Who is "really" handicapping? What is "really" handicapping? And does it really even matter? I didn't want to contribute to thread creep in the Is a Positive ROI Attainable... thread so I started a new one.
The fundamental question, as Jim Cramer always said, is "How do I take money out of this race?" At the extreme ends, many people believe that for most races, you can't. Fill in the blank with your favorite reasons - crooked trainers, crooked jockeys, drugs, whales have ruined the sport, the markets are almost completely efficient, etc. Fewer believe that for most races, you can. Love them or hate them, rebates to large volume players has certainly influenced how one goes about answering the question "How do I take money out of this race?"
As a first distinction, rather than using words like "handicapping" and "assessing value", I'll make 2 distinctions:
1. Determining what bets to make in a race or sequence of races (as in daily doubles, pick 4s, etc.)
2. Executing those bets
Step 1 might result in
- 2 horses to make a win bet on
- An exacta bet
- Multiple trifecta bets
- A place bet
Step 2 might be
- Walk up to a parimutuel window at a track and read the bet(s) to the clerk
- Call up your bookie
- Log onto an ADW platform like Twinspires or DRFBets and make bet via interface
- Upload a conditional odds list of bets to your ADW so the bets are only executed if odds thresholds are met
- Bet diectly through an API that an ADW makes available to (typically) large volume customers
These steps are certainly not independent. The ability to bet through an API directly into the pool affects the types of bets one may assemble and the manner of assembling them. So with an API, one can work with a model with hundreds and even thousands of probability combinations for all the available bets in each race and compare all probablities and current odds (or projected odds) in milliseconds and then execute bets that meet preset thresholds at very close to the fnal flash before pools close. This can mean that the decision on WHAT to bet is made in milliseconds before the bet is executed, determined by all the programming and modelling that most likely took years of work and testing.
On the other hand, a player who has a spot play that generates a play in race 6 might very well bet before the card even starts and then go off and do something else until checking the results the next morning. Step 2 is just not as critical for this player. Just execute the bet before the race starts. Bet execution is literally trivial.
In addition, the determining what bets to make portion may be fluid, such that the list of bets to make changes from one flash to the next, because of real time odds changes. This might occur for somebody who has a win bet in sight and is watching the tote board at the track for favorable odds. It might also occur for high speed API betters where thousands of probabilities are repeatedly compared by computers to current and projected odds, each flash perhaps giving a different answer to the question "what bets to make?"
Ask the player with a DRF in hand what bet(s) she is making, you'll probably get a stright answer, like the 5 horse to win. Or maybe the 5 horse to win if the odds are higher than 3 to 1. Ask an API bettor and there is literally no answer until bet execution time, or the answer is so complex as to be meaningless. The bets assembled 5 minutes before final flash could be completely different from the bets actually made at the final flash. And the API bettor just (just!!!) needs a program that compares model probablities with current odds or sequence of odds and makes the bets through the API, either manually, with the press of a button, or automatically. There is no "Who do you like in race 6?" and maybe not even "What did you bet in race 6?" Whatever the model and odds and the bet generation rules spit out.
Whatever the model and current odds and the bet generation rules spit out? Hmm. The analytical work was done before hand. The hundreds and thousands of hours spent developing, refining and testing the model and keeping it tweaked. Incorporating ever new data in the models to maintain an edge. Database maintenance. Programming the analytical engines. Bet generation software. API programming. Tracking bets and results. Reconciling anomalies. Bug fixes (Ugh!). New ideas (yes, thinking does not stop). This is NOT the lazy person's way to riches (or the poor house). A betting team or group can allow the work to be spread out. A one person shop can be brutal.
Is the API player (or the model player who executes bets through an interface like Twinspires) a "real" handicapper? It just doesn't matter.
Well, give the API pseudo-handicapper-model-programmer a DRF and send them to the track and the result might be comical (though not always). Give the DRF handicapper MATLAB, some cloud servers, a robust programming platform and several terabytes of raw data and the result might be comical (though not always). Give both of them a popular handicapping program and send them to the NHC tournament and they probably finish mid to bottom pack. This is not to elevate one and denigrate the others. Its fantastically different approaches to answering the fundamental question "How do I take money out of this race / these races / this sport?" In my mind, when done with skill and expertise and success, all of these are things of beauty. Whatever you call them.
OK, OK, there's nothing beautiful about a complete novice betting his kids birthdays and hitting a 2 million dollar pick 6, while you've been struggling for decades to hit anything. On the other hand, maybe there is...
Ron Tiller
HDW
