Comment author: satt 26 June 2014 03:55:59AM *  2 points [-]

If I understand Taleb correctly, his objection is that if X's distribution's upper tail tends to a power law with small enough (negated) exponent α, then sample proportions of X going to the distribution's top end are inconsistent under aggregation, and suffer a bias that decreases with sample size. And since the Gini coefficient is such a measure, it has these problems.

However, Taleb & Douady give me the impression that the quantitative effect of these problems is substantial only when α is appreciably less than 2. (The sole graphical example for which T&D mention a specific α, their figure 1, uses α = 1.1). But I have a hard time seeing how α can really be that small for income & wealth, because that'd imply mean income & mean wealth aren't well-defined in the population, which must be false because no one actually has, or is earning, infinitely many dollars or euros.

[Edit after E_N's response: changed "a bias that rises with sample size" to "a bias that decreases with sample size", I got that the wrong way round.]

Comment author: Eugine_Nier 27 June 2014 01:36:27AM 0 points [-]

But I have a hard time seeing how α can really be that small for income & wealth, because that'd imply mean income & mean wealth aren't well-defined in the population,

Um no. They're not well defined over the distribution, they will certainly be well defined over a finite population.

which must be false because no one actually has, or is earning, infinitely many dollars or euros.

You seem to be confused about how distributions with infinite means work. Here's a good exercise: get some coins and flip them to obtain data in a St. Petersburg distribution notice that even though the distribution has infinite mean all your data points are still finite (and quite small).

Comment author: Lumifer 24 June 2014 05:43:55PM *  1 point [-]

Well, the US forces actually attempted not to rig them.

No need to, the locals can do everything necessary. The US forces just provided the money and prevented the "undesirables" from playing.

Whichever laws you invoked when you said implied that "old-style colonialism won't work in our time" is a reasonable hypothesis.

I did not invoke any laws of nature. I think that in the current social, political, informational, military, etc. global environment the old-style colonialism is highly unlikely to work. No laws of nature are involved in this assertion.

Comment author: Eugine_Nier 26 June 2014 04:23:11AM 4 points [-]

I think that in the current social, political, informational, military, etc. global environment the old-style colonialism is highly unlikely to work.

Can you be specific about what you think is the relevant change?

Comment author: Daniel_Burfoot 24 June 2014 01:33:06AM 3 points [-]

People always mention the Butterfly Effect as if it is an unmitigated disaster for humans: it's kind of like Nature saying FU, you are never going to be able to predict me.

And it's true that the Butterfly Effect makes it very hard to make good predictions about the weather. But it also has an upside, because it means if we somehow do figure out how to make good predictions, than we should also be able to easily control the weather. If a butterfly flapping its wings in Tokyo can cause a tornado in Kansas, and we know this, then we should be able to prevent the tornado in Kansas by having another butterfly flap its wings in Osaka (or whatever).

Comment author: Eugine_Nier 26 June 2014 04:20:44AM 2 points [-]

The problem is that the weather also has many variables. While the chaos implies control principal works for low dimensional chaotic systems, e.g., the three body problem in orbital dynamics, I'm not sure how well it would work for weather.

Comment author: Ander 26 June 2014 12:11:25AM 1 point [-]

Excellent post!

Regarding ways that the questions might be disanalogous: For temperature data, I don't think that many people would question the data, average temperatures seem like good, hard facts to me. But some people might question unemployment data that they were presented with, stating that the measure of unemployment is flawed because it only measures people actively looking for work who are still eligible to receive unemployment. Some people 'fall off' and just become long term unemployed that no longer get counted in the statistics.

Perhaps you might note that some measure of 'percentage of population that is employed (adjusting for demographics changes)' would work better as the 'data' for some people?

Also, the post made me realize that in both of these two cases, the belief that I actually have (agree with both hypotheses), were formed due to the theory, and not due to looking at any empirical data. That is, when I hear empirical data in support of climate change, I think: 'well, obviously!', not 'here is the data that should be strengthening my belief in climate change'. I also realize that I haven't investigated and seen any data either way regarding whether minimum wages really do increase unemployment or not, and maybe I should do that.

Comment author: Eugine_Nier 26 June 2014 02:26:04AM *  -2 points [-]

For temperature data, I don't think that many people would question the data, average temperatures seem like good, hard facts to me.

Not really. The problem is we don't have uniformly spaced weather stations all over the earth. Furthermore the locations of the stations we do have tend to change over the time period of interest. (The various proxies suffer from similar problems.) Thus it's necessary to apply weights to the data we do have to correct for this. Unfortunately, the weights are semi-arbitrary in practice and as we learned from the leaked climategate e-mails frequently have the warming built in.

That is, when I hear empirical data in support of climate change, I think: 'well, obviously!', not 'here is the data that should be strengthening my belief in climate change'.

What's your reaction to the data that shows a lack of warming over the past 17-years?

Comment author: Suryc11 25 June 2014 08:07:15PM 1 point [-]

Hm, I'm confused. I agree that at least part of the disagreement was over Arthur's willingness to lie for his cause, but how is that not captured by Viliam_Bur's post?

Lying for a cause or otherwise playing "dirty" to win for your cause, as Arthur seemed to be advocating, seems to straightforwardly line up with Viliam_Bur's theory about "Nice Greens", "Nasty Greens", "Nice Blues" and "Nasty Blues"; specifically, in this theory, Arthur would be a "Nasty" player on the side of progress/civilization/neoliberalism-ish/etc. and Yvain would be a "Nice" player on the same side.

I guess I'm not sure what you mean by tone?

Comment author: Eugine_Nier 26 June 2014 02:07:17AM 5 points [-]

The nastiness Viliam talks about is mostly in the form of trolling or making insulting statements with little semantic content. Lying of the type Arthur advocates goes beyond that since it injects false statements into the discussion and tends to result in one's side filling up with people who believe said lies and thus willing to lie further.

Comment author: TheAncientGeek 25 June 2014 12:42:09PM *  1 point [-]

Europe is more equal on empirical measures such as the Gini Coefficient.

The comment in Talebs aphorisms does not refute that, because it is not evidence based. Instead, Taleb is making some sort of circular, ideology driven argument...that Europe is "socialist" and under "socialism" the state runs everything., therefore no healthy competition, therefore stasis..but no. in Socially Democratic Europe, the government does not intervene in the boardroom.

What's more, the empirical evidence actually contradicts Talebs untested expectation:

"according to the latest Global 500 CEO Departures™ study by global public relations firm Weber Shandwick, departing European chief executives were also more likely to be forced out of office than North American and Asia Pacific CEOs during this 2007 time period."

http://www.reputationrx.com/Default.aspx/CEOTURNOVER/GLOBAL500CEODEPARTURES%E2%84%A2andCEODEPARTURESSTUDY%E2%84%A2

Comment author: Eugine_Nier 26 June 2014 02:00:34AM 2 points [-]

Europe is more equal on empirical measures such as the Gini Coefficient.

Here is Taleb's paper about the problems with measures like the Gini Coefficient.

Comment author: Suryc11 24 June 2014 07:58:38PM 2 points [-]

Relevant SMBC. It illustrates my political theory that in every political conflict that seems to be between Greens and Blues, there are actually four sides of the conflict, let's call the "Nice Greens", "Nasty Greens", "Nice Blues" and "Nasty Blues". And there is more than one line of conflict.

Officially, "Nice Greens" + "Nasty Greens" and "Nice Blues" + "Nasty Blues" should be the only existing coalitions. But there is also the value of niceness, which somehow connects "Nice Greens" and "Nice Blues", and puts them into often unconscious opposition against the "Nasty Greens" and "Nasty Blues". Being nasty is a personality trait... for a "Nasty Green" it is often easier to become a "Nasty Blue" (different slogans, but generally the same behavior) than a "Nice Green" (different everyday behavior both among the enemies and the allies).

Yup, and that is almost exactly what (at least part of) the relatively recent disagreement between Yvain and Arthur Chu was about. See http://www.patheos.com/blogs/hallq/2014/02/on-some-criticism-of-lesswrong/ and http://slatestarcodex.com/2014/02/23/in-favor-of-niceness-community-and-civilization/

Comment author: Eugine_Nier 25 June 2014 03:56:14AM 4 points [-]

The disagreement wasn't just about tone. It was about Arthur Chu's willingness to lie for his cause. The only reason it appeared to be mostly about tone is that Yvain didn't make the strongest argument he could have.

Comment author: James_Miller 24 June 2014 05:05:20PM 3 points [-]

If the falling price of gene sequencing lets us determine a lot about how genes influence human behavior social scientists, I predict, will get a lot better at figuring out the causal effects of social programs.

Comment author: Eugine_Nier 25 June 2014 02:26:35AM 6 points [-]

Once social scientists get past their taboo against genetic explanations.

Comment author: TheAncientGeek 24 June 2014 08:29:38AM -2 points [-]

Europe is also more equal than the US. The counterargument put forward to that is that the Iron Curtain countries were not particularly egalitarian. My countercounterargument is that social democracy is not commensurable with single party state socialism.

Comment author: Eugine_Nier 25 June 2014 01:44:36AM *  2 points [-]

Europe is also more equal than the US.

That is precisely the claim being disputed. In particular, as Taleb points out in the document I quoted in the great-grandparent, when you stop trying to use static measures of inequality and instead base it on the amount of turnover at the top, you see that Europe is much more unequal (almost an oligarchy) than the US.

Comment author: Lumifer 23 June 2014 05:16:38PM 1 point [-]

They held elections and put the people who got majority into positions of power.

Elections are no big deal. Mugabe holds elections, Putin holds elections, hey, even Assad recently held elections.

Old-style colonialism wasn't based on crushing military superiority

Yes, it was. Certainly, it wasn't just military superiority, especially once the colonies were established, and the British, for example, became masters of control through political and financial means as well. However the military strength was the underlying bedrock.

Are you saying that the laws of nature somehow changed over the past century?

Which particular laws of nature do you have in mind?

Comment author: Eugine_Nier 24 June 2014 02:26:29AM 1 point [-]

Elections are no big deal. Mugabe holds elections, Putin holds elections, hey, even Assad recently held elections.

Well, the US forces actually attempted not to rig them.

However the military strength was the underlying bedrock.

Disagree. Military strength was based on a bedrock of competent management.

Are you saying that the laws of nature somehow changed over the past century?

Which particular laws of nature do you have in mind?

Whichever laws you invoked when you said implied that "old-style colonialism won't work in our time" is a reasonable hypothesis.

View more: Prev | Next