Creating Your Own Systems & Angles

Recently, after telling a Facebook friend that my Win Factor and Pace Profile reports were best used in conjunction with one’s own handicapping or as the basis of a system or angle, I was asked a great question: “OK, so how do I come up with an angle?”

Indeed. It is a question that has no easy answer. To me, developing a winning system — and the “winning” part should be stressed; anybody can develop a losing system — is part science, part art, part long nights and strong booze.

Let me start by giving some advice on what notto do: Don’t try to string a bunch of positive impact-value factors together. Although this sounds like a great idea, in reality it doesn’t work.

For example, according to my database studies, horses with the best last-race Brisnet speed figure have an impact value of 2.05 — meaning that they win approximately twice as often as expected (impact values are computed by dividing the percentage of winners showing a particular characteristic by the percentage of entrants possessing that characteristic). Those same studies also reveal that horses with the highest BRIS Power Rating have a 2.49 IV.

Now, according to the more ardent impact-value handicappers, combining these factors should result in an IV in the neighborhood of 5.10 (2.05 x 2.49). Others are less optimistic, but still believe that the IV will be considerably higher than 2.49 (the higher IV of the two).

The actual IV? 2.82.

To put this in perspective, race favorites (excluding entries and/or multiple top betting choices) yield a 2.76 IV. Worse, none of these impact-value angles account for the odds, or ROI. And, as I stated at the outset, that is ultimately what we are concerned with. Horses with the best last-race speed figure yield a 17.1 percent loss; animals with the best BRIS Power Rating show a 14.9 percent loss; and the two factors combined produce a 15.9 percent loss.

Of course, many impact-value handicappers have a solution to this value conundrum.

“Just insist on higher odds,” they say, presumably with an indulgent smile.

OK, let’s do that. Since horses with the highest last-race Brisnet speed figure and best BRIS Power Rating win approximately 38 percent of the time according to my studies, let’s insist on odds of 9-5, which, in theory, should result in a 5-6 percent ROI. (No point in being greedy and seeking higher odds.)

Sadly, this “fair odds” requirement didn’t produce the money tree that we’d hoped for. In fact, it produced just the opposite. Not only did we lower the impact value of these combined factors to 1.76, but we also increased the negative ROI to -19.6 percent, worse than any of the individual factors.

Naturally, this raises the question: Is there a way to stem all this red ink or am I simply trying to squash everybody’s dream of racetrack riches?

The answer is no… no, I’m not trying to squash everybody’s dream of racetrack riches, that is. There are solutions to the issues I presented. While I don’t have the space to get very in-depth, here are some general guidelines to adhere to when developing your own systems and methods:

1) Try to use unique criterion. One of the reasons that impact values don’t ascend like some handicappers expect them to when positive-IV factors are combined is because the factors overlap.

For instance, I am reasonably sure that the BRIS Power Ratings take speed into account, so combining them with the best last-race speed figure as we did above is, in effect, double-weighting speed as a factor.

That’s OK in some instances, but one should at least be aware of what they are doing.

2) Related to the above, make sure that your rules make sense and that they work together. If you think last-race form is important — I certainly do — than you absolutely, positively need to impose a date requirement on that race. After all, is it really logical to credit a horse for a great last race (however one defines that) when it was run two years ago? I think not.

3) Follow the KISS (keep-it-simple-stupid) principle whenever possible. Many a novice player goes astray by trying to make their angles/systems too complex. Unless you are using mathematical regression techniques or other advanced mathematical procedures to weight the factors you are using, stick with what you know.

None of my angles take body language into account. Why? Because nine times out of 10 I can’t spot it, much less quantify it.

4) This has been a key to my own success: Try to make your techniques as broad and universal as possible — don’t over-optimize.

You will find that you can usually unearth profitable angles in any subset of data, but in doing so you run the very great risk of what some have called the “backtest fallacy.” In other words, as you pare down the data, you wind up fitting it to your preconceived notions, as opposed to the other way around.

Hence, when you apply your “wonder system” to new data, it crumbles like Anthony Weiner’s political aspirations… again.
I hope these rules help. I’ll end this piece with a sampling of my own angle/method spot plays for Wednesday, July 31:

Performance Rating Angle Plays
Bet to win at even odds (1-1) or greater

BEU1: 2-Maestro Miss (2-1 on the morning line)
DEL1: 5-Its Looking (7-2)
DEL7: 7-Upside Down (3-1)
NP2: 5-Kool Shazoom (8-5)
NP4: 4-Passion Red (8-1)
TDN2: 7-Sky Kerridge (3-1)

Top Form Plays