Yes, you should develop a number of distinct handicapping methodologies, each race will be approached from a different angle (handicapping methodology) depending on which Factor or combination of a couple of Factors determined the Favourite in Todays race and the strength of gap over the 2nd favourite.
Then you would go about determining which of your methodologies outperforms Todays race Favourite.
This involves multivariate analysis / cluster analysis.

To use an example Moeee uses Time in his handicapping, his performance with his factor is running around the 35.2% for his top rater. Joe Public's performance is running at 34.8% for pretty much the same rated races.
So when he an the public agree he performs pretty much even, and he has to try an find overlays in a dynamic market, but when they differ and the Public's Favourite is on top either thru a Factor based on Class or Box Position he starts to have a very good edge and these are the events that become your target