Saturday, January 24, 2009

Handicapping - Layoffs

One of the toughest things to handicap is the horse off an extended layoff. At the track I will often hear, "I will never bet a horse who is coming off a qualifier with a long layoff" or in the runners "200 days or more in an automatic pitch." Let's face it, this is often correct. In harness racing, a complaint from some is the 'he will need one' line. In thoroughbred racing, if a trainer wanted to put one over on someone it is a fine time to do it with a first time starter, or a long layoff horse with no form, because people tend to believe they have no shot.

So, is this true? Are long layoffs truly detrimental? Are they sucker bets? Should we bet the odds board? Well we have some numbers. First, courtesy our friend Ray, some from harness.

In 2008, here is some data:

Days off.....Starters......ROI

29-40 .......2496 .........0.42
41-60 .......1230 .........0.65
61+ ..........246 .........0.72


With a blind bet ROI of about 0.78, each of these layoff sections have a less than one impact value.

The number that jumps out to me is the 29-40 angle. That ROI is horrid, and about as bad as you can find anywhere. That is an ultimate fade angle. I would think that this is intuitively not surprising, however. I think that if a horse is off 29-40 days that horse must have had a fairly troublesome issue. If a horse is sick, he might be off twenty days, and be back in. He is probably 100% sound, and had a good schooler to get ready and he can still race well. Same thing with a minor foot or hock issue. But for a horse to be out over a month I think the issue is more than one, or much more serious. The horse probably could not get much work into him, and he would need a tightener, or perhaps is not the same horse at all. Regardless, these horses on face value seem to be very bad bets. The crowd is not discounting the obvious problem enough.

As time grows, so does ROI. Again, this is not surprising, if we break it down. The average handicapper hates long layoffs, and now they do factor in the fact that the horse had a serious issue. Further, I believe they are prone to think old school - "that horse will need one, and I am not touching him." When looking at it objectively, there are some good bets on long layoff horses, eventhough they do not win very much.

How can we up those ROI's, and perhaps catch a nice bet or two in pick 4's? I think there is a few ways:

1) The trainer - Statistics on trainers off layoffs in harness are not readily available. You have to know your circuit and know the guys who school long layoff horses and have them ready to go.
Noel Daley and Ross Croghan are ROI positive off 30+ day layoffs
2) The odds board - Is a horse that should not be live, live?
Today in race 5 at AQU a very long layoff horse, not overly liked by public handicappers, won at 3-1. Why was he so low? Because he was ready.
3) Warmups/Scoring - If the driver looks live on a bad form long layoff horse, there is a chance he may be live.

In thoroughbreds, this is 2008 data from a pretty large database:

As we notice, again long 200+ layoff horses do not hold up very well.

Starts........ Wins.........Win %....... ROI
1634 ......... 161 ........ 9.8% ....... 0.72


When just favorites are run, however, this jumps to 0.83 ROI:

Starts........ Wins.........Win %....... ROI
141 ......... 45 ........ 32% ....... 0.83


There is a slight bias against layoff chalk. If they are chalk, they are actually very good chalk. We have to think contrarian. The odds board is not stupid.

So I guess the lesson for me is that there are no bad bets just because the horse is off a long layoff, although the general numbers overall show this. Since the general handicapping bias is against these horses, that means the crowd is. If the general crowd all thinks something is sure to happen, you do not want to be with them all the time. With some filters (odds action; trainer data) you can weed out bad bets, and find some good ones. If you do you will be rewarded for it.

1 comment:

nazko said...

RE: Layoffs.
"With some filters (odds action; trainer data) you can weed out bad bets, and find some good ones."

Indeed! In my small universe some filters worked very well. I assume that productive filters would differ from track -to- track.

In smaller parimutuel pools "action" is a good indicator, unfortunately the odds self destruct.