I decided I wanted to try binary ARV again, since it is somewhat easier to get hits than the multiple-target horse races we have been doing, where each horse gets its own target.
(I had done some binary ARV before, awhile back, for the stock trading group.)
So I did a race yesterday.
The way it works is I divide the field into two parts, each of about 50% total probability of winning, assuming that the probability can be derived from the morning line odds.
(probability = 1 / (odds+1) is the formula.)
I call the two sets of horses Favorites and Longshots.
In this race (8th at Suffolk, 2010-09-27) the top 3 morning line choices (program numbers 6, 7, 9) added up to 50% of the probability, and the other 7 50%.
The ARV session was a good one, I got a direct hit on one of the photosites, and the other did not match hardly at all.
Targ CR score was 5 to 1.
(See below for sketch and winning photosite – the match to the lower portion of the sketch is clear.)
The way I bet multiple horses is to bet proportionally on the 3 horses, again according to their normalized morning line odds, such that I would profit the same amount no matter which of the 3 horses won.
In this case the horses’ morning line odds were 3.4, 4.0 and 4.7 to 1; their morning-line probabilities, .19, .17 and .15; fraction of bet to be placed on each, 38%, 33% and 29%; so after rounding, the amount to bet on each per $10 total was $4, $3 and $3.
(Note: I had to renormalize the track handicapper’s morning line odds – they added up to “tote probability” of 1.9, and it should only add up to 1.2, including the 0.2 track take.)
The theoretical return on investment if the bet wins for this race turned out to be +$0.66 profit per dollar bet, or $6.60 profit for a $10 bet.
It’s theoretical because it assumes that the horses post time odds will be the same as their morning line odds.
The morning line favorite (3.4-1) won the race, so I got a Hit – nice 😉
Even better, his post time odds were higher than the ML – he went off at 4.2 to 1, so my actual return on investment for this race was +$1.08 – I more than doubled my money.
It would make for better and more accurate results if one used the post time odds instead of morning line odds to do these calculations, of course.
Unfortunately, these days odds vary a great deal right at post time, mostly due to the effect of betting bots used by very large bettors.
While morning-line odds will likely be less optimal, I will use them for awhile and see what happens.
I can compare them with the results I would have gotten had I used post time odds,
It is not clear to me at this point how much the return-on-investment from either ML or post-time odds will vary, depending on the odds of the horses.
I will be doing some research on this point to find out.
By the way, one uses PRECOG for binary ARV, not PRECOG10.
It is easy to set up your own event in PRECOG.
(If you want to try this yourself, you will need a PRECOG login – it is separate from your PRECOG10 login.)
Winning Photosite - 2010-09-27 8th @ Suffolk