This Wired article about the Netflix prize provides an important hint about a valuable result of understanding human biases: Couldn’t a pure statistician have also observed the inertia in the ratings? Of course. But there are infinitely many biases, patterns, and anomalies to fish for. And in almost every case, the number-cruncher wouldn’t turn up anything. A psychologist, however, can suggest to the statisticians where to point their high-powered mathematical instruments. "It cuts out dead ends,"
yes, yes, I think you guys have stumbled upon the new secret to beating the market! We are hard wired to behave this way. And for the lucky few who can rise above our primitive constraints untold riches await...
I would be surprised if the Netflix data contain enough information to identify all relevant human biases. Unlike chromodynamics, the Netflix data are selected for ease of data collection rather than for explanatory power.
I want to clarify my above comment in light of Eliezer's article about reductionism:http://www.overcomingbias.c...
The Netflix data which consists of client's movie choices contains implicitly all the biases that influence those choices(that's my assumption).
So you don't need to program in the biases explicitly. In the same way that Eliezer's chromodynamic 747 model would contain all facts about airflow implicitly.
So if we have a perfect analyser of the data we should be able to make optimal predictions without the need to explicitly factor in biases.
I was wondering, if a perfect bayesian mind analysed the data would it discover all the human biases that are influencing/contributing to it? If we program in these biases in addition to analysing the data statistically aren't we counting the same thing twice?
Prospect Theory? Which suggests that investors would take riskier actions as a market trends downward. After all there hasn't been a run on a bank to invest into it. Or in evolutionary terms, one will run from a lion at slightest provocation, but will take their time hunting/gathering.
yes, yes, I think you guys have stumbled upon the new secret to beating the market! We are hard wired to behave this way. And for the lucky few who can rise above our primitive constraints untold riches await...
signed
your friends on Wall Street
I would be surprised if the Netflix data contain enough information to identify all relevant human biases. Unlike chromodynamics, the Netflix data are selected for ease of data collection rather than for explanatory power.
I want to clarify my above comment in light of Eliezer's article about reductionism:http://www.overcomingbias.c...
The Netflix data which consists of client's movie choices contains implicitly all the biases that influence those choices(that's my assumption).
So you don't need to program in the biases explicitly. In the same way that Eliezer's chromodynamic 747 model would contain all facts about airflow implicitly.
So if we have a perfect analyser of the data we should be able to make optimal predictions without the need to explicitly factor in biases.
Does that make sense?
Regarding the Netflix prize:
I was wondering, if a perfect bayesian mind analysed the data would it discover all the human biases that are influencing/contributing to it? If we program in these biases in addition to analysing the data statistically aren't we counting the same thing twice?
I plan to keep it infrequently mentioned, as I don't have much of an idea how increased awareness of it would affect my ability to profit from it.
Yes, please, what is that "one infrequently mentioned result"? :)
Prospect Theory? Which suggests that investors would take riskier actions as a market trends downward. After all there hasn't been a run on a bank to invest into it. Or in evolutionary terms, one will run from a lion at slightest provocation, but will take their time hunting/gathering.
What's the one infrequently-mentioned result?