Risks and Rewards
Posted by Sachin Garg on 30th November 2005 | Permanent Link
After reading the very refreshing article pointed by Mark Nelson where John Allen Paulos takes us on a tour starting with complexity and ending with incompleteness, I went on to read some more from the archive.
I came across this more recent article on Risks and Rewards and how humans percieve it, here John explains how we usually choose the more certain scenario even when mathematical probability of two events is same.
It made me wonder if human bias towards more certain options has limited the optimizations we apply to improve PPM models for predicting the next symbols. Has our bias towards more certian models reduced the risks our algorithms take and thus reduced the rewards we get?