A Handicapping Revolution? Introducing Betmix

Many times on this site we have broached the subject of how to get more people involved the sport of Horse Racing. And many times I have said that what I think inhibits its growth is that there is so much minutiae involved, so many things to learn, and so many ways to process information. The Form, and its tiny, black type intimidate some people. And even with mastery of its unique language,the facets of handicapping themselves -- speed, class, form, pace, trainer intention, and pedigree -- can leave the casual player a bit bewildered.

I make media, and occasionally trawl the internet for archival footage of things and remixes of songs. I do periodic searches of words like "footage" and "mix" and scroll deep to see what they bring up. In an odd collision of my interests, my search for media brought me a horse racing website. Last week, I came across Betmix.com. Betmix, which was created by a man named Dave Williams, is a fascinating piece of software that enables a user to handicap a horse race using any combination of 33 factors that the user may weight in importance. Betmix's interface (pictured below) is elegant. Users enable a factor by turning it "on" and then use a slider to set how much weight they give that factor. Any number of the 33 may be selected and weighted, thereby creating a "mix." Say, you value late pace in a race, but you also pay attention to earnings. You could select those two factors and give them equal weight, and then Betmix will order the horses according to your "mix." You can save any number of different mixes and use them quickly to look at any race in a variety of ways.

This is such a simple and intuitive tool, and coupled with its elegant interface, one that I think could help the sport gain all kinds of new fans. And it isn't just for newbies. I think it offers real value to seasoned handicappers as well. In fact, I think it could be nothing short of revolutionary in its power and ease-of-use. I reached out to Mr. Williams with some questions and he was kind enough to respond.

JP: First things first, Dave. Tell me a little about yourself.

DW: I grew up in Louisville, a lifelong fan of horse racing. I became intrigued with PP's and stats as a kid. I have owned several horses in partnerships through the years, and worked on a few horse farms. I have a background in programming, moved to Chicago as consultant with IBM, bought a home near Arlington Park, and then moved back to Kentucky when my first kid was born.

JP: What was the genesis of Betmix?

DW: I have always been interested in creating systems and seeing how long I could make them profitable. I created Betmix because I am a firm believer in using consistent methods. You have to be exact in your approach to handicapping to know if what you are doing is working. I love to talk to people at the races and pick their brains about why they selected a certain horse, or what their handicapping method is. It's always a little bit of this, and a little bit of that. But they have trouble defining the methods they used, and what worked for them once may work for them again, but they can't seem to quantify what they did. If they tell me that they are speed handicappers but also like to use trainer/jockey stats in conjunction with Speed Figures it is hard for a typical person to verbalize how important each factor is to them. Most people do not have hard and fast rules that they handicap by. I'm not arguing that intuition isn't important, or that if you don't use an exact method or process that you are doing things the wrong way, I'm making the case that if you can't measure what you are doing then you will never improve.

So, I had developed several spreadsheets that I used to input data point, and rank horses. I went through some exercises where I tried to quantify how important each factor was to me in my decision making process (in my head) and to see if what I thought I was doing could be supported by how the math added up. I found out that I was often wrong when looking at multiple factors and assuming that I had correctly identified the horse I should bet on based on the conditions I had set out.

I was struggling with some of the math behind the weighting, and getting extremely tired of inputting data. I then decided to start importing Bris data files and hooked up with a physicist at VA Tech that I had worked with on a previous project for some help on the math. At first I thought working with someone who had absolutely no knowledge of horse racing would be difficult, but I found that it was actually a positive thing. Math is math, and it doesn't matter where a number comes from or what a data point is called. He didn't have any prejudices about speed, pace, class, etc. and was able to treat every factor as equal. That is when it really clicked - what I think may be important may actually not be to someone else. Betmix allows you to decide which factors you want to consider in your decision, and also how important each one should be. You and I may choose completely different factors and weigh each of them differently yet still end up with the same horse on top. We can both be right as long as we picked the winner.

JP: I think that is a tremendous concept, because there really are many paths to the winner, if you will. I know some people who go to the track with nothing but pedigree stats and find longshots with consistency, and then I know others who look at nothing but the earnings box, and still others who average speed numbers at distance or surface. I respect what you have done here. My wife and I did something similar: creating a database that looked at 14 factors we value from over a thousand races we had handicapped. What I find most interesting is how your program allows the user to set his own parameters. That makes it viable for people new to the sport as well as those who already know what they value in the PPs. Was there a moment when you thought to yourself, I am really on to something here?

DW: Betmix was originally a project that I was going to keep to myself. I showed it to a few friends, and before long, they had given up their old methods of handicapping and started using Betmix exclusively. I wasn't sure there was a market for a product like this until one day when I was at the track with a group of people, several of whom had never been before and who were completely overwhelmed by trying to read the DRF. I had my laptop with me and let them handicap by using Betmix, and it made much more sense to them at that point. They enjoyed using the program and had fun manipulating the sliders. The real breakthrough was when I noticed how they were using Betmix after the race, to see what factors and settings they could have used to pick a winner, and then they applied those same settings to the following race. Although each race is different, it made sense to them to handicap each race the same way, or at least to consider the factors that picked the winner in one race again in the next. History does repeat itself, and that is very true with horse racing as well. With Betmix you can save as many specific mixes as you want and then apply that mix to any future race you see by selecting it in your saved Mixes dropdown. By doing that you are looking at races consistently and without prejudice.

I can't tell you how many longshots I have come up with using Betmix that I never would have bet before. I may have liked the horse on paper, but thrown it out because I didn't like where it ran its last race, or wasn't too fond of the trainer. Now, I know that I can trust the numbers. If the horse is on top I don't second guess myself.

JP: How long has Betmix been out there, and how has it developed since it went live?

DW: We made Betmix available to the public in September of 2012 and have been upgrading the product ever since. The first big upgrade was with MixMaker. That allows you to pull up a previous race, tell the computer who ran 1st, 2nd, and 3rd, and it will attempt to create a Mix for you that would have accurately predicted that finish. It's a great starting point for users who may not have a deep background with handicapping and lets them see what factors were most predictive in different race types. We are going to continue working with the algorithms behind MixMaker and see if we can make it smarter. We also created the MixLibrary which allows users to submit mixes that they have created and allow other users to take them and use them in their handicapping. Many of our updates have come from user suggestions, and we have a long list of things that we are working on to improve the product. We are growing steadily and mainly by word of mouth, which is great. We haven't spent much on marketing, everything is going back into development at this point. Based on feedback from users we know that we are providing a great product and service and are headed in the right direction. Once a user gets comfortable with the product they become very heavy users, spending hours a day on the site looking at past races and coming up with new mixes.

JP: Is there a secret to creating successful mixes?

DW: I have found that when you are creating a general mix you really only need to use a handful of factors to get positive results. When you are creating a mix for a specific race, say 1 mile turf ALW @ Saratoga, you will want to use several factors. As I mentioned before, every user approaches things differently, and different methods can produce similar results. It's a matter of preference, but the key thing is that you are able to handicap a race the same way each time. By doing that you can figure out if its working or not. If it's not, you can adjust your mix and then go and apply it to all of the races in our database to see if it works. There is some trial and error involved, but when you get it right, it's well worth it.

JP: Was there a process in how you came to choose the factors that you did, and are there more factors being added in the future?

DW: The 33 factors that we currently have are based on some studies that we looked at that isolated handicapping factors and ranked their impact value. Obviously not every factor is applicable to each race (some of them would be very specific) but in general the group that we have out there would cover almost every handicapping angle. That's not to say that we are finished - far from it. User feedback is very important to us and we have heard from users about other factors they would like included as well as different ways to use the current factors. We will continue to improve the software, and will be adding new factors and features before the Saratoga meet.

JP: I was wondering, as the creator of such a powerful handicapping tool, do you still use PP's at all? And were you an angle player as a handicapper before Betmix?

DW: I exclusively use Betmix. I do not use any other product. In the typical development of a new customer we find that at first they compare the Betmix results to their normal handicapping method, whether that is another software product or a hard copy of the DRF. It takes a new customer a few weeks of handicapping to fully trust the system and the mixes they have built before they are willing to give up their old methods. Part of that development takes place when they create mixes and test them against past results. If they are very serious handicappers then they have some records that they have maintained about the bets that they made in the past, and they can use Betmix to go back and handicap those races and determine if they would have done better with us than with their old methods.

We have had numerous emails from customers that let us know that they have stopped buying the form, and that they are relying on Betmix exclusively at this point, and that tells us we are doing things right.

I was an angle player before Betmix , and was an enthusiastic developer and tester of "systems". The desire to have a truly measurable and consistent handicapping method was the spark that inspired Betmix. Betmix is a great tool for angle players, you can spot check every race on the card in seconds to find scenarios that you want to investigate further (using the All Races Report).

JP: Thanks very much for the time, and congratulations on a great product. I really think it could have a widespread impact in the sport, and could help it gain lots of new fans.

DW: You're welcome! I'm very excited about the site's potential and I greatly appreciate the exposure! Thank you very much for the opportunity. I wish you and your site's followers a very successful Derby!

You can check the site out for yourself at Betmix.com. There is a free Race of the Day offered each day where you can experiment with the software.