MSGI:
You are missing the point. Let's take it to the extreme. Let's say there are only three contests. The first one has an average return of 50% (great market). The second one has an average return of 0% (average market). The third one has an average return of -50%. (horrible market).
Person A only participated 1&2. Person B only in 2&3. Person A had totals of +40% and -10% - thus, he was below average in both.
Person B had +10% and -40% - thus, he was above average in both.
The "ranking" list would show person B ahead of Person A as it should since person B was in the upper half of both contests and Person A was in the lower half of both contests.
The percentage return list, on the other hand, would wrongly show Person A ahead of Person B because Person A had an average return of +15% while Person B had an average return of -15%.
That's the fundamental flaw in using ratings (returns) when they are all over the map and not everyone participates in all contests.
Your argument is in error because you are not factoring in whether those four were in contests that were skewed up or skewed down.
As I said, there could also be a built in flaw for using ratings as opposed to rankings. That would be if the number of participants were all over the map. Say you only had 20 for one contest and 100 for another. A person who got 20th sucked in the first one, but a person who got 25th in the second one was pretty darn good.
Gilead's idea is probably the best of all worlds. I suspect I'll get rid of both rank and rate the way I have them and go to just that one in the future.
Len
News 
Market Data 
Discover 
