From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Sun Jun 15 2003 - 00:29:41 MDT
Olga Bourlin wrote:
> From: "Eliezer S. Yudkowsky" <sentience@pobox.com>
>
>>Olga Bourlin wrote:
>>
>>>What kind of evidence and how much evidence would you need? We can all
>>>find statistical evidence for investors and traders in the stock market
>>>- but that's not the point here. There *are* people who do well, so
>>>they must be doing "something right."
>>
>>Does not follow. There are people who win the lottery; it doesn't mean
>>that those people alone have discovered the secret pattern of winning
>>numbers. It means that a lot of people tried, and a few succeeded, at the
>>chance probability. Are there more investors who do well than chance
>>would lead you to expect?
>
> Sorry if I wasn't as clear as I could have been. Yes, that's exactly what
> I'm saying - that there are investors who do better than chance would have
> them do.
Some investors will, by chance, do better than chance would have them do.
I.e., one out of 20 investors will perform at the p < 0.05 level, one
out of 100 investors will perform at the p < 0.01 level, and so on. An
awful lot of research on this type of problem shows that people tend to
overestimate the significance of effects, to underestimate the degree to
which runs of good luck appear in random strings, to selectively recall
successes but not failures, and so on.
-- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Sun Jun 15 2003 - 00:38:51 MDT