Can data confirm/refute this bit of racetrack lore?

[Milkshaker]

Back in the day, an old-timer told me he believed favorites were always slightly (but consistently) overlaid in the last race of the day (on the theory that bettors trying to salvage the day would, as a group, bet a touch more on long shots).

This always seemed plausible to me, even though I lack the skills to prove it via data.

So I'm wondering if this theory is (or ever was) valid statistically.

Or, perhaps more relevantly, has this axiom changed since the advent of simulcasting--when there is always "another race" from somewhere to bail out on (and the "last race" at a live track is no longer so significant)?

[Stillriledup]

I've heard that too. With most money coming from offtrack these days, you're right, its not most people's last race.

[Augenj]

You know, I've tried to explain this in another thread but favorite stats at one track are not the same as another track's favorite stats. You just can't lump favorites across all tracks in one big bucket and hope to come to a generalization that favorites always return 32 or 33 per cent each year no matter what.

So, regarding your question, you need to know what the favorites win per cent at a certain track is in order to compare the last race favorite win per cent. And do you compare last years favorite win per cent or do you look at two years ago? Believe me, they can be radically different even though when you average 20 different tracks together, the result will be the same useless (for handicapping) 32 or 33 per cent. Sorry, I guess this sounds like a rant but the dogmatism elsewhere got me in a rant mode. :D

[ultracapper]

Back in the day at Longacres you could always get a great price on the favorite the last race of the day. Longacres ran 10 races a day Wed-Sun, big allowance races for the feature 9th Wed-Fri, and a stake/hcp for the feature on the weekends. The 10th was a "hangover" race, drunk with your winnings from the day, or sick from your losses, and everyone was trying to make a killing in the 10th.

Data on overlays is a tough thing to come by because an opinion is needed for the value attached. "True" value of a horse in a race is a difficult thing to assess, and I couldn't even imagine where you could find that kind of info. You could just use averages, but that's not telling you whether they're overlaid or not. The average favorite may pay more in the last, but that doesn't tell you if it's an overlay or not.

[davew]

Back in the day at Longacres you could always get a great price on the favorite the last race of the day. Longacres ran 10 races a day Wed-Sun, big allowance races for the feature 9th Wed-Fri, and a stake/hcp for the feature on the weekends. The 10th was a "hangover" race, drunk with your winnings from the day, or sick from your losses, and everyone was trying to make a killing in the 10th.

Data on overlays is a tough thing to come by because an opinion is needed for the value attached. "True" value of a horse in a race is a difficult thing to assess, and I couldn't even imagine where you could find that kind of info. You could just use averages, but that's not telling you whether they're overlaid or not. The average favorite may pay more in the last, but that doesn't tell you if it's an overlay or not.

you could compare ROI on favorite on last race vs ROI on all favorites

[ultracapper]

Sure, but that still doesn't determine whether they were overlaid or not. Just determines that the favorites are paying better. If they are weaker favorites, they should pay better, therefore, they aren't overlays.

[thaskalos]

In today's full-card simulcasting environment...which race is the last race of the day? Plus...with the volume of races that we have available to us today...the average horseplayer is probably stuck for more money by the end of the day than he could possibly recover with a win wager. He'd probably need to cash a trifecta...

[BIG49010]

I believe William L. Quirin introduced this concept in one of his books, and I believe it was true back in the days before simulcasting, etc.. Today with another race from somewhere in 10 minutes, I doubt there is the need to get even on the last race of the day.

[Robert Goren]

I believe William L. Quirin introduced this concept in one of his books, and I believe it was true back in the days before simulcasting, etc.. Today with another race from somewhere in 10 minutes, I doubt there is the need to get even on the last race of the day.
The theory was around before Quirin. It predates me. I started betting in 1964. I think it may have been true at one time. The theory was that losing bettors for the day was chasing longshot. I think it isn't anymore. The chasers now bet exacta, tris and supers. You might catch a good price with a favorite in one of those pool though.