$97,000 Superfecta ?

[km]

Our local paper horse racing column in the O.C. Register had a byline: "Whopping $97,000 Superfecta Payoff"

Maybe the writer should have checked his facts first.
The entire superfecta pool was only $85,000.

So nobody went home with 97k in that race, but it looks like 6 people did have a 10-center and got $9,700 each. The payoff on the charts is calibrated to a $1 base unit.

Our charts and Results files have the same $97,000. Even the Equibase (new) Charts don't make this distinction and the $97,000 payoff for $1 is completely misleading. Even if only one bettor hit the super for 10cents, his "all burger" payoff would have been around $60,000, but the charts would have shown a $1 payoff = $600,000 !!

It is futile to research superfecta results because of this.

[MikeDee]

If only everyone would agree to report ALL pay offs in the lowest betting unit available then you wouldn't have this problem.

Trying to find a spot play for supers is futile though regardless of the how they report payoffs. In my opinion almost every horse in the race has a shot at 3 & 4th so trying to handicap them is futile and a all, all in the bottom 2 holes is a quick way to breaking the bank unless you can hit the win place on top cold.

[MVM]

I have kind of an opposite take on both items here (I hope that doesn't make me obnoxious).

I think doing research on superfecta betting strategies has (at least for me) been worth the time, if only because the majority of players either don't do it, or do it incorrectly. I think the key is to account for your own wager and it's effect on the payout.

As far as using a large number of horses in the 3rd and 4th spots, I have had the best results doing just the opposite. I have a philosophy that if I don't have an opinion on who is going to finish 4th, then I would play the trifecta. If I have no opinion on who is going to finish 3rd, then why play a tri.

I think there are probably merits to both methods, so it probably just goes to show that (once again) there is more than one way to shave a cat (skinning is just too barbaric).

[km]

MV, good idea if you are merely establishing a key to finish 3rd or 4th at high odds. 20/1+ shots that are Fr3=1 finish 4th as often as the favorite does, so a % tally on that would be enlightening.

But if you try to run an automated ROI on a history of those horses in the super, it cannot be done with confidence because the payoffs are distorted.

For example, as mentioned, Fr3 =1 above 20/1 will finish 3rd and 4th at far higher rates than they finish 1-2. But if I ran my db on this and tallied the Equibase superfecta prices,the resulting ROI is worthless. Not only because of the 10cent, but the "ALL" on the bottom as well cannot be ascertained.

The only way I see around this is to have data fields for $1 superfecta price and 10c superfecta price as 'Dee mentions. Also that the official finish position is followed by the payoff if the horse was inolved. So a horse finishing 8th in a superfecta ALL, would have the superfecta prices listed next to it also, which establishes that the horse was involved in the payoff.

Having been given all that, i still think it might be useless to research for ROI. The reason why is that the superfecta price 'going forward' would be far different if all the people researching it now had made a bet on it. The key to making money in the super is to drop a bomb and score big. But if our research shows a pattern on a backfit, the big price in the future won't be realized because too many bettors figured it out by researching!

So MV's approach makes the most sense, find a factor that promotes longshots into the 3rd and 4th spots more often than expected and single it there and cross your fingers.

[MVM]

I have a field in my ExotPay (Exotic Payout) table that called TxSold (Tickets Sold). In the case of a superfecta all payout, I have a record with a 0 in the TxSold field for the superfecta itself, as well as a record for what is the equivilant of a superfecta consolation payoff for the same race with a NumCor (Number Correct) value of 3 or less.
I keep the actual Pool Info in a seperate table to allow a 1-Many relationship between a pool and a payout.
Coding the calculation for determining the tickets sold on a dead-heat was a bit painful, so I have never bothered to extrapolate it out in case of 3 or more seperate payouts for a single exotic.

Rather than make the payout value reflect the 10 cent or 50 cent minimum, I keep everything at the $1 payout, and set the TxSold field as a single decimal value. In the case of the race inquestion, TxSold value is .6.

I got away from nth Normal Form database design a few years ago. I calculate and store a lot of values during the import routine that could technically be derived from values in the db, it really speeds up any kind of research.

In the case of all exotics, I have fields indicating that betting interests value as a multiplier.
I have tables for Entrys (an individual horse), as well as BetIntrst (a betting interest). This allows me to (again) establish a 1-M relationship between individual horses and the betting interest (multiple horses in the case of a coupled entry). I have the usual fields for the $1 W-P_S payouts in the BetIntrst table, as well as a Boolean field in the Entry table that indicates whether or not a payout was "inherited". Example the 1 wins a race and the 1A runs 3rd, the 1A inherited the Win and Place payouts.
As far as exotics go, I have fields for ExMult, TriMult, SupMult,, P2Mult etc. etc.
In the case of an exacta, ExMult is simply the Exacta price/Win price, TriMult is Trifecta Price/Exacta Price, etc. Same case for the multi race bets.
The really obnoxious thing I do, is in cases where I played the race, i take my money out of the pool. Since my wagers are in the same database as the race data, this is fairly simple to do. This way, when I run a betting simulation, I am running against data that does not include my effect on it. At the same time, the simulation puts my pretend money into the pool, and calculates my pretend winners effect on the payout. Therefore, the payout values stored in my db sometimes differ from those in the chart (on those rare occasions i actually cash).

[km]

MVM: brilliant!

Like the fact that you have explored alternative methods of research.

I have always been fascinated with finding the "perfect zero". This means I'm looking for a factor or combination of factors in HTR that will produce the largest sample size without a single winner occuring. It is far more difficult than you expect, we sort of proved that with the @@ (removed from latest version). While a meager 1-2% of them won, those winners seemed to be key longshot bombs showing up at the most opportune times!

How about a little contest? $100 prize to the first person that can locate a 1000+ sample of races with specific verifiable parameters that produces zero winners.

1000 consecutive losers is a tough proposition. But if you want to post it here and the rest of us can verify it, i'll send you a Benjamin and maybe we'll learn something.

The best one i think i uncovered was a play that went 3 for 678 or something. Using thinkgs like K=9, "R" running style, E/P = 9, JKY < 100, etc. But if you get too picky, the sample size drops way down.

This is not a geek-ercise. Finding out what does NOT work is just as important as understanding what does. The outcome of most races with a big longshot is typically a "perfect storm" for the winner. The ability to eliminate hopeless horses is critical so that even if everything does fall their way, they still don't win.

[fred4now]

MVM, That is really original-cool.

KM, "1000+ sample of races", over how many years data am I allowed? :D

[km]

Hey F4N, you find a 1000 straight losers over any period of time and I'm interested!