PDA

View Full Version : One Day Spot Check of 'Cramer Exacta'


stu
11-09-2007, 03:23 PM
From reading Marc Cramer's works, I interprete the 'Cramer exacta' a hedge better. When betting a non-favorite, some hedge with a place bet. Cramer suggested considering an exacta with the post-time favorite over your horse.

I did a quick one day study to spot check the concept. This is something I do periodically to make sure tools in my toolbox still work. I used the results for North American thoroughbreds on Thursday, Novemver 8.

I compared the place payoff for the second place finisher with the $2 payoff of the 'Cramer exacta' for races where the favorite didn't finish second. If the exacta didn't hit then $0.00 went in the exacta column.

The sample included 139 races. 43% had the favorite winning. The place prices ranged from $2.40 to $37.00. The exactas ranged from $5.60 to $188.40.

Place ROI = 2.91 (where 0=break even)
'Cramer Exacta' ROI = 7.58

Just sharing an observation,
Stu

ryesteve
11-09-2007, 04:06 PM
43% had the favorite winning.
I'd sample more races until that % comes down to a more reasonable number. I think there's probably value in the approach you described, but that 43% is what's making the difference look so enormous right now.

stu
11-09-2007, 04:53 PM
I added Wednesday November 7 to the sample.

44% exactas from 240 races

Place ROI = 3.03
Cramer ROI = 6.79

stu
11-09-2007, 04:58 PM
I'd sample more races until that % comes down to a more reasonable number. I think there's probably value in the approach you described, but that 43% is what's making the difference look so enormous right now.

43% isn't that out of line considering that I excluded all races where the post time favorite finished second.

Lets consider the old rules of thumb. The favorite wins a third of the races. Let's then assume that the post time favorite finishes second 20% of the time.

If we exclude 20% of the races (those with the fav placing), then the favorite should win 33.3/(100-20) of the races or 41.6%

I think that I am in the right ballpark for a spot check.

DJofSD
11-09-2007, 04:59 PM
I will use this strategy. It certainly cuts down on the woulda-shoulda-coulda factor.

ryesteve
11-09-2007, 05:02 PM
43% isn't that out of line considering that I excluded all races where the post time favorite finished second.
Good point...

gallahadion
11-09-2007, 05:07 PM
...to use this strategy (and I have cause i read cramers books years ago)
you have to think the favorite is legitimate or there would be no reason to bet this particular combination.

Stu, it might take some time but what if you looked at all those races and decided if the favorite was legitimate or not, would that raise your already outstanding roi?

stu
11-09-2007, 05:37 PM
True enough if you limited the play to legitimate favorites.

I must note that ones real life ROI will be much lower because
you must subtract the losses from the races in which the key
horse wins (as this is only a hedge bet) and from the races where the key horse finishes third or worse.

TEJAS KIDD
11-09-2007, 06:22 PM
STU,

Why are you throwing out the races where the favorite finished 2nd. If you're assuming that your horse finished 2nd in every race, then shouldnt you assume that your horse won the race where the favorite finished 2nd.
I did the same research because I am interested in this theory. On NOV 7th, I only found 133 races on Drf's site. The favorite won 48 at 36% rate. The return was $1425.50. The place horse in all of these races (including the winner's place price in races where the favorite finished 2nd) added up to $1002.50.
I am sure the ROI goes up much higher in races with bigger fields (as the exacta payoffs would be more), however, the % of hits would certainly decrease. I think that 133 or even 240 races is too small a sample, one big exacta can really taint the overall results. The place prices would remain at a constant small return.
Still an interesting theory. I would like to see a larger sample, if you're willing to do it. Took me too long to do 1 day.

DJofSD
11-09-2007, 07:33 PM
Why are you throwing out the races where the favorite finished 2nd.

If I recall correctly, it is because the original way Mark Cramer wrote about the bet was to play your long shot pick under the favorite -- not to box. Play your long shot to win with the hedge bet being the favorite+your pick exacta.

TEJAS KIDD
11-09-2007, 09:02 PM
If I recall correctly, it is because the original way Mark Cramer wrote about the bet was to play your long shot pick under the favorite -- not to box. Play your long shot to win with the hedge bet being the favorite+your pick exacta.


You didn't understand what I was asking. I understand Cramer's theory. What STU was doing was comparing the exacta payoffs ROI to the place payoffs ROI.
The exacta payoffs represent Mr CRAMERS thoery. The place payoffs represent what you would do instead of Mr Cramers Theory.

For example.
Player 1 plays $2 exactas putting the favorite on top of his selection
Player 2 plays $2 place bets on his selection (the same selection)
What STU was doing was assuming that every selection of player 2 was a winning wager and that Player 1 hit the exacta when the favorites won over the same selection.
What I was asking was why he dropped the races where the favorite finished 2nd. If he was assuming that all of the place wagers were winning wagers, then he should assume that the horses that beat the favorites (who finished 2nd) were the players horses that won.

chickenhead
11-09-2007, 09:11 PM
I agree, I think you would want to include the winners place price when the fav finishes 2nd for a more apples to apples comparison.

stu
11-09-2007, 09:38 PM
I am comparing scenrio in order to optimize.

My scenario is that I am betting #4 - Big Shot Bob to win provided that #4 is not the post-time favorite. Then I am hedging.

Scenario A - I hedge by betting #4 to place

Scenario B - I hedge by betting a one-way exacta the post-time favorite over #4

In both hedge scenarios, the only way that I collect is if the #4 finishes first or second. Scenario A will pay 100% of the time that the #4 finishes exactly second. Scenario B only pays when the post-time favorite wins and the #4 finishes exactly second. I can easily compare when my horse finishes second.

More later.

chickenhead
11-09-2007, 09:56 PM
right....but you are excluding a group of races where:

Scenario B has a guaranteed 0 return

Scenario A has a guaranteed non-zero return (unless you pick 0 winners)

in real life that will make a difference. how big...I don't know.

stu
11-09-2007, 10:23 PM
right....but you are excluding a group of races where:

Scenario B has a guaranteed 0 return

Scenario A has a guaranteed non-zero return (unless you pick 0 winners)

in real life that will make a difference. how big...I don't know.

Let's use Thursday only.

Post #1 describes getting paid 100% to place and x% Cramer Hedge when the key horse finishes exactly second.

139 Bets produced a gross of:

Place = 1087.00
Cramer = 2385.2

I acknowledge that I misaccounted for when the key horse finishes first and the place bet pays 100% of the time and the Cramer pays 0%. To estimate this I added the place payouts for all non-post-time favorite winners in the same set of races regardless of where the post-time favorite finishes. I acknowledge that is sloppy but it is still a sanity check.

101 Bets produced a gross of:

Place = 729.50
Cramer = 0* (*The only time the exacta also pays is when the key deadheats with the post-time favorite)

Combinations when key finishes first or second:

Place = 1816.50
Cramer = 2385.20

Place ROI = ((1816.5-(2x240)/(2x240)) = 2.78

Cramer ROI = ((2385.2-(2x240)/(2x240)) = 3.96

I will have to sleep on it for about a month or longer to figure out how to formally (dis)prove it.

In the meantime, let's assume that the population of place payouts for a key horse is roughly identical when the key horse finishes 1st as it is when it finishes second. In this case if the sum of the cramer exacta payout is greater than twice the place payout when the key horse finishes second, then Cramer's strategy is optimal.

My brain hurts,
Stu

chickenhead
11-09-2007, 10:41 PM
my wager is the place wins by a small margin (a little less than the difference in take).

not necessarily in real life..but if this test was carried out into a big sample.

I can't think of a good way to test it.

Robert Fischer
11-09-2007, 10:54 PM
interesting stuff , and good work by those who noticed the place-win payout issue...

Hey Stu have you ever tried using something artificial as a "selection process" for your studies? I notice you like to cut to the chase and skip that step.

I don't know if it would make any difference to use something like a beyer or a bris or whatever rating you wish ... and then use for example the the next highest ratingHorse other than the post time Fav as your "pick".

TEJAS KIDD
11-10-2007, 12:02 AM
No matter how this plays out, I would still not use this wagering advice. If I spend the time to handicap a race and have a strong enough opinion to INVEST my money, then I would most certainly not throw my money into a BLIND wager.

I have friends/family that do that sort of thing regularly. I'll call them and tell them that a horse is a lock. They ask me "Who do you like for 2nd?". I respond, "I'm not sure, just play him to win" After the race, I ask " Did you do well?". They say " No, I boxed him with the favorite"

Robert Fischer
11-10-2007, 12:14 AM
No matter how this plays out, I would still not use this wagering advice. If I spend the time to handicap a race and have a strong enough opinion to INVEST my money, then I would most certainly not throw my money into a BLIND wager.

I have friends/family that do that sort of thing regularly. I'll call them and tell them that a horse is a lock. They ask me "Who do you like for 2nd?". I respond, "I'm not sure, just play him to win" After the race, I ask " Did you do well?". They say " No, I boxed him with the favorite"

i see your point
most of the time either the posttime fav. is going to be part of your big exacta , or else is a big underlay that you want to oppose...

BMeadow
11-10-2007, 03:19 AM
From the August 1998 issue of Meadow's Racing Monthly (www.trpublishing.com):

Exacta As A Place Bet

One angle recommended by several racing authors is the “exacta as a place bet,” in which instead of betting a longshot to place you simply play him to win, then in the second slot of the exacta with the first and/or second and/or third choices. The idea is that if the horse runs second, the exacta is likely to pay much better than the place payoff.

True, but there are two major problems:

• The takeout is higher for exacta bets than it is for win bets.

• The horse might win, or run second to someone else, so you lose a bet you should have won.

Jim Cramer of Handicappers Data Warehouse reviewed tens of thousands of races to see if, indeed, a partial backwheel of a longshot in the two-hole might be a good strategy. We considered a favorite any horse ranked in the top three odds, and a longshot to be anyone ranked fourth through seventh (longshots eighth or worse in the field yielded many fewer plays and thus offered more skews, so we eliminated them).

Below are the results, with $2.00 break-even result averages rounded off to the nearest penny. For a comparison basis, betting all fourth through seventh choices to win yielded $1.59 and betting them all to place returned $1.49.



1st choice wins 4th choice runs second 2554 hits in 44,301 races $1.63

1st choice wins 5th choice runs second 1917 hits in 44,220 races $1.66

1st choice wins 6th choice runs second 1295 hits in 42,201 races $1.65

1st choice wins 7th choice runs second 780 hits in 35,473 races $1.58

2nd choice wins 4th choice runs second 1443 hits in 44,301 races $1.54

2nd choice wins 5th choice runs second 1065 hits in 44,220 races $1.54

2nd choice wins 6th choice runs second 756 hits in 42,201 races $1.60

2nd choice wins 7th choice runs second 479 hits in 35,473 races $1.59

3rd choice wins 4th choice runs second 988 hits in 44,301 races $1.46

3rd choice wins 5th choice runs second 708 hits in 44,220 races $1.37

3rd choice wins 6th choice runs second 496 hits in 42,201 races $1.45

3rd choice wins 7th choice runs second 328 hits in 35,473 races $1.60



There did not seem any significant advantage in the “exacta as a place bet” strategy

gallahadion
11-10-2007, 05:00 AM
I think you guys are missing the point. Cramer never intended for his exacta as place method to be analyzed this way. Each race is independent on its own merits and a particular longshot that you may play may not be the same that I would play. And, you don't know which races would be played and which races would be skipped. In summary, your analysis is flawed Sir.

I know Mark Cramer sir and you are no Mark Cramer! Good day sir!:D

dav4463
11-10-2007, 06:40 AM
If you handicap the race and your picks are normally longshots at 8-1 odds or higher, you can also handicap the three or four horses that you feel are most likely to beat your pick. Pick the top two out of those and put them on top of your pick. In this case, the exacta as place bet is a good bet if you make the right picks.

ryesteve
11-10-2007, 08:23 AM
There did not seem any significant advantage in the “exacta as a place bet” strategyDepends how you define "significant". Based on these figures, playing the horse under the favorite is providing a return that's 7-8% higher than betting to place. That's an advantage I'd like.

alysheba88
11-10-2007, 08:46 AM
Its not a "hedge". Just like betting place is not a "hedge"

The idea here was better to play the exacta in the second spot then place bet. But neither are "hedge" strategies

INFRONT07
11-10-2007, 09:47 AM
Thanks for the post;some of these guys just don;t get it:mad: If they want to follow a guy who rides a bike&lives in Paris good luck to them.you have proved the point.
THANKS:jump:

INFRONT07
11-10-2007, 10:00 AM
I think you guys are missing the point. Cramer never intended for his exacta as place method to be analyzed this way. Each race is independent on its own merits and a particular longshot that you may play may not be the same that I would play. And, you don't know which races would be played and which races would be skipped. In summary, your analysis is flawed Sir.

I know Mark Cramer sir and you are no Mark Cramer! Good day sir!:D
NO HE;S A REAL PLAYER &THE BEST MIND IN RACING.YOU SIR ARE A FOOL:bang: :bang:

Tom
11-10-2007, 10:55 AM
My 2 cents - if you want to test a method, use the PPs not the charts.
Write done every play before the race is run, then you will have meanigful results.

And Infront, I forget, what are some of the titles you have published?

shanta
11-10-2007, 11:07 AM
My 2 cents - if you want to test a method, use the PPs not the charts.
Write done every play before the race is run, then you will have meanigful results.


Right on the money Tom.

Light
11-10-2007, 11:17 AM
Jim Cramer of Handicappers Data Warehouse reviewed tens of thousands of races to see if, indeed, a partial backwheel of a longshot in the two-hole might be a good strategy.

Your results are what one would expect with longshots based soley on odds,(otherwise everyone would have caught on by now and bet this way if it was profitable)The real study should be based on longshots who are also top rated in some form of handicapping method be it man or computer generated.

chickenhead
11-10-2007, 01:38 PM
thanks for sharing that Barry. I agree with rye, the 7-8% boost in place ROI is interesting. Worth checking further with real plays.

Kelso
11-11-2007, 01:04 AM
I have friends/family that do that sort of thing regularly. I'll call them and tell them that a horse is a lock. They ask me "Who do you like for 2nd?". I respond, "I'm not sure, just play him to win" After the race, I ask " Did you do well?". They say " No, I boxed him with the favorite"


I think you should cut off those ungrateful (expletives). Tell me, instead. I promise I'll bet every one of 'em on the nose. Then we'll share a hearty laugh on 'em!! :cool:

pandy
11-12-2007, 11:39 PM
Here's an example of what I do. Besides my regular prime bets, I make many small wagers on longshots in maiden races, usually first time starters. They win at low percentage but show a nice profit. I can spot these plays pretty quickly. Once I've indentified the key horse, I spend just a minute or so handicapping the race, looking for a horse that jumps out at me as the main contender. About 70% of the time the main contender is the obvious favorite anyway. But regardless, I box it with my key horse. If the race is contentious and a main contender doesn't stick out to me, I simply bet my longshot win and place. I think you'll find that if you are a good handicapper you will do better picking your own exacta horse rather than arbitrarily using the favorite over your key. I also think that you should box it, not just use the favorite on top. If you are going to show a profit on win, you will show a profit on the exacta too. Try it on paper. Even if you just use the favorite, box it.