Forum for Playing Strength and Gender
Discussion of the main page [#2372]
A forum for discussing Playing Strength And Gender
I wish you would recognize that the particular presentation is what I object most. The framing of the whole page by your observations etc. - E.g. I observed how a female player left a college club of say not very communicative guys right after ten minutes, but I doubt you are willing to concede that genderized college spaces are a factor affecting participation rates (without argueing about every i dot in that sentence). I observed in other activities that sexist remarks by only one or two male participants may result in women stopping to attend, however I doubt you are willing to concede easily that sexism affects participation rates. I observed boys doing less household work than girls in male dominated societies, leaving more time for leisure activities etc. etc.
Other notes: Professional players are not self selected, but by examinations of subpopulations. E.g. with the Nihon Ki-in there is 1 place reserved for female players out of 6, sometimes but rarely one of the other qualifying spots is a women as well. I would argue a large self-selected sample (with the criterion being tournament participation) is better than two smaller samples selected by several uneven procedures (different qualifying tournaments by region and sex, association).
Top players winning tournaments are only very few. A significantly smaller sub-population (female players, western go players) is not unlikely to contain none of them.
tapir: I wish you would recognize that the particular presentation is what I object most.
Velobici: Would be interesting in hearing from you how you would like to present the material. Was hoping that the current manner of separating out the observed, assumptions, limitations, related materials would be clear and easy to read. If there is a clearer manner of presentation, would be very interested in it.
tapir: The framing of the whole page by your observations etc. -
Velobici: They are not my observations. The statements may be verified without asking me. Do you believe that any of the observations listed incorrect? Is so, which?
tapir: E.g. I observed how a female player left a college club of say not very communicative guys right after ten minutes, but I doubt you are willing to concede that genderized college spaces are a factor affecting participation rates (without argueing about every i dot in that sentence). I observed in other activities that sexist remarks by only one or two male participants may result in women stopping to attend, however I doubt you are willing to concede easily that sexism affects participation rates. I observed boys doing less household work than girls in male dominated societies, leaving more time for leisure activities etc. etc.
Velobici: The
discussion on sexism is located on the page that you, tapir, created. The
tapir: I doubt you are willing to concede easily that sexism affects participation rates.
Velobici: Why do you resort to personal remarks ? Don't believe that it helps us discuss the matter or contributes to clarifying. Please refrain from personal remarks in the future. Should I have made this error or make this error in the future, please accept my apologies and point it out so that I might correct the matter.
tapir: Professional players are not self selected, but by examinations of subpopulations. E.g. with the Nihon Ki-in there is 1 place reserved for female players out of 6, sometimes but rarely one of the other qualifying spots is a women as well.
Velobici: Agree that professional players are not a self selected population. Those that try to become professional are self selected. Club and national organization players are self selected. The general population of go players is self selected, as well. They could have chosen to spend the time they use for go in a different activity.
tapir: I would argue a large self-selected sample (with the criterion being tournament participation) is better than two smaller samples selected by several uneven procedures (different qualifying tournaments by region and sex, association).
Velobici: Are you suggesting national associations or the entire EGD population vs the professional populations of the three countries: China, Japan, and Korea ? I do not understand what course of action, or direction of discussion, you are suggesting.
tapir: Top players winning tournaments are only very few. A significantly smaller sub-population (female players, western go players) is not unlikely to contain none of them.
Velobici: The number of players winning tournaments is a comparatively small population. It appears to contain a few that win repeatedly, while others get only one title once.
(I would prefer not breaking up posts, commenting in between. It will be troublesome after some replies.)
Presentation & Observations
I recognize that you try to make the material more objective by dividing observations, limitations, assumptions etc. The point at stake is however the questions and by whom they are defined. All else depends on that, which observations are relevant, which are not... So it is not the fact that there are only three 9 dan female professionals that is disputed, that would be trivial, but we obviously do not agree on what that means. Also it is likely that the overall professional winning rate being 50% that the female professional winning rate is somewhat less, with the winning rate of the best females being quite good however, again we do not agree what this means.
What you wrote before, they way how you formulated summaries, how a page on discrimination was turned into a page on genetics left me with the strong impression that you are going to proof a lesser disposition of women towards Go no matter what arguments people give in the discussion.
That is the reason I prefer a discussion mode, which is honestly subjective and does not claim clarity before it is reached in argument.
My observations
The wild guess about your acceptance of such observations was prompted by your evasive answer here. I prefer communicating expectations once in a while to a no-nonsense, facts-only approach which does not guarantee settlement either when interpretation is at stake and usually proceeds to search for external rulings. Offense was and is not intended.
I did not create any new page regarding gender discrimination, i tried to sort out topics for discussion in different threads on an otherwise unified page - separating only the Rui Naiwei discussion which was quite independent of the rest. If this was/is a bad idea in your opinion, I wish there would have been another way of communicating it.
Self selection
We agree on that.
Sample selection
For measurement of skill by gender a more appropriate course than comparing professional population by gender would be to take a professional population of equal composition as the general go playing public. (Which is somehow hard to define, because of different intensity of play for sure.) That is e.g. if the female ratio among the public is 1:9, and among professionals 1:5. It would be reasonable to compare only the best half of female professionals with all male professionals.
Self selection - giving players of about the same intensity, who care going to a tournament - may be a better selection criterion to look for a gender difference among equally intensive players. (We know that intensity of play / spent time is a factor of strength, already. So no need to construct a sex* difference by accounting a possible difference by time spent as such.)
Top players
We agree on that and our hopes for Michael Redmond.
- I dislike the term gender here, difference of disposable time may well be a difference by gender, because women tend to have less disposable time on average. What people are supposedly concerned about on this page is not only gender but whether differences by gender (incl. all the side effects of being female in this world) are difference by sex (being female in a merely hypothetical world of equal opportunities) as well. On a second look this may be the main issue.
(Don't know how to address the numerous points you make in a single post other than by separating each on out and responding to each. Suggestions welcome.)
Presentation & Observations
tapir: The point at stake is however the questions and by whom they are defined. All else depends on that...
Velobici: Which questions we choose to ask is an important consideration. The parent page Playing Strength and Gender does not, at this time, have any questions. Have chosen not to add to that section, till we have spent some time working on the other sections.
tapir: ... by whom they are defined. All else depends on that...
Velobici: Understand, how a racist asking a question about race might be offensive to some. I am assuming that no one at Sensei's Library holds such positions and attitudes, but rather each person is seeking to understand the subject of each page without bigotry. Hope that all members or Sensei's Library are doing the same.
tapir: they way how you formulated summaries
Velobici: willing to work with you to rewrite the summaries.
tapir: how a page on discrimination was turned into a page on genetics
Velobici: Perhaps you are referring to the comments regarding Lance Armstrong and his remarkable genetic foundation for aerobic exercise greatly enhanced by his formidable determination and rigorous training schedule.
Go playing skill seems to run in some families: the three Honda sisters, the three Yamada brothers, Fujiwara with three generations of strong players two professionals, Kobayashi family, etc.
Positing that genetic endowments might not be limited to physical attributes but rather might extend to go playing ability, I commented on the parents and grandparent of Kobayashi Izumi, who married Cho U. All very skilled go players. Might have an effect on their offspring. Do you believe that genetic endowment plays no role in of go playing skill?
tapir: left me with the strong impression that you are going to proof a lesser disposition of women towards Go no matter what arguments people give in the discussion.
Velobici:
First, would prefer if you would limit your judgements of me to those based upon what I have written rather than your impression of what I might or might not write.
Second, correlation is not causation. Finding a correlation between two observations does not mean that one causes the other. Both may be caused by a third item. Each may be caused by distinct items. Please read/study the material in
My (tapir's)observations
tapir: The wild guess about your acceptance of such observations was prompted by your evasive answer http://senseis.xmp.net/?posting=7922#P7922.
Velobici: That post contains four questions and one answer. The answer was to Ian Davis' question What factor are you referring to? prompted by reading How might we rule sexism out, or discount sexism, as a factor? The answer I wrote is factor in accounting for the differences in the male and female populations, ranks, and other observable, measurable characteristics. A direct answer to Ian's question. Perhaps Ian read a factor in discrimination as opposed to a factor in creating differences. Please explain to me how this is evasive. What question was being evaded?
Sample selection
tapir: For measurement of skill by gender a more appropriate course than comparing professional population by gender would be to take a professional population of equal composition as the general go playing public.
Velobici: equal composition?? Do you mean equal numbers of individuals in each group? The mathematics of statistics makes that unnecessary. Given a sufficiently large population, a random selected sample of a population selected without bias that results in an acceptably small 95% confidence interval, there is no statistically requirement that the two population samples have an equal number of individuals.
tapir: That is e.g. if the female ratio among the public is 1:9, and among professionals 1:5. It would be reasonable to compare only the best half of female professionals with all male professionals.
Velobici: The numbers 1:9, 1:5 leading to one half...don't understand them at all. Could you explain how you came to use these values and their relationship to one half?
tapir: That is e.g. if the female ratio among the public is 1:9, and among professionals 1:5. It would be reasonable to compare only the best half of female professionals with all male professionals.
Velobici: We might compare the best half of professionals to amateurs. Its a valid question. Might be more interesting as a follow up question once we have compared a sample of go professionals to a sample of the entire go playing population. That would help us to determine how much stronger professionals, as a population, are than the non-professional go playing population. Then we might ask, what about the top half of go professionals.
tapir: Self selection - giving players of about the same intensity, who care going to a tournament - may be a better selection criterion to look for a gender difference among equally intensive players.
Velobici: We could use the EGD and AGA Go Database for this purpose.
Top players
tapir: I dislike the term gender here...difference by sex...
Velobici: This may be a generational matter. There are two genders: masculine and feminine, at least in the grammar of some languages, one is used for males, the other for females. Gender is an adjective and a noun. By contrast, sex is a noun and a verb. Have been using gender in preference to sex focus on the male/female rather than use a word that brings with it verb meaning. It
tapir: On a second look this may be the main issue.
Velobici: Oh! Did not realize that this. If the main issue is sex versus gender, we could rename the page Playing Strength and Sex. It seems that the title might be misunderstood. Playing Strength and Sex is similar to such titles as Sex and the City, Sex and the Single Girl, Everything You Wanted to Know about Sex, but where Afraid to Ask. Are you sure that Playing Strength and Sex is better than Playing Strength and Gender? Is there another word that we might use?
Velobici: Go playing skill seems to run in some families.
Well, this is the typical example why the whole discussion about correlation not being causation is lost to most people. They instantly look for causation and among different possible causes for genetics way too early. When this is pointed out, they will tell you "correlation is not causation" but they will behave as if they found a cause. (E.g. the World Bank did motivation training as development effort because they found that motivated people are the ones more likely to start a business later.)
Velobici: The numbers 1:9, 1:5 leading to one half...don't understand them at all. Could you explain how you came to use these values and their relationship to one half?
There is no way to discuss statistics, if we can't cope with this little mathematics. The female players are a part of the total population. If you subtract the half of 0.2 (1:5) from the whole (1), you have 0.9 left, as 0.1 is the half of 0.2 -> 0.1/0.9 = 1:9.
Some people never fail to comment on the female qualification tournament being easier. If that is the case, then the idea of comparing all male professionals with all female professionals to determine strength by gender is statistically nonsense. In populations of even strength the top 0.001% of one population should be expected to be better than the top 0.005% of another population.
Professionals are also not randomly selected, but supposed to be the strongest of the population. That makes it crucial that you compare according to the base populations.
AFAIK this is the same fallacy which leads Germans to conclude that federal states where only a smaller percentage is allowed to higher education, have better education system. Because their top 10% are better than the top 40% of other federal states.
More later.
tapir wrote
- Comment 1: For statistics of playing strength and gender it is not the male:female ratio of professional players which is important, but the male:female ratio among the more general Go playing population.
- Comment 2: Playing strength of female professionals != playing strength by gender, exactly because the uneven composition of the Go playing public. I.e. even if the result is, that on average female professional play slightly worse than male professionals, this does not tell anything about the question whether playing strength is influenced by gender.
- Comment 3: Statistically it is more sound to look for winning percentages/ratings than to rare title events.
- Comment 4: Even if a statistically significant difference in playing strength by gender would be found, this does not indicate any special reason for that difference.
Attempted to incorporate these into the main page through an edit of "Assumptions" and addition of "Limitations".
"removed obvious nonsense, added goratings.org data analysis," however the limitations section and supportive links are not obvious nonsense. It's hard to avoid the conclusion we should revert the last changes.
- As for the limitations section
Perhaps it's not-so-obvious nonsense, in that case let me explain why it still is.
"it would be desirable to have... of the entire.." - It would be, but that statement itself adds nothing of value and helps nobody. If it's meant as limitation, it should limit something. There is nothing to limit, ergo we can scrap it.
"correlation is not causality" - yes, but no argument was made that involved correlation or causality, so this is irrelevant here.
"statistics obtained from self-selected" - nonsense. What is "self-selected" even supposed to mean? "...may not reflect the statistics..." - before my addition, there was not a single statistic (but hey, 1 frequency: 4 female 9p) mentioned on this page. And again, since this statement is not related to anything on the page, we can scrap it as well.
"since we don't know the ratio of male professionals:male players" - irrelevant. We can only use data we have or gather data we need. No point is made that requires these ratios and no attempt is made to acquire the data, therefore this point is moot as well. The WHOLE IDEA of statistics is that you DO NOT KNOW the entire population. If you did, you would not need statistics (where you work with estimators), you would only ever need frequency analysis.
Actually, there are pretty good data available, but the previous author made no attempt to obtain said data. Laziness is no excuse and certainly doesn't qualify as "limitation" to whatever, other than (presumably) herself. There are at least 2 useful, extensive databases, EGD and AGAGD. There are numerous professional game repositories and with a bit of work you can find out the sex of the players. But yes, it's work.
- As for the "supportive links"
I only scrapped the obviously unrelated links. "The mother-daughter relationship of Korean women", some chess links... come on. Let's keep this relevant, shall we? This is senseis library, not self-help library.
- Lastly, concerning fallacious reasoning
"Since women constitute 50% of the population, 50% of all go players should be female." - False and fallacious. In statistics, a population is the entire set of entities you want to make inferences about, as opposed to the sample, which is the set of entities you can test or measure in some way. You use the sample to infer something about the population. If we see 1000 people enter a Go tournament and 100 of them are women, we have a large sample that suggests about 10% of the population of Go playing individuals is female. Our assumption here is, of course, that both males and females are equally likely to participate in tournaments of this sort. Until we find evidence to the contrary, we now have a set we might call "our best data on the sex distribution among (tournament-attending) Go players".
1) Your writing has not elucidated gender inequality any more than the original.
2) Not argument was made that involved correlation or causality - perhaps that is itself rather a limitation then??
at least it is shorter.
to answer one point by mysandwich:
"since we don't know the ratio of male professionals:male players" - irrelevant. We can only use data we have or gather data we need. No point is made that requires these ratios and no attempt is made to acquire the data, therefore this point is moot as well.
the point was made time and time again. in this thread, on sl (discrimination in go is the original page) and elsewhere:
"If you take a random 10% of Go players, the top 100 of these 10% will be weaker on average than the top 100 of the rest of the population. This may not explain everything, but before diving into venus & mars or east & west type of explanation, you definitely should account for the different size of player populations first and figure out whether any significant difference to be explained remains after that."
if you don't understand why the ratio is important, you better not preach about proper statistics. further reading: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2679077/
(western go players have worse results than women, they have won 0 titles and only a few of them even qualified as professionals. yet, i don't see the same eagerness to attribute it to a difference in innate abilities.)
1. If the point was made time and time again, you should probably take the hint.
2. You can link to articles all you like, but you don't show any understanding of them, so until you make a valid point, you should refrain from talking as if you did. Do you even have a background in statistics? I majored in research methodology (including continuous multivariate statistics), how about you?
3. "If you take 10%..." It seems whoever made that point does not understand sampling very well, so I will explain it to you. If you sample 10% of all Go players -at random-, ad infinitum, the cumulative sampling distribution will approach the population distribution. That's why bootstrapping works, by the way.
Moving on. First, the sentence you quoted seems irrelevant to the discussion. Do you assume that women will never be in the top 100? What's your point? Second, you seem to be convinced that women are equally strong at Go as men are, just that the numbers are wrong and the statistics are flawed. That's what researchers like me call "bias".
4. "western go players have worse results than women" I don't know what data you're referring to, please specify. Who makes up the set of "western go players" and who makes up the set of "women"?
5. I don't know where you get the idea from that anyone claimed the (evident) difference in average strength (of at least the set of professionals I referenced) was innate. You're the first, actually.
please read, what is actually written, before answering. it helps. sometimes.
The red flag has to be raised when we see an "I majored in this course" appearing in the discussion. I think that bringing statistics into this page is not doing it any favours. You can look up a lot of papers on the subject of such disparity in Chess / Sciences / Math if you want to. There is no peer researched paper on the topic in Go. Frankly, I'd prefer we don't try to make one on this page. If an author wants to make a diatribe, sorry, well researched article, why don't they make it on this user/subpage first?
What we can do is simple list places were we see discrimination. I'd say a hairdryer being awarded as a prize in the Women's World Amateur might be a good choice there.
![[Welcome to Sensei's Library!]](images/stone-hello.png "Sensei's Library")