When looking at the DRF, there are so many things to consider (the type/class of a race, most recent performances, surface, distance, trainer, recent workouts, trainers/riders record coming off 60+ layoff, etc.). Which has the most impact on the outcome?
A combination of the class of the race and the past performance of the horses in it. For example, in a $3,500 claiming race it is entirely possible that a Longshot will win if its past performance is better than the Favorite's (think Maiden race), so then an overlayed horse may actually be a better choice than an odds-on; while in a stakes race I would never assume that – I'd bet the Favorite to Show. I'd want the horse to have failed in the past at the wrong length or wrong surface rather than at its customary one; I'd want it to have had at least one miserable performance (without injury) that it recovered from somewhere in its racing history. I wouldn't get myself too bogged down with trainer of record in the last race unless he/she was so lousy that the owner changed, which is a sign that the owner is really trying to win and not just trying to condition the horse. Same for the jockey. A lot of times the horses with the best trainers/jockeys have their odds kicked down so low you can't win anything significant yet you can lose it all if they miss. As for surface, never bet a turf favorite to win if the race is being moved off-turf due to rain – that's because turf specialists usually have larger feet, so they bog down worse in mud. In all the stats I have analyzed it seems that Turf Stakes races involving Fillies and Mares running at 1 to 1 1/2 miles are the most erratic in the results of the odds of the winner. Workourts are meaningless as they can be rigged – a friend of mine swears that sometimes they take horses out to a secret pasture and show the serious bettors/potential buyers what a horse that "has no chance" can really do. Coming off a layoff is VERY inportant. In all my studies a horse that has run a good performance within the last 15 days has the edge over a horse with a similar but older performance; this is especially true if the oldest race happened 6-9 months ago, in which case the horse with the oldest race may come in last – I've seen it.
In my opinion the easiest general way to handicap is to pick those that finished in the top 33% of the achievement average of all the horses in their last race, then eliminate the oldest performances, then consider a solid underlayed first favorite and a horse that is a strong contender yet an unjustified-overlay (should be bet on but the public is passing – I don't mean it went from 5 to 2 up to 50 to 1!) But make sure to check the class level of past races and watch for drops/raises in class or to see if any of the horses in the particular race have run against each other – if so, unless one of them won that race, the one that finished in the money or outranked the competitor has a slight edge. The reason I don't like to bet one that won against today's competition is that it may suffer from the "bounce" – but I would put it in an Exacta/Trifecta box.
Hope this helps.