Single Status Update
Congrats on the IPREDICT thread - very interesting.
I wonder if you could improve profit performance by excluding longer odds selections. For example what are your returns if you exclude any matches where your estimated chance of the true odds is above a certain level.
I was calculating profits on all matches over the last few months where your prediction was 2.10 or less true chance. That was returning a 15% edge whereas your overall profit figure was negative in same period.
Of course this is a bit of "data mining" but I wonder if the more likely selections are also more likely to produce profit.
Yours in sport,