The Racing Betting Market Gets Sharper, Even with Bombs

Great post I saw today (h/t to @dennycaps) about the favorite-longshot bias. The author looked at over 32,000 races run in Australia and New Zealand, and their expected wins by odds level at the exchange. 

It found what's historically always been found about the favorite-longshot bias in racing literature - longshots are terrible bets. 

It's funny because coincidentally I noticed this yesterday watching Keeneland. Someone on the teevee noted they didn't mind a certain horse (I can't remember which one) who was 15-1 on the board. On the exchange the horse was 80-1. According to this study, he should've been even higher. 

Historically on the exchange I don't think it's always been like this.

When I was playing full time for almost ten years I played betfair religiously, until it was pulled in 2013 here in the Tundra. I kept my personal statistics by bet level and was lifetime ROI positive on horses over 20-1. 

This, of course, had selection bias. I built longer shot models, which clicked just enough to be profitable. I generally played longer horses who could show speed, and the dirt tracks at the time were harder and somewhat faster where speed could carry. Klein's DRF book "The Power of Early Speed" with stats from the 1990's were kind of still relevant in certain subsets. Getting these horses at 70 when they were 20 on the tote board was something good. 

Clearly today that would not work. But it doesn't work with even more longshots than yesteryear and I think this study proves it. 

Why?

I think the CAW's know that field size doesn't equal field depth, and the 9 or 10 horse fields today are peppered with some serious no-hopers. These horses are literally 1000-1 or more to win. Sharp players are fading them to more reasonable levels, and no doubt making a profit on them on the exchange. 

As @quantum_sport noted to me on twitter awhile back, and I think he's super sharp, the teams do not use these horses at all, and it's reflected in exotics. There's no alls for fourth in their super tickets. 

As usual, I think they're onto something. 

Does this mean we should never use longshots? Returning to subsets, not on your life - because these horses are not covered by the teams. 

Longshots clunk up supers. And when they do, they pay, so if you're on one you love, using them in the two-three slots on supers is gangbusters. 

Just yesterday in the 6th at Keeneland, the super paid $10k for $1. Logical horse in first, logical horse third. The two bombs in the two and four slots resulted in a monster payoff. If you liked one of those bombs and swung them 2nd, or 3rd or 4th, for a modest 10 cent ticket investment you got paid. 

Markets move and adjust over time, and our betting markets have done exactly that, exchanges or otherwise. Yes, longshots - whether it be in exotics, on the board, or even on exchanges - are terrible bets if you bet them all. Fortunately for us though we don't bet them all and if we like one, we can still take a poke, and bet a little to win a lot. 

Have a great Monday everyone.