Belmont 20-20

This week has been one of indulgence. My wife and I generally devote a good deal of time to staying fit. We hit the gym, play semi-competitive soccer and my wife has even begun to run in long distance races. But not this week, this week we've been watching old JAG episodes and working our way through two tubs of ice cream. Almost like a mini vacation right here at home.

I suspect the Kennedy household week of indulgence will continue right through to Saturday night because no week devoted to pleasure could be devoid of horse racing. The Belmont Stakes and undercard is as tasty as the finest fudge crackle I've been sampling. Another favorite thing of mine will come into play this weekend and that is the maiden launch of my new Belmont 20-20 system. My goal was to have a 20-20 system ready for all three Triple Crown races this year. The Derby model worked out okay without being profitable while the Preakness was just about dead on. I'm not sure what to expect from the Belmont. It's such a wild and varied race, it was unlike any race I've looked into before.

In case this is the first time you've heard of anything of this nature the 20-20 systems are statistically based handicapping systems that can help you pinpoint exactly which horses fit the historical and statistical mold of past winners. The name 20-20 is derived from the fact that my first version of the Derby system had 20 statistical factors and 20-20 is a description of perfect vision. It seemed to fit well since the goal of the system is to see the entire field ranked in order of probability, hopefully with the most likely winners towards the top. Although the model is very different the concept is similar to what you'd find at The Downey Profile or Crushing the Cup.

I have been working on the concept in the Kentucky Derby and Breeders' Cup since 2003. The first working models were published online in 2006. Since then I've dabbled around with races like the Preakness, Belmont, Epsom Derby, Santa Anita Handicap and the Woodbine Mile. My hope is to eventually have 20-20 profiles for many of the major stakes events in North America.

I do not have as much back data for the Belmont as I do for the Derby and Breeders' Cup. It was only possible to create a profile from the year 2000 and onwards. I also threw it together in a few weeks as opposed to the several years it took me for the other projects. So as I share the results here I do warn you that this is still a work in progress.

Without further ado you can check out the results of the Belmont 20-20. Just a tip to help you read the spreadsheet: Every category with a green cell earns the entrant +1, every red cell is worth -1. The cells that are left blank are worth 0.

Certain cells are blank if that stat is not applicable to that entrant. For instance factor #10 has to do with Tomlinson Distance Ratings and the ratings themselves denote that certain horses have an insufficient sample size to draw an accurate conclusion. It doesn't seem fair to award or subtract points based on inconclusive data.

The system can be used in many ways, helping you to shape your view of the race, pick winners and even have the right horses for the exotics. You may notice the column entitled "WAGER". That is simply a guideline to show me if the points assigned by the statistical system are actually worthwhile. The theory is that you could bet the designated amount to win on each entrant and that would hopefully give you a chance to make money without even thinking. In the Belmont you'd have made 219.59% from betting the system blindly and outperformed equal win wagering on all contenders by an astonishing 122%. It's not intended to be bet blindly, hopefully by applying some judgment you can beat the results of this system. It's merely an attempt to help shape your thinking.

One way I like to use the system is to focus on the horses that are "perfect" qualifiers. That is, those horses with no strikes against them. You can see from the small table on the side of the spreadsheet that there have been 11 perfect qualifiers in the Belmont not including Charitable Man this year. Those 11 horses combined for a 11-6-2-0 record in the Belmont and would have yielded a 1069% ROI.

Payouts in the Belmont have been massive. No race I've studied has had such a healthy dose of longshot winners in it's recent history. Simply betting every horse in the race since the year 2000 would have made you a 97% profit. In other words almost anyone could turn a profit in this race if they stayed away from the favorites. Still favorites have gone 9-2-3-1 during that time and are definitely useful in the exotics.

It is interesting that none of the four horses that had a chance at the Triple Crown were well regarded by the system. There is no factor written into the system that automatically eliminates Triple Crown candidates it just so happened that none of them fit the historical mold. Some lacked experience over the track, some were over the edge fitness wise and War Emblem was simply not bred for it, among other things.

Charitable Man is the horse tabbed by the system this year as the only Perfect Qualifier. He is already going to be the hot horse on the board. I can't remember this much hype around a non Triple Crown race winner in the Belmont since Empire Maker. Let's hope he does as well as Bobby Frankel's colt did.

Calvin Borel may have guaranteed victory on Mine That Bird but he certainly isn't the most obvious choice from a statistical point of view. Mine That Bird is ranked in the lower half of the field. Ironically his father Birdstone was the only horse to win the Belmont while ranked in the lower half of the field by the 20-20 system.

The system is far from infallible, especially since this is the first version of it but if anyone is interested in checking out the past results of other projects you may notice that the Derby 20-20 and Preakness 20-20 are available in the same spreadsheet as the Belmont 20-20. You can also check out the Breeders' Cup 20-20 if you like.