4 ms·

But Taleb is specifically arguing against simple linear correlation measures like R^2. (See figure 1 in Taleb's article.) Feels like you haven't grasped the spe

by tw1010 7y ago

But Taleb is specifically arguing against simple linear correlation measures like R^2. (See figure 1 in Taleb's article.) Feels like you haven't grasped the specifics of his argument without taking that into consideration.

cheez 7y ago
Not only that but the best argument shows almost (but not quite) marginal contribution.
1000units 7y ago
Taleb's IQ is simply too high.
vipa123 7y ago
I don't understand. Are you saying the author shouldn't point out these errors, given what Taleb is arguing against? Or they aren't actually errors, given that the author hasn't grasped the specifics of Taleb's argument?
- tw1010 7y ago
  Taleb argues against R^2, and the author of this article seems to just ignore those points. I think Taleb has valid points, but the author here seems to be doing just the same conventional analysis that people writing psychology papers tend to do, which Taleb specifically argues is an invalid way of doing things (because assumptions built into linear correlation measures doesn't apply in this situation).
  - jamesmadison66 7y ago
    This is a really frustrating issue with criticism/response to Taleb. Taleb: "There are problems with A because of B reasons" Response: "B reasons prove A" I really don't understand why this is so common. I don't fully agree w/ Taleb that the academic demographic he tends to pick on is so lost to navel-gazing that it is completely unable to understand the context of his arguments, but it's really odd how common this is. To clarify, I'd love to see someone who actually understands Taleb's arguments to push back at him from the same premise. He says some really controversial (and potentially hugely important) stuff that needs to be challenged and peer-reviewed just to get a basic second opinion on it. This has yet to happen really because of what I described.
    - tompccs 7y ago
      I used to think Taleb was too harsh on academics. Seeing how academics respond to his well-worked critiques, I now think he's not harsh enough.
      - mcguire 7y ago
        His critique here isn't particularly well-worked. But the author here doesn't seem to have read his post....
spamizbad 7y ago
This really needs to be the top post. It's increasingly acknowledged, even in the "soft" sciences, that r-squared is not an appropriate analysis tool. Anyway, wasn't the IQ test originally created to determine whether or not someone was mentally disabled? It seems like with more modern statistical methods it still could be useful measurement for that purpose, even if it doesn't accurately "scale" beyond 80 points. Ultimately, if you just want to answer the question "Does this person have significant cognitive deficits?" it still has its place even if it's not a good measure for broader intelligence.
- reasonaway 7y ago
  but it does scale beyond 80 points. why ignore all the available evidence and trust instead the Taleb's evidence free assumptions?
zazaraka 7y ago
Exactly. For example recently Taleb made a lot of fun of these linear fits, which are pretty ridiculous even for a non-statistician like me: https://twitter.com/aedmans/status/1147494932183179264 https://twitter.com/aedmans/status/1147494932183179264

by tw1010 7y ago

cheez 7y ago

Not only that but the best argument shows almost (but not quite) marginal contribution.

1000units 7y ago

Taleb's IQ is simply too high.

vipa123 7y ago

I don't understand. Are you saying the author shouldn't point out these errors, given what Taleb is arguing against? Or they aren't actually errors, given that the author hasn't grasped the specifics of Taleb's argument?

tw1010 7y ago
Taleb argues against R^2, and the author of this article seems to just ignore those points. I think Taleb has valid points, but the author here seems to be doing just the same conventional analysis that people writing psychology papers tend to do, which Taleb specifically argues is an invalid way of doing things (because assumptions built into linear correlation measures doesn't apply in this situation).
- jamesmadison66 7y ago
  This is a really frustrating issue with criticism/response to Taleb. Taleb: "There are problems with A because of B reasons" Response: "B reasons prove A" I really don't understand why this is so common. I don't fully agree w/ Taleb that the academic demographic he tends to pick on is so lost to navel-gazing that it is completely unable to understand the context of his arguments, but it's really odd how common this is. To clarify, I'd love to see someone who actually understands Taleb's arguments to push back at him from the same premise. He says some really controversial (and potentially hugely important) stuff that needs to be challenged and peer-reviewed just to get a basic second opinion on it. This has yet to happen really because of what I described.
  - tompccs 7y ago
    I used to think Taleb was too harsh on academics. Seeing how academics respond to his well-worked critiques, I now think he's not harsh enough.
    - mcguire 7y ago
      His critique here isn't particularly well-worked. But the author here doesn't seem to have read his post....

spamizbad 7y ago

This really needs to be the top post. It's increasingly acknowledged, even in the "soft" sciences, that r-squared is not an appropriate analysis tool. Anyway, wasn't the IQ test originally created to determine whether or not someone was mentally disabled? It seems like with more modern statistical methods it still could be useful measurement for that purpose, even if it doesn't accurately "scale" beyond 80 points. Ultimately, if you just want to answer the question "Does this person have significant cognitive deficits?" it still has its place even if it's not a good measure for broader intelligence.

reasonaway 7y ago
but it does scale beyond 80 points. why ignore all the available evidence and trust instead the Taleb's evidence free assumptions?

zazaraka 7y ago

Exactly. For example recently Taleb made a lot of fun of these linear fits, which are pretty ridiculous even for a non-statistician like me: https://twitter.com/aedmans/status/1147494932183179264 https://twitter.com/aedmans/status/1147494932183179264