I was reading over this thread and recalled a discussion I had a with friend regarding chaos races. My friend noticed less predictable results when handicapping a race where entrants had not run to PAR. I remember doing a brief test about 10 years ago after our discussion. What I did was compare my projected figs to the race PAR. In my case it was the Beyer PAR. Some time has passed so I don't have the actual test data but I do recall seeing more unusual results in those races.
Has anyone conducted any similar tests? — Tony Kofalt
Every study I have ever done indicates that no variant outperforms any variant. — Dave Schwartz