and how, providing them more tools for generating new knowledge. A highly generalized statement, but what I'm aiming at is that an intelligent person might be more likely to be asking those fundamental questions and not be satisfied without that knowledge. Does that mean that a person of a slightly below average iq couldn't achieve a doctorate degree? No, it is certainly possible. More challenging and less likely, yes.
Both education and iq are influenced with (almost) unlimited other factors, external and internal, not the least of which is culture. If one would like to truly measure the role of iq in chess performance, the study would need control groups with standardized studying resources and maybe even backgrounds in more detail.
Some already got into the questions of what do the iq tests truly even capture. That's a whole another topic but they do capture something when the sample size is appropriately big. When an individual person says their iq is either 109 or 125, I'd say it's unclear what conclusions you can draw out of those numbers.
That's an interesting comment. I'm aware that you can have a research Masters degree or a wholly taught one. The taught ones will usually contain a small element of basic research work.
Regarding doctorates, my son did a theoretical physics PhD after his MMath and he told me that everyone who works diligently is almost sure to get the PhD. That is because they do tailor the difficulty of the research project according to the ability of the candidate. He also told me that less able candidates were tending to opt for cosmology-based research projects. Why? Because there's much less positive feedback in cosmology (positive meaning evidence-based) and so many of the checks and balances are evaded or avoided.
His project was to depict magnetism in terms of fermionic spins. He found it intensely difficult even though he was the equal best mathematician at St Andrews (equal with a female German candidate) and also because some of the criticism of his work was due to the assessors' lack of understanding of the subject, which was cutting edge at the time.
I think he would tell you that maybe a large majority of PhDs do not imply being able to think fundamentally or to break new ground. I've talked to some physics PhDs and found them not particularly able to understand what I'm talking about. That means they were not over-bright.
IQ is, more or less, a measure of mental clarity, mental fluidity and the ability to apply that to diverse problems, with time being of the essence. This means that to get a high IQ score, you have to apply it and get the appropriate answers VERY fast, which means total self-confidence. And some people wonder why I have a lot of self-confidence!
Interesting anecdote. One thing to consider is that there are levels above levels to fundemental thinking. If your son happens to be a genius, he might stand out among impressive people. That is not to say PhD's on average aren't more advanced thinkers than the rest. Confidence by the way is a very good and underestimated aspect of cognitive performance. A highly intelligent person wont be able to access half of their potential if they (severely) lack confidence.
Along with things like art so while there could be a correlation between better cognitive thinking it's not really the case it's basically learning a skill like an instrument and learning art or a sport hobbies are a good thing