7 ms·

Chess statistics do not show evidence of a significant difference: https://en.chessbase.com/post/what-gender-gap-in-chess https://en.chessbase.com/post/what-ge

Chess statistics do not show evidence of a significant difference: https://en.chessbase.com/post/what-gender-gap-in-chess https://en.chessbase.com/post/what-gender-gap-in-chess

satokema 4y ago
"(6.1%) were female" This article doesn't prove the point you thought you were making.
- Jabbles 4y ago
  Your comment is so ambiguous that I'm not sure how to interpret it, but I encourage you to read the article completely, as the number you quoted is not part of the conclusion.
SamReidHughes 4y ago
That is obviously garbage statistics, isn't it? The conclusion they've drawn about the female distribution is based entirely on the performance of one female player. And they're obfuscating rank ordering with rating gap. Also, for an unexplained reason, they only consider Indian players. That's weird, isn't it? Why Indian? Let's investigate. Using the Oct 20th 2020 data, since the Oct 6th 2020 data isn't available on the FIDE website, we could look Chinese players (birthday <2000 to follow their criteria): I picked China because they're a large country and because the top active female player is Chinese. Of the top Chinese players, the ordering is 8 males, 1 female (Hou, Yifan), 12 males, 1 female (Xie, Jun), then 3 M, 1 F, 5 M, 1 F, 1 M, 1 F, 2 M, 1 F, and so on. The overall Chinese female percentage is 30.47%. Here's a table of percentage over rating threshold, with cumulative sums shown. M F M Sum F Sum FSum/(MSum+FSum) 2800 0 0 0 0 - 2700 5 0 5 0 0.00% 2600 10 1 15 1 6.25% 2500 15 4 30 5 14.29% 2400 38 10 68 15 18.07% 2300 55 10 123 25 16.89% 2200 76 31 199 56 21.96% 2100 89 40 288 96 25.00% 2000 48 29 336 125 27.11% 1900 42 17 378 142 27.31% 1800 42 25 420 167 28.45% 1700 27 15 447 182 28.93% 1600 27 11 474 193 28.94% 1500 13 14 487 207 29.83% 1400 9 2 496 209 29.65% 1300 3 4 499 213 29.92% 1200 1 3 500 216 30.17% 1100 2 1 502 217 30.18% 1000 0 3 502 220 30.47% Here's the data for India: M F M sum F sum FSum/(MSum+FSum) 2800 0 0 0 0 - 2700 3 0 3 0 0.00% 2600 10 0 13 0 0.00% 2500 16 2 29 2 6.45% 2400 38 0 67 2 2.90% 2300 68 9 135 11 7.53% 2200 150 11 285 22 7.17% 2100 321 31 606 53 8.04% 2000 556 57 1162 110 8.65% 1900 591 42 1753 152 7.98% 1800 750 66 2503 218 8.01% 1700 920 53 3423 271 7.34% 1600 1321 74 4744 345 6.78% 1500 1585 117 6329 462 6.80% 1400 2135 129 8464 591 6.53% 1300 2671 141 11135 732 6.17% 1200 2962 165 14097 897 5.98% 1100 2566 139 16663 1036 5.85% 1000 1592 139 18255 1175 6.05% What is going on there? It is worthwhile to check out India having filtered out the inactive players: M F M sum F sum 2800 0 0 0 0 - 2700 3 0 3 0 0.00% 2600 9 0 12 0 0.00% 2500 13 2 25 2 7.41% 2400 30 0 55 2 3.51% 2300 36 8 91 10 9.90% 2200 44 6 135 16 10.60% 2100 47 7 182 23 11.22% 2000 56 11 238 34 12.50% 1900 51 8 289 42 12.69% 1800 72 7 361 49 11.95% 1700 128 6 489 55 10.11% 1600 179 3 668 58 7.99% 1500 234 7 902 65 6.72% 1400 370 8 1272 73 5.43% 1300 493 10 1765 83 4.49% 1200 605 14 2370 97 3.93% 1100 513 18 2883 115 3.84% 1000 296 20 3179 135 4.07% There is a big pile of low-rated Indian men with FIDE ratings, while mediocre women aren't interested. Note that the article fallaciously compares the male average to the female average as if that means something. If we were to compute the article's bogus statistical argument about the top woman, a more appropriate female percentage to use would be more like 12.69%, the maximum on that list, before the hoard of mediocre males, instead of 6.1%. It would make more sense to base our statistics on the set of all the world's active players, instead of one woman from a particular country against an irrelevant percentage. Here is that data: M F M sum F sum Fsum/(Msum+Fsum) 2800 2 0 0 0 - 2700 32 0 32 0 0.00% 2600 186 1 218 1 0.46% 2500 437 11 655 12 1.80% 2400 1049 39 1704 51 2.91% 2300 1928 87 3632 138 3.66% 2200 3385 156 7017 294 4.02% 2100 5699 231 12716 525 3.96%
- SamReidHughes 4y ago
  Oh, and one other thing, at a meta level: The entire construction of the author's argument is to pick a statistical measure with a wide standard deviation, instead of a better one that would use more data to get a low standard deviation. Then he finds that one real-life outcome is within that wide standard deviation. Not only that, he defines it via the exponentially increasing ELO rating distribution, instead of rank order, and that would have the effect of piling up a bunch of lower ranked players together within one standard deviation of the mean. This means it would be virtually impossible to get a result 1 standard deviation out, let alone 2.
- Jabbles 4y ago
  You seem passionate about proving something but I can't tell what it is.
  - AdrianB1 4y ago
    > intelligence distribution skews more male at the higher end
    - SamReidHughes 4y ago
      Yep. In fact, something even narrower. That it would do so without any sexual dimorphism between men and women. (And I'm not even here to have this argument -- it's just a basic fact behind one of the reasons why some tech companies would skew more male than others -- but then some people decided to argue.)
  - SamReidHughes 4y ago
    Specifically here it's that the article you linked is total garbage, poorly reasoned, and that can be discerned without even checking the data. And more generally Adrian's reply.

Chess statistics do not show evidence of a significant difference: https://en.chessbase.com/post/what-gender-gap-in-chess https://en.chessbase.com/post/what-gender-gap-in-chess

satokema 4y ago

"(6.1%) were female" This article doesn't prove the point you thought you were making.

Jabbles 4y ago
Your comment is so ambiguous that I'm not sure how to interpret it, but I encourage you to read the article completely, as the number you quoted is not part of the conclusion.

SamReidHughes 4y ago

That is obviously garbage statistics, isn't it? The conclusion they've drawn about the female distribution is based entirely on the performance of one female player. And they're obfuscating rank ordering with rating gap. Also, for an unexplained reason, they only consider Indian players. That's weird, isn't it? Why Indian? Let's investigate. Using the Oct 20th 2020 data, since the Oct 6th 2020 data isn't available on the FIDE website, we could look Chinese players (birthday <2000 to follow their criteria): I picked China because they're a large country and because the top active female player is Chinese. Of the top Chinese players, the ordering is 8 males, 1 female (Hou, Yifan), 12 males, 1 female (Xie, Jun), then 3 M, 1 F, 5 M, 1 F, 1 M, 1 F, 2 M, 1 F, and so on. The overall Chinese female percentage is 30.47%. Here's a table of percentage over rating threshold, with cumulative sums shown. M F M Sum F Sum FSum/(MSum+FSum) 2800 0 0 0 0 - 2700 5 0 5 0 0.00% 2600 10 1 15 1 6.25% 2500 15 4 30 5 14.29% 2400 38 10 68 15 18.07% 2300 55 10 123 25 16.89% 2200 76 31 199 56 21.96% 2100 89 40 288 96 25.00% 2000 48 29 336 125 27.11% 1900 42 17 378 142 27.31% 1800 42 25 420 167 28.45% 1700 27 15 447 182 28.93% 1600 27 11 474 193 28.94% 1500 13 14 487 207 29.83% 1400 9 2 496 209 29.65% 1300 3 4 499 213 29.92% 1200 1 3 500 216 30.17% 1100 2 1 502 217 30.18% 1000 0 3 502 220 30.47% Here's the data for India: M F M sum F sum FSum/(MSum+FSum) 2800 0 0 0 0 - 2700 3 0 3 0 0.00% 2600 10 0 13 0 0.00% 2500 16 2 29 2 6.45% 2400 38 0 67 2 2.90% 2300 68 9 135 11 7.53% 2200 150 11 285 22 7.17% 2100 321 31 606 53 8.04% 2000 556 57 1162 110 8.65% 1900 591 42 1753 152 7.98% 1800 750 66 2503 218 8.01% 1700 920 53 3423 271 7.34% 1600 1321 74 4744 345 6.78% 1500 1585 117 6329 462 6.80% 1400 2135 129 8464 591 6.53% 1300 2671 141 11135 732 6.17% 1200 2962 165 14097 897 5.98% 1100 2566 139 16663 1036 5.85% 1000 1592 139 18255 1175 6.05% What is going on there? It is worthwhile to check out India having filtered out the inactive players: M F M sum F sum 2800 0 0 0 0 - 2700 3 0 3 0 0.00% 2600 9 0 12 0 0.00% 2500 13 2 25 2 7.41% 2400 30 0 55 2 3.51% 2300 36 8 91 10 9.90% 2200 44 6 135 16 10.60% 2100 47 7 182 23 11.22% 2000 56 11 238 34 12.50% 1900 51 8 289 42 12.69% 1800 72 7 361 49 11.95% 1700 128 6 489 55 10.11% 1600 179 3 668 58 7.99% 1500 234 7 902 65 6.72% 1400 370 8 1272 73 5.43% 1300 493 10 1765 83 4.49% 1200 605 14 2370 97 3.93% 1100 513 18 2883 115 3.84% 1000 296 20 3179 135 4.07% There is a big pile of low-rated Indian men with FIDE ratings, while mediocre women aren't interested. Note that the article fallaciously compares the male average to the female average as if that means something. If we were to compute the article's bogus statistical argument about the top woman, a more appropriate female percentage to use would be more like 12.69%, the maximum on that list, before the hoard of mediocre males, instead of 6.1%. It would make more sense to base our statistics on the set of all the world's active players, instead of one woman from a particular country against an irrelevant percentage. Here is that data: M F M sum F sum Fsum/(Msum+Fsum) 2800 2 0 0 0 - 2700 32 0 32 0 0.00% 2600 186 1 218 1 0.46% 2500 437 11 655 12 1.80% 2400 1049 39 1704 51 2.91% 2300 1928 87 3632 138 3.66% 2200 3385 156 7017 294 4.02% 2100 5699 231 12716 525 3.96%

SamReidHughes 4y ago
Oh, and one other thing, at a meta level: The entire construction of the author's argument is to pick a statistical measure with a wide standard deviation, instead of a better one that would use more data to get a low standard deviation. Then he finds that one real-life outcome is within that wide standard deviation. Not only that, he defines it via the exponentially increasing ELO rating distribution, instead of rank order, and that would have the effect of piling up a bunch of lower ranked players together within one standard deviation of the mean. This means it would be virtually impossible to get a result 1 standard deviation out, let alone 2.
Jabbles 4y ago
You seem passionate about proving something but I can't tell what it is.
- AdrianB1 4y ago
  > intelligence distribution skews more male at the higher end
  - SamReidHughes 4y ago
    Yep. In fact, something even narrower. That it would do so without any sexual dimorphism between men and women. (And I'm not even here to have this argument -- it's just a basic fact behind one of the reasons why some tech companies would skew more male than others -- but then some people decided to argue.)
- SamReidHughes 4y ago
  Specifically here it's that the article you linked is total garbage, poorly reasoned, and that can be discerned without even checking the data. And more generally Adrian's reply.