Dave Schwartz Teaches Hybrid Handicapping (Podcast Transcript, Part 4)
Over the years there have been a handful of key handicapping concepts that have dominated horse racing. In my lifetime it was Class, then Speed took over in the 1970s, then Impact Values in the 80s, Pace in the 90s, andÂ powerful regression systems in the new millennium.
What will be next? Where can you begin now to get an informational advantage over the competition?
What if you could get ahead of the curve?
To get an idea of what you might track – and this is not going to be without effort – you can pool resources. You might find a partner and share the data work. There are all kinds of ways to do it, but for starters you need to know what you are going to track.
Start with this: Consider any statistical table you have ever seen that is readily available. Think of any model, track profile or anything else that crosses your mind, even things from my Percentages and Probabilities book or William Quirin’s book, Winning at the Races.
Then, mentally take it apart. Ask yourself what is wrong with it. The answer will almost always be that it is too general. The true value of a statistical table, that is, its true value to the handicapper, lies in the shortcomings of the table.
The result of a statistical table can be improved in the hands of the skilled craftsman. Let me give you an example: post position. Many of us have picked up the newspaper, racing form, or back in the old days, a daily newspaper that showed post position.
Generally, you see that there are more winners stacked towards the top than there are at the bottom. So what came from that knowledge is that inside posts have an advantage. Is that true? William Quirin wrote about how inside post positions were better than outside posts. However, in his impact value tables he had a calculation mistake.
Anything that is a ranking factor there is a logical, natural way to use percentage of winners divided by percentage of starters. In such a case, there is a natural bias towards lower numbers. For example, when you compare post position one to post position twelve, post position twelve is always in a twelve horse field while post position one is sometimes in a five horse field.
Logically, there were times when post position one had a natural 2.0 impact value against post position twelve. That has been repaired in the newer calculations that I do in my Percentages and Probabilities book.
But the real point is that while an inside post does run a slightly shorter distance in a race, the truth is that, at some tracks, outside posts are much more viable. There are even tracks where the middle posts are more viable. You have to change the post position point of view. Instead of seeing numbers like one through fourteen. What you need to measure is â€œinsidedness.â€
In other words, you might look at post positions 1, 2 and 3 together as an inside post. Then you have everything outwards. This is how I do it in my software. Everything outwards is clumped together so you wind up with two values, 1-3 and 4+.
Then you do the same thing with outside posts so that you have 1-3, meaning that in a twelve horse field it would be 12-11-10 represented as â€œOutside,â€ then 9 and below goes into the â€œInsideâ€ area. Think about that and it should make sense to you.
12, 11 and 10 are the most outside posts in that race. In an eight horse field it would be 8, 7 and 6. Also, you need to build a â€œmiddlenessâ€ table so that you have an â€œinsidedness,â€ â€œoutsidednessâ€ andâ€ middleness.â€
In other words, if you have an eight horse field, the horses in post positions four and five would be â€œmiddleâ€ and everything else would be â€œnot middle.â€ In a nine-horse field, it would be 4, 5 and 6 as â€œmiddle,â€ and â€œnot middleâ€ would be the rest.
You have now taken a statistic that everyone has and you have looked at it in a way they are not looking at it. Will it add value? Make it track specific and it probably will have value to you andÂ the greatest value will be when your model says the outside post has an advantage and the public thinks it is the inside post.
I guarantee you that will happen. I am going to carry it even a step further. From any post position where there is just one table, you now have derived three tables – inside, middle, outside. What happens if you cross reference those post positions with the running style of the horse?
Is there anyone out there who is not aware that having the inside post at 6 furlongs could be the kiss of death at some tracks when the horse cannot get to the front? That just naturally screams that an off pace horse will not be as negatively impacted in the number 1 post as a front runner. As soon as you can see a logical distinction so that it lends itself to multiple tables you are now getting different opinions than what the public has and you now have your own unique set of post position data.
Someone is out there saying yes, but that is a lot of work. Of course it is! Some of this stuff will not even turn out to have value, but that is how you get unique statistics. By the way, I will guarantee that if you go out and cross reference post position as I just said, (inside, middle and outside) with running style, you will find something you can use that will give you an informational edge.
I actually had a user long ago in 1996 or 1997 whose approach to the game was post position statistics cross referenced with early speed points. In other words, he used Quirinâ€™s early speed points. It is simple-yet-powerful.
He actually made his living at that and, specifically, looked at seven and eight-point Quirin early speed horses – those that are pronounced front runners. He wanted to know how they do. His particular breakdown was posts 1, 2, 3-6 and outside. I believe that inside, outside and middle are better, but it was good enough for him. The point is to create unique information.
Again, I am not saying that these do not come with a workload. There really is no free lunch out there, but these are ideas. For example, when it comes to speed ratings, can you imagine that the value of speed ratings could also change based on the running style or Quirinâ€™s early speed points of the horse at a particular track-surface-distance?
In other words, if you had a Quirin-type table – what I use in my book is an impact value table that 1st, 2nd, 3rd, front half and rear half. It is something that simple. Can you imagine that there might be a difference in the value of 7 and 8-point horses, depending on the speed rank? Can you imagine that it might be more important to have big speed ratings if you had a lot of early speed?
In other words, it might be more important to rank high if you are an early speed horse than if you are a closer. I am not saying it is. I am saying it might be.
Why would I say that? When do closers typically win? They win when the front runners have died off and when the front runners die off the races run slower times.
The par for a race will be higher when an early horse wins than it is when a late horse wins. In other words, you can throw out early horses at a higher number. So, a race where your intended par is, (say) 85, might only apply to front runner, while a closer might only need an 80.
You might find that, in reality, a front runner has to run an 85 and a closer only has to run an 80. I am making these numbers up, but how do you find out? You test, keep some data and figure out the best way for you to do it.
Will it take some work? Yes, it will. I am not going to turn this into a speech about collaboration, but this screams out for collaboration. Find someone with whom you can work.
I have been talking about speed ratings. Speed is so obvious, yet there is an untapped possibility there. Trainer ratings can be important. Everyone looks at trainer ratings and they try to pick out this angle and that angle for help.
I just did a class last night for our software users and I actually invited the general public to it. We were talking about trainer stats and one of the problems is that you have all these categories. The question is which trainer stat is most important? Why not use them all or a subset of them. Find out which ones are most important.
How do you do that? Keep a few statistics. You only have to keep them on the top trainers to determine which ones have the greatest value. You do not actually know right now. Just take a bunch of them and average them together. As surprising as it sounds, it is a killer approach. That is what I demonstrated last night.
Trainer handicapping is the core of my own handicapping approach. The list of examples will go on and on. Post position is the least understood bias out there. That is a gift from me to you because if you do the work on post position, track surface, distance and more, you will find information that nobody else has.
You have to build some data. It is going to take you some time, but the good news is that by combining the three inside posts together, by the time you have ten races you do not have just a sample size of 10 in the â€œinsideâ€ category, you have 30. You also have the converse of that: the outside and the middle. This is doable.
There are other statistics for which you can build a table. Get creative.
In the first six months you are doing it, you may try five different factors and probably four of them will have no value whatsoever, but you will find one item that is a keeper. If you are an everyday player you will build data much quicker. If you are looking at data here and there you will see correlation from one track to another. If you are doing multiple tracks, you are really putting effort into it, but the payoff can be immense.
When it comes to jockeys, I am in the process of adding deep jockey statistics right now to our software. I have done this in the past, but not since we moved to Windows almost two decades ago.
Did you know there are jockeys that will absolutely only win from an inside post when they are on a front runner? Why? Chris McCarron, one of my favorite jockeys of all time, was one of them.
It was a statistical fact that when he was on a rail horse that was a closer, he was going to have to circle the field because he was just not coming through on the rail. It is the hungry jocks that come through on the rail.
Do you want to know if a jockey is really good? Track how well he does in photo finishes. Look at every single photo finish in every single race. By the way, this is not original idea of mine. I have to give the credit to Gordon Jones, a very intelligent man. You look at every horse that finished within a length, record the number of jockey starts and jockey wins and you will be amazed at that as a statistic.
All of this started with Hybrid Handicapping.
I hope this got your juices flowing and your mind working a little. Go out there and find something to track. If nothing else, work on building hit rate and value ratings for each of your horses. All you have to do is rank them, determine which one has the best chance of winning and what, in your estimation, is the most undervalued horse by the public.
You take those two, multiply them together and track that. What is a 1 x 1, 1 x 2 and a 3 x 1? How do those horses work for you? Do they make money or do they lose money?
You know what the Greeks used to say: know thyself.
Those Greeks were really smart.
Check back often.Â Did you know that the best stuffÂ goes directly to our email community?Â Be the first to hear about special offers, new products, and new material.Â Sign Up Today.
Join the discussion in ourÂ private Talking Handicapping with Dave Schwartz Facebook Group.