Cliff's Northfield

The answer is to somehow quantify the relationships between the post positions. It would be helpful to know statistically how much a horse's chances have improved from a move from, let's say pp4 to pp1, or how much the change from pp2 to pp6 will adversely affect that entry. Printed every day in your program is the raw data concerning number of starts at each post position, number of wins and in the money finishes. There is also the same data in percentage format. I don't know about you, but I can stare at that matrix of numbers all day long and still not really grasp the direct numeric relationship between the post positions. Sure, it's obvious that it's better to start from pp1 than pp8, but how much better? That is the question.

Let's put this data in a different format. By the way, I can't take credit for this method. I'm sure it must have been around for a long time. Another handicapper demonstrated it to me, and it made so much sense, I had to adopt it for my own use. Let's consider a nine horse field at Northfield. If post position made no difference, that is, it's just as likely to win from any post position, then over the course of time the number of winners from every post would be (statistically) the same, just as randomly drawing the numbers from a hat. The winning percentage for each post would eventually converge at 11%. Note: yes, I know 9 X 11% is only 99%, but let's keep the math simple, shall we. It's the concept that's important, not the arithmetic.

Looking at the stats in the program, one can readily see that post positions do not converge around 11%, but vary widely from nearly twice, to less than half that percentage. Some simple math is all that is required to establish the relationships between the post positions Take the difference between the post position's winning percentage and what would be expected if post position didn't matter (11%) and divide by that same expected 11%. The resulting value will be either positive or negative and reflect the percentage difference above or below what would be expected if post position were no factor at all. Did you catch all that? Not that hard, really.

For example, if the winning percentage of a post were 17%, we subtract 11, then divide by that same 11:
(17-11)/11 = .55 or 55%

If the winning percentage were 4%, we subtract 11, then divide by that same 11:
(4-11)/11 = -.64 or -64%

Big difference, I would say. By the way, the two examples are typical of the difference between pp1 and pp8 at Northfield. What the numbers are telling us is that there is a 55% better chance of winning from pp1 than if all post positions were equal, and conversely, there is a negative 64% likelyhood of winning from pp8 than would be expected from a theoretically neutral post.

There are a few problems with just crunching the published program numbers, as I have suggested above. First of all, every race does not have a nine horse field, and since the program lists all starts, the percentages tend to be skewed upward for the inside horses, as there are generally always at least five horses starting, and their winning percentages increase in shorter fields. The program numbers don't reflect scratches which actually change the post positions of the horses from a handicapping point of view. And in some races the post positions are not randomly drawn, i.e., handicapped races where the better horses are placed outside by design.

These problems aside, the concept is sound. In the past I have kept my own stats, thus eliminating the concerns from using published program numbers. After a couple thousand races the issues from using program numbers are ameliorated, and I have convinced myself that using program stats gives a close enough approximation for handicapping purposes. It should also be noted that over a short period there are wide swings in the percentages. Horses starting with pp3 many win five or six times during a given card, for example, but over the long haul, the figures remain very consistent. I periodically check the figures, but I haven't found much variance over the last ten years or so. Here are the numbers that establish the relationships between post positions at Northfield:

I've left the percent sign off for clarity; the fact that the numbers are actually percentages is not really pertinent. The numbers are also rounded, as it is the basic relationship between the post positions that is worth noting.

Let's analyze these numbers from a handicapping point of view. Of course we knew that having the rail is quite an advantage, but who would have expected it to be twice the advantage as drawing the 2 hole? Or almost four times the advantage of the 3, 4, or 5? So yes, the move from any of these "positive" posts to the rail cannot be ignored. But how about a move from pp5 to pp3? If you trust these numbers there is really no statistical advantage to be gained with that move. Now here's a biggy ...look where the post position changes from a positive number to a negative number. As a horse moves out from pp5 to pp6 the number changes from a positive 15 to a negative 20 ! That's a swing of 35 points. More on this little tidbit in part two. A couple other items jump out. Unlike the inside posts 3, 4, and 5 whose numbers are identical, the outside posts 6, 7, and 8 get progressively worse the further out they go. Post 7 is twice as bad a draw as post 6, and post 8 is almost twice as bad as 7! They not only get worse, they get exponentially worse. (Slight exaggeration, pp8 would have to be -80 to be a truly exponential progression, but you see the point, I hope.)

A word about pp9. Post position 9 is a special situation. Most of the time the position is a positive, somewhat less an advantage than the other "positive" posts. However, so many other factors influence that post position, it can be treated as a negative post in many situations. In fact as this is being written in May, 2002, pp9 has been statistically negative during the current Northfield meet. (I told you I checked the numbers periodically ) It's beyond the scope of this piece to talk much more about pp9, but I'm planning on writing an article devoted entirely to pp9 in the future.

Let me end this and move on to part two by reminding you that post position is only one factor, and sometimes not even the most important one to consider when handicapping. I should also point out that the numbers here are specific to Northfield. Similar analysis can be done for other tracks, but the results will vary with the track. As a general rule, post position doesn't play as big a statistical factor at the larger tracks as it does at the half milers like Northfield. Also at most of the 5/8's tracks such as Raceway, Scioto, the Meadows, the most advantageous posts are the 4 and 5, not the rail, as it is at Northfield.

I hope this look at post position, Cliff style , provides some insights that you can add to your bag of tricks. I know once I fully understood the relationships between the post positions at Northfield, my handicapping improved markedly. Example: all post position changes are not created equal. A move of two places, from the 5 to the 3 means nothing, while that same two place move from the 7 to the 5 is dramatic.