both of you brought up some very good and valid points worth consideration. I have no stats on this method whatsoever, so I might just try getting some this weekend for say 1,000 races or so.

Seems to have quite a good strike rate. If you could solidly get 30% of favs winning without horror runs, and could achieve average returns above $3.30 then there may be something to this.

Lumbarsua, you're quite right. Big race trifectas always pay over the odds, in my opinion. If you can cull a great proportion of the field and leave in only what you consider to be the best chances, if you manage to snare the trifecta, you'll sure get value.

Becareful, even on country tracks this seems to work quite well, because the 30% rule always applies on averages. Therefore if you can sort out where that 30% is most likely to occur, then you've got it.
Similarly, potting the fav in lower class races is something that could be worth investigating. There is a lot of value in those races, if you overlook the favourite.

Especially on heavy tracks, when the favourite is questionable in the going. I.E. Not failed in the heavy, but no previous form on heavy tracks.