In 1979, a book by Dr William Quirin changed thoroughbred handicapping forever. The title was Winning At The Races, Computer Discoveries in Thoroughbred Handicapping.
In the book, Quirin wrote on many aspects of race analysis and was the first to document his suppositions with computerised data from sizeable race samples. Almost immediately, terms such as "impact value" and "ROI" became part of the vernacular of modern handicapping culture.
But it was an obscure chapter near the end of the book that energised a transition from handicapping as an art toward becoming a pure science.
Chapter 25 of Winning At The Races contained the first published computer algorithm for analysing a horse race without human intervention. It was different from other mechanical approaches (systems) to beating
the races in several important respects.
First, the computer was used to examine past race results and create a set of criteria that was common to the winners of those races. Next, it tested those winning factors in combinations until the most effective approach was found.
Finally, it could take in data from a current race and produce the most probable outcome according to what it had learned from the past. An accurate odds-line was an additional key benefit.
The formulas, one for sprints and one for longer trips, were based on multiple regression analysis of the results of 646 sprint races and 300 longer-trip races.
Of course, there was no guarantee (or evidence) that they would work on an entirely different group of races, a problem that continues to haunt handicapping researchers today.
However, the fact that a computer did the research, not a human, was significant, and inspired legions of computer programmers to look for similar formulas.
More than 20 years later, the advances in technology and computing power have blown the science of picking winners wide open.
Inexpensive sources of mass data are easily available on the Internet.
Tens of thousands of races can be tested in a matter of minutes with a high-speed PC. Tremendous storage capacity of hard disk drives and databases allows every aspect of thoroughbred racing to be dissected and quantified.
Yet, has the rapid evolution of information processing made picking winners and making money at the track any easier? I'll answer that question while exploring ideas on computer usage that can help us become highly skilled horse bettors.
It's been said that the first step toward becoming a billionaire is to keep a daily log of every cent Spent. That kind of detailed record keeping is the most basic form of self-analysis and a firm source of feedback on one's financial health.
Successful horse players are also dedicated record-keepers. In my own 25-year odyssey as a gambler, my most profitable years came when I studiously used a spreadsheet to record my actions at the casino, racetrack or betting shop.
‘The future with technology is potentially amazing’
Spreadsheets (e. g. Microsoft Excel) are nothing more than an organised group of boxes that allows us to compartmentalise information on the computer. They have enormous advantages over manual record-keeping methods and are easy enough to learn from a book.
After creating some simple columns, the spreadsheet software takes care of organisation and computation for us. But are we too lazy or tired at the end of a raceday to perform 10 minutes of typing?
Sometimes it's tough to face those inexplicable losses and the impetuously stupid betting decisions we make at the track. The computer becomes an effective therapist if we can face the task.
The spreadsheet display will not lie to us if losses are mounting. But computers are wonderfully disarming when delivering bad news in our face. They don't pass judgement and never make cynical remarks.
If the ship is off-course, it must be righted. Changes in race selection (age/sex/class) or betting strategy (exotics/win bets) are the most common positive result of spreadsheet feedback for punters.
If you're betting too many races or seem to have no luck with trifectas, while your longshot win bets are doing well, a spreadsheet will reveal the truth. And that truth will be different for every player.
For me, it was my frequent losses in the Pick 6 that convinced me to abandon this bet, other than for token play.
Many punters expand their use of the spreadsheet into areas of data analysis, particularly by keeping trainer and jockey records. But spreadsheets are too cumbersome for large tasks. That kind of grunt work is best done by a database program, which is best suited for the more demanding chore of managing and making sense of large amounts of historical horseracing data.
Microsoft Access is the most popular such software.
A database is really a gigantic well-organised storage container. Information is placed in records that can be retrieved and utilised instantly. Each individual record is made up of separate data fields.
A data field contains a single type of information such as the date, or name of the racetrack, or the distance of a race.
The key task of a database program is to sort and query the information. Suppose we kept a simple database that contained all the records of all North American (or Australian) races with just the following fields: Date, racetrack, winning trainer, distance of the race and the win prices.
By sorting the trainer field, we can retrieve all instances of "Freedman, Lee" in a split second. We could then query the Freedman records to tell us which track he won the most races at, or which distance is his best, or perhaps how often he wins with a longshot, and when he does win with longshots, where, when and how does he do it?
The search possibilities are endless, particularly as the number of fields is expanded. The handicapper with this capability will know more about the trainer than the man himself.
The use of database software has become vogue in recent years among successful punters.
Independent research can uncover critical facts the public cannot perceive through regular sources. The speed and storage capacity of modern home computers, combined with inexpensive data sources, has enabled any punter of modest means to become a warehouse of knowledge.
And a warning from Barry Meadow - Remember the days before you could download every known fact about every horse in every race, everywhere? When it was just you, the formguide, and a lamp?
Now, you can buy computer programs not only to help you handicap, but to do all the handicapping for you. Some will analyse your betting strategies.
And for system players, others automate the drudgery of doing the calculations required for some systems.
Dozens of programs are available and more enter the market each month. But which ones might help you and which are a waste of money? Not so easy to determine.
Handicappers may be looking for a black box, a program that spits out winners, but despite some questionable claims there doesn't seem to be any such thing on the market. But that doesn't necessarily negate the value of computer handicapping programs.
After all, the Daily Racing Form (or any formguide) doesn't spit out winners, either. They simply provide information. And that's what most computer handicapping programs do.
And many do it quite well, offering a dazzling variety of ways to analyse that information.
Some players swear by the computer handicapping programs they use, while others swear AT them. Certainly, when you decide to get into a computer program, you're looking at a commitment, of money, of time, of a new way of looking at data.
Whether this commitment will earn you any money at the track is another question altogether.
In some ways, computer programs are vastly superior to the human brain. As calculations get more complex, computers are virtually required.
Is it any easier to win using computer programs? Hard to say. Nobody's ever come up with accurate numbers to show who winning players are, how much they win, what tools they use, or much else.
Naturally, the programmers and vendors sing the praises of their products, but that doesn't mean much.
By itself, a software program probably won't make a losing bettor into a winner. But a good one is a useful tool that will take much of the mathematical drudgery out of your daily handicapping chores, and help you analyse races through an entirely different point of view.
Or maybe dozens of different points of view.
Yet how valuable is this historical information? The database may uncover past patterns of success but this might be dangerous to the unsuspecting user.
Small samples are a particular problem. Let's say that a certain trainer has started, in America, nine dirt-to-turf starters over the past year with zero wins. Doesn't sound promising, does it?
But what if two of those were 40 /1 and lost by a nose apiece? Had either won, the ROI (Return on Investment) for this trainer would be huge in this category. Had they both won, he'd show, say, 22 per cent wins with a gigantic ROI. So is the trainer good with dirt-to-turf moves or not?
However, what if over a five-year period we did uncover that a certain trainer is, say, 0 for 77 with Maiden first-time starters at Saratoga (or any Australian track). That would be a significant pattern and confidence would be high if we bet against it in the future.
How do we deal with the unique dynamics of today's races on our computer? That is the function of "handicapping software".
The essential role of the handicapping computer program is to relate key facts about a field of horses and compare them side by side. Doing so can reveal Strengths and weaknesses that are not readily seen by perusing the regular past performances.
Computers are not street-smart handicappers and they cannot understand the insightful nuances of an individual race the way an experienced player can. Yet they do have great advantages over the human brain in terms of calculation speed and organisation.
There are many key distinctions that can be uncovered with handicapping software. Among the best I have found for profitable play are:
- The demonstration of the power of early speed over final time speed ratings, and the knowledge of which tracks and distance,; that early speed is dominant over all other factors.
- The ability to quantify every class level and an,, set of race conditions onto a comparative scale of integers or pars that allow immediate class appraisal of any race shown in the past performances.
- A clear understanding of the factors commonly prevalent when longshots win.
- An insight into the surprising pace patterns and running styles that make up the back end of highpriced exacta and trifecta payoffs, and how the mix of running styles in each race affects the outcome.
- The ability to extract plays from all tracks on a given day in an instant.
- The ability to keep an ongoing record of results and to uncover profitable betting strategies.
The most essential lesson that it learned from the regular use of handicapping software is the recognition of how difficult the game is to beat.
The experience of scanning and scrutinising thousands of races makes it abundantly clear that betting low-priced, standout horses will result in severe losses.
Successful handicappers must hit a home run with the occasional longshot to fatten the bankroll, and the ego.
Finding overlay winners has become increasingly more difficult to do with standard pen and pencil methods such as speed figures or trainer and pedigree stats.
Quality handicapping software with a variety of tools can fill that void.
Can we say for sure that computers have increased the number of winning punters? No.
The percentage of those who realise realistic profits from horse betting has probably remained the same for a hundred years. But hightech handicapping has made the game more competitive, and a far more interesting pursuit.
Everyone realises that to make money betting on thoroughbreds requires some knowledge that the masses of other bettors are not aware of. Computers have a decisive role in that undertaking.
The future with horse-racing and technology is potentially amazing. The punter, anywhere, who connects his home-grown software to the tote may only be the beginning of new high-speed functionality. In a few years, handheld cyber-units will be common and they will be connected to a wireless Internet.
Along with built-in handicapping software, punters will find late odds, live races and touch screen betting in a device they can carry anywhere ... never to miss out on another betting opportunity.
Ken Massa writes computer software for HTR (Handicapping Technology & Research) and provides a download service and monthly newsletter out of California. His website, which has many useful articles, is www.homebased2.com/km
By Ken Massa
PRACTICAL PUNTING - FEBRUARY 2002