|
|
 |
|
|
|
Fun with numbers
By Doug Parry Some people spend their winters skiing. Some spend them traveling to nice, warm places. I spent part of mine making speed figures. Some of the best authors I've read, such as James Quinn and Steve Davidowitz, say a serious horseplayer has to try it sometime. I decided to see if I was serious. Even if your speed figures turn out to be money-losers you can't use, making them will teach you so much about the value of the popular speed figures on the market, and how much margin of error is built into them. Why make your own figures?It's tough to set out to make your own speed figures without first defining your goals. It's not enough just to make them. They have to be better than what you can get when you plunk your money down for past performances in the Racing Form or online. Do you want your speed figures to take into account factors such as ground lost on the turns? Maybe you want the effect of a fast pace to register in your figures, rather than just using the final times. In my case, I've always thought the big speed figure makers are too distant to pay much attention to Emerald Downs. Some of them are even made by computers. By making my own speed figures, I could take into account subtle class differences and pace situations that the big guys might miss. I'd also have my own angle, as opposed to the overexposed figures in the Racing Form. The first item of business in making your own speed figures is to figure out where to begin. Where to start: Par timesMost of the best speed figures out there today owe their origin to Andrew Beyer's methods. Beyer was the one who popularized the use of par times -- the times you'd expect a group of, say, $10,000 claimers to run. Based on how fast or slow a day's races are run in comparison to par, you can get an idea of how the track surface itself is affecting horses' performances. After adding or subtracting this effect, you can get a basic speed figure. Beyer's principles seemed like a good place to start. If I was going to have accurate speed figures, I needed to have super-accurate par times. I got a set of charts for last summer's meet and got to work. I looked at the running time for every race of the meet, 875 races in all, and grouped them into class levels. It's important when doing this to separate ages, sexes, distances and classes within classes. Straight $10,000 3-year-old fillies, for example, are NOT the same as $10,000 nonwinners of two races, and the running times you'd expect from them are very different. If looking at every class level sounds like too big of a task, you can get similar results by looking at the most popular class levels and extrapolate the rest. The track runs dozens of races for bottom-level $3,200 claimers, but races for $50,000 claimers are so rare it's hard to get any meaningful numbers out of them. Get the variantArmed with par times for each class level and your set of charts, you can look at every race from a given day and find out how fast the track was playing. Then adjust the running times accordingly. Times were extremely slow across the board on July 22 of last year, for example, so it made sense to shave the day's times to make them faster. On Aug. 8, on the other hand, the day's running times tended to be faster than expected. Horses' speed figures should be penalized, because they came on a track surface that was helping them along. The amount you add or subtract to the running times is called the track variant. Getting down to numbersOnce you've got par times and track variants in hand, the figures themselves are relatively easy. Start with a single class level that you believe you have a solid par time for, one that comes around often at several distances. The bottom-level $3,200 claimers seem to run all the time, but $8,000 4-year-olds and up might be a more consistent group to use. Next, assign a speed figure number that's par for the group. Par for the $8,000 4 and up claimers in my system is 83. The number you assign should be the claiming level's par at any distance, short or long. Decide how much time one point is worth in your system. If one second is worth 10 points, two seconds are worth 20 points, and 2/10ths of a second is worth 2 points. You may want to make a single point worth less time in sprints and more in routes. After your point system is in place, you can pin a speed figure on every race, based on the times you've already adjusted. That gets you the winner's speed figure, but what about the horses trailing behind? Make yourself a chart saying how much a length is worth in your system at different distances. Beyer's book "Picking Winners" and others have good beaten-lengths charts to use as a starting point. And remember, one second is worth about six lengths, not five lengths as many believe. Grading your workWas it all worthwhile? Depending on how much work you put into your speed figures, this can be the moment of truth. Fortunately, there's a way to find out right in the charts. An easy way to do it is to look at the money return from your top figure horse in a sample of races. Throw out the horses shipping in from other tracks or coming off of long layoffs. If you want to get deeper into details, use the last three speed figures for each horse, giving more weight the most recent one. The bigger the sample, the better; and you might get better numbers from using the money return for win, place and show -- not just for the win pool. To get the $2 ROI (return on investment) on the top figure horse, divide the money return by the number of races. (Divide again by three if you use WPS instead of just the win pool.) If your top figure horses return $90.00 in 50 races, your $2 ROI is $1.80. An ROI above $1.70 beats the track takeout; an ROI over $2 is profitable. In my case, I was happy to record a $2.11 ROI on a 100-race sample. If calculating your ROI doesn't sound like fun, another way to study your speed figures is to find the best ones and see if they seem right. Wild Wonder set a track record in winning the 1998 Longacres Mile. If makes sense if his speed figure is the best of the year. Click here for the top 20 performances of 1998 under my system. The last wordIf there's an angle you think you can add to speed figures, include it. (For obvious reasons, I didn't include all of mine here.) If there's a flaw you see in the popular speed figures, address it. You just might come up with a winning angle all your own. |