Thoroughbred Handicapping:
Learning, earning and loving the track.

NetCapper Store
New Stuff

The Grandstand
Capper Demo
Spot Play Demo
New Features
Track Tracts
TTs Archive
Contact Info
More Books
Capper Email























To The Cappers
A Message to Users of The Capper - October 31, 2003

Effective Payback

I did a study of over 16,000 horses recently.  I was looking at something I call "effective payback."  I've always thought in the back of my mind that this was important, but I think I've been missing how important it is.

We know that, in the win pool, the track takes out around 18% of each dollar bet.  It varies from track to track, but that's a rough average.  Breakage probably adds another penny or so to that total.  But when you look at a large sample of horses and calculate your effective payback if you bet every single horse, it's around .78 (78 cents returned per dollar bet).  Why is 22% of each dollar bet not being returned to the bettor?

It's because of something called the favorite/longshot bias.  I've read about this in a couple of handicapping books, but the studies are pretty old at this point.  The idea is that favorites tend to have a higher effective payback than longshots because favorites are relatively underbet and longshots are relatively overbet.  So, if you bet even amounts on every horse, your effective payback is lower than the .81 or so you'd expect it to be from the track take because you're betting proportionally more on longshots than the public does.  (The public only bets a small fraction of its money on longshots.)

I remember unsubstantiated estimates that betting every favorite would return 90 cents on the dollar.  But it's been my impression over the last 10 years or so that favorites aren't returning anywhere near that much.  They seem to be bet down lower than ever.  The favorite win percentage still hovers around 33%, but the prices seem lower.

Overall we're dealing with an effective payback in the win pool of about .78.  Not a very good situation, really, but this is the game we're playing.  This .78 is the field we're playing on.  We have to do 28% better than average (1.00/.78 = 1.28) before we even get to break-even. 

The assumption that almost all handicappers go on is that the playing field is level.  But is it?  Wouldn't you like to handicap a situation where you could bet every horse and break even or maybe show a profit, rather than the typical situation where you start out 22 cents in the hole? 

My study looked at the range of track odds from 1/9 to 99/1 for these 16,000 horses and calculated the effective payback for each odds level.  The graph below displays the results.  The horizontal axis of this chart is Track Odds, going from 1/9 on the left to 99/1 on the right.  The vertical axis is Return On Investment.  The green line highlights a 1.00 ROI or the point at which you'd break even.:

A few conclusions stand out for me after looking at this chart:

�The effective payback on low-odds horses is better than average.  But the effective payback on favorites is, not the .90 that's been advertised.  Favorites, for some reason, actually have a lower effective payback than non-favorites at the same track odds level.  Looking at the raw data, my study showed that 1961 favorites had an effective payback of .80, only slightly better than the .78 for all horses.  The conclusion: favorites really aren't underbet, but low-odds non-favorites may be.

�If you stick to horses going off at around 6/1 to 10/1, you're dealing with a situation that's practically break-even.  Ever wonder why winning handicappers often seem to concentrate on horses around this odds level?  It may be because they've found an area that effectively offers a fair game even before handicapping.  Recommendation: try to focus on horses going off around 6/1 to 10/1.  If you can improve on random betting even a little, you're in positive territory.

�Avoid very long longshots, those going off at 35/1 or higher, unless you have a large sample telling you that you can overcome the extremely poor effective payback you suffer at those odds levels.  Overcoming an effective payback of .78 is one thing.  Overcoming a .40 effective payback is almost hopeless.

The low effective payback on 1/9 horses can probably be ignored due to a small sample size for that odds level.

Remember: effective payback is the "playing field" you're dealing with as a bettor.  Effective payback gets even more interesting when you break it down by track or race type, etc.  It looks like you can bet every horse at certain tracks during certain periods and make a profit.  That's because so many longshots are winning that the effective payback at that track is actually positive.  Whether it's possible to jump on these situations and make a profit is still an open question.

This Effective Payback Chart will be a feature in the next version of The Capper. 

Talk to you later,
Gordon Pine

Copyright �2003 NetCapper Inc. All rights reserved.

Tech Tips: Using The Capper

�You can set the performance qualifiers in the Qualified Spot Plays tab of the Setup dialog to B for Best.  For instance, if you set the Minimum LT Flat ROI to B, a qualifying spot play will only display if it is from the subcategory with the best long-term Return On Investment.  If, for the Fundamental Contenders spot play, MSW is the subcategory with the highest LT ROI, then a Fundamental Contender spot play will only show up in the Performance Qualified Spot Play report for MSW races.

�To keep The Capper running as fast as possible, keep your database lean and mean.  The best way to do this is to delete your race data as soon as possible.  After you've downloaded, imported and recalculated the binary charts for a particular day, you can delete the old race data from that day.  Here's what I do: every day, before I download yesterday's binaries and today's race files, I go into the Delete Race Data tab of the Utilities dialog and use the calendar to delete all race data from the day before yesterday and before.  Then I quit The Capper, start the Database Utility and hit the Compact/Repair button.  That compacts the database and helps speed it up significantly.  Then, when I download and Import, Recalculate and Print New Races, everything goes really fast.  With a relatively new computer, each race file will generally import in about half a second.  Also, note that I pretty much never delete model data (using the Delete Model Data tab) -- only race data.

�If you have a reasonably fast computer, it's a good idea to download all 1 Month Surface Model files whenever you're downloading race files.  That way, when you import and recalculate, the TSDs (Track/Surface/Distance adjustments) will be as current as possible.  A good alternative is just to download the 1 Month Surface Model files for the tracks you're handicapping each day, and then, once a week or so, download all the tracks.  This will ensure that shippers from tracks you don't normally handicapped will be judged properly by The Capper.

Copyright �2003 NetCapper Inc. All rights reserved.
Information in this document is subject to change without notice.  Other products and companies referred to herein are trademarks or registered trademarks of their respective companies or mark holders.

Back to Top