It's funny because he claims to be so smart just because he's white, but he did a really stupid thing by gambling in the stock market.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731146)

Nobody owes you an intellectual response you kooky fag. Your dumb and poor.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731738)

Date: May 15th, 2014 9:42 PM

Author: """''""''''

give it 10 years

(http://www.autoadmit.com/thread.php?thread_id=2568609&forum_id=2#25570609)

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731842)

and that he has experience regression to the mean with his children and financial portfolio.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731189)

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731382)

You probably don't believe in efficient markets either.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731333)

I'll throw $5 in the kitty for today's fucklaw screeds.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731355)

http://www.nature.com/scitable/topicpage/estimating-trait-heritability-46889

My main problem with this statistical model is that you just are not going to get good estimates if you regress blacks, whites, Asians, Jews, and the Darwin-Wedgwood family to the same place. I've said this before.

I am aware that you can make charts as in figure 1 in the link and draw a regression line to get estimates of regression to the mean, where the heritability coefficient is related to the slope of the line. Sure. I'm not sure what IQ data looks like plotted like that because I've never actually run across it (this is what I was inquiring about in the beginning), but I suspect that a single line is not necessarily the best estimate--this is basically a another way of saying that we should be careful about applying the same regression to every group.

Similarly, the regression estimates might be much more accurate in the middle of the distribution than at the ends. Remember that we're talking about some pretty extreme outliers here, and extending estimates that work in the middle to the extremes is not necessarily going to be valid. I haven't seen actual data on this, but I would be careful about applying that kind of a simple estimate on the ends of a distribution.

Separately, I think a lot of people refer to regression to the mean when they have information on only one parent. At least when we're talking about casual conversation, I think a lot of people think of regression to the mean when they're actually seeing the results of non-assortative mating. I would hope that academic papers would not make this error, but this possibility exists.

Anyway, let's say you have data that doesn't have these problems. You've gotten a good sample of 1%ers, and the linear model holds up at the end of the distribution. Ok. I don't know whether the data actually supports this or not (I'd like to see it), but let's just assume that we do. I accept the use of the heritability coefficient to calculate a reasonable estimate of regression to the mean in that case. (Obviously if you have this data you don't need the regression to make an estimate--just take a bin around 137 and find the average of the offspring--but let's say that all you have is the heritability coefficient: you don't have the data itself but you can be assured that it satisfies the conditions here.)

If this is the case, do I still have any objections? Sort of. My objection then is not to the use of a coefficient of heritability--instead I doubt that 0.6 is a good guess of what the coefficient actually is. I'm basing this doubt on personal experience with smart families and on some sources I've seen showing substantially higher coefficients of heritability.

My intuitive sense is that 137 to 130 or 132 would be reasonable, but I have a pretty hard time believing 137 to 122--hence the previous request for data on this. I also suspect that the calculated expectation of the regression to the mean is going to be weighed down by some retards, so the more typical regression to the mean--and the amount of regression we would most commonly see--is going to be lower than the calculations would indicate.

I would not assume that the Terman data is not properly normalized without more information about the study. Steven Hsu admits that he doesn't know whether it is or not in the original link as well. I'm inclined to believe that the original study was appropriately performed and not second-guess the original study's authors since the results sound quite plausible to me.

Early on in this discussion (last year) I admit that I wasn't particularly clear about my thoughts, and I was rusty on how the coefficient of heritability is calculated. I think this should clear things up.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731389)

Also just as a common sense test: if you're right, smart people would breed into ULTRA GENIUSES over the course of a few generations. Think about it.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731459)

fucklaw is 180

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731362)

fucklaw is annoying because he doesn't really want to discuss this shit on an intellectual level. he's too dumb. It's just a thin veneer over angry racism.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731539)

Also even if fucklaw was 100% right, what action could we take?

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731569)

to be fair, there probably aren't a lot of people who are "intellectuals" that are into "race realism"

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26731553)

- bizarro j. Edgar Hoover

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#26736288)

and he became obsessive over all this after joining the funemployed a few years back. prior to that he was just kind of a weird faggot that posted naked pics of his wife, and has posted since at least 2007 iirc.

(http://www.autoadmit.com/thread.php?thread_id=2728362&forum_id=2#28587422)