Need a reality check

[dav4463]

I've been on a roll ! I'm trying out a new (for the sake of a better word) method for selecting my exacta horses. Through 54 races using 3 horse exacta boxes: $648 bet $1593.60 return. I've never made this type of profit on exactas in my life. Can it possibly continue even close to this? The key is ranking the horses and then throwing out the obvious and looking for the best chance to pull off an upset. So far.... 15 exacta hits in 54 races with an average payoff over $106. I have a feeling I've just been lucky, but there is a method that I am following and so far so good....I'm not above enjoying the profits of a short-term result!

[midnight]

You have the right idea, but I doubt you'll continue at anything close to the same rate of ROI.

[Geekyguy]

I've been on a roll ! I'm trying out a new (for the sake of a better word) method for selecting my exacta horses. Through 54 races using 3 horse exacta boxes: $648 bet $1593.60 return. I've never made this type of profit on exactas in my life. Can it possibly continue even close to this? The key is ranking the horses and then throwing out the obvious and looking for the best chance to pull off an upset. So far.... 15 exacta hits in 54 races with an average payoff over $106. I have a feeling I've just been lucky, but there is a method that I am following and so far so good....I'm not above enjoying the profits of a short-term result!

I'd say keep playing the method until it gives you a reason not to.

Whatever you do, DO NOT SELL THE METHOD as that will kill its value.

[twindouble]

I've been on a roll ! I'm trying out a new (for the sake of a better word) method for selecting my exacta horses. Through 54 races using 3 horse exacta boxes: $648 bet $1593.60 return. I've never made this type of profit on exactas in my life. Can it possibly continue even close to this? The key is ranking the horses and then throwing out the obvious and looking for the best chance to pull off an upset. So far.... 15 exacta hits in 54 races with an average payoff over $106. I have a feeling I've just been lucky, but there is a method that I am following and so far so good....I'm not above enjoying the profits of a short-term result!

I don't think it's just the wagering strategy that's winning for you, it's good handicapping as well, just baced on what you said. I might add, if your skipping races where the risk is higher, your on your way to becoming a good gambler. Myself, I'd take more tickets on the ones where as you put it, " chance to pull off an upset."


Good luck,

T.D.

[Light]

The key is ranking the horses...

Ranked according to what?

[Handiman]

The four legged ones get top ranking..the three legged ones not so much...:lol:


Sorry, just needed a little levity this morning, the way my wife and kids have been treating me, plus the horses, I feel like a warm steaming pile that just fell from the south end of a last place finsher on a very cold morning.

Handi

[cato]

Congratulations! Run with it as long as you can!
Be sure to document what is working--write out the esact procedures you are using. My tendency has been when I uncover something like this and have success for awhile, I become overly aggresive and bend the rules to get more bets in and then --boom the edge is gone--

SO stay true to what works

Having said all that my guess is you start coming back to reality in another 50 races or so, reversion to the mean and all that, but as Geekguy said go with it until it proves to not work any more.

Cato

[dav4463]

Ranked according to what?

The main thing I use to finalize contenders is class moves as I try to see what the trainer is doing and whether or not the horse is spotted well.

The rankings are very basic. I have four main categories that I use to rank the horses. If the horse ranks one or two in any of the four categories, he is a possible contender. I use speed (I like the Beyers!), form, running style, and connections. Nothing fancy, but the final contender selection is pretty easy once I've narrowed it down to three or four horses. If the horse's four rankings add up to ten or less (example: ranks number one in all four categories= 4 points) or the horse has the best last race Beyer, I throw them out of the exacta if I can find anything wrong with them. If I can't find a reason to throw them out, I pass the race. If a horse does not rank top 4 in a category, I call it a six. example: rank of (1, 3, out, 4) would equal 14 (higher than ten means possible contender unless he has best last race Beyer).

I then play exacta boxes of top three horses when the probable payoff is $50 or more at 3 mtp. Occasionally I will include a fourth or fifth choice for second if it looks like a $100+ exacta chance.

[NYPlayer]

I've been on a roll ! I'm trying out a new (for the sake of a better word) method for selecting my exacta horses. Through 54 races using 3 horse exacta boxes: $648 bet $1593.60 return. I've never made this type of profit on exactas in my life. Can it possibly continue even close to this? The key is ranking the horses and then throwing out the obvious and looking for the best chance to pull off an upset. So far.... 15 exacta hits in 54 races with an average payoff over $106. I have a feeling I've just been lucky, but there is a method that I am following and so far so good....I'm not above enjoying the profits of a short-term result!

Report back after you have 500 races. I'll be waiting.....

Good luck.

[andicap]

In terms of ranking horses, you should read Dave Schwartz's stuff on Reynolds rankings -- of course don't read it unless things go south!! Don't change a thing now.
Its a different way of rating horses once you have final rankings. Instead of adding the rankings you multiply them, I believe.

Another thing to consider is if you have 1 low odds horse and two longshots you may want to play the race. By low I mean 2-1 or higher depending on field size, etc.
Check out what a 2-1/10-1 can pay in a 10-horse field.....

[Hosshead]

Dave, Congratulations! Keep on doing what you're doing. Only time will tell if you're on a "streak".
If you want to analyze it further, I would say, try to "nail down" what you really are doing, as much as possible. Especially the part in the "grey" (subjective) area.

Possible grey areas might include: (underlined)

"The key is ranking the horses and then throwing out the obvious and looking for the best chance to pull off an upset".

"I try to see what the trainer is doing and whether or not the horse is spotted well".

"I use speed, form, running style and connections"..

"I throw them out of the exacta if I can find anything wrong with them".

"If I can't find a reason to throw them out, then I pass the race".

I think if you can quantify some of these decisions, it will take some pressure off. Otherwise you'll be going with your "gut feeling" in the "grey" area, which a lot of cappers do. But can cause a "stomach ache".

.... I've never made this type of profit on exactas in my life. Can it possibly continue even close to this? !Beyer called this the "Messiah Complex" <g> :) Good Luck.

[dav4463]

Thanks for all the great replies! I will definitely report results every 100 races or so. The last 14 races this weekend were pretty good. I missed 10 in a row, the top Beyer horse won seven times! Then hit for $66, missed two more and then hit for $176 . I didn't get to play Sunday or Monday, but will keep plugging away. I also caught a $21.40 winner, but missed the exacta. I still should remember to put something on the win end I guess! Only 2 for 14 this weekend, but still a profit.

[dav4463]

Where do you get the Reynold's Rankings? I like to read everything.

[bobbyb]

Good Fortune Dave!!

You can read the Reynolds article @

http://www.horsestreet.com/freestuff/articles/reynolds/index.html

bobbyb

[Light]

Interesting article but I see 1 big flaw. I also use modeling but not so general. My modeling is narrowed to a track based on class,distance,surface,sex, and age. Can't see lumping all the different races on a card and getting a handle on anything. As I said in another thread,I have started modeling the individual BC races over the last 3 years with very good results. It actually reduced the contenders of the Pleasant Home race to four which included the top 2 finishers. (Allthough I didn't bet it). If you took the entire card as a model as Dave suggested,there would just be too wide a range with favorites and bombs so that you could never get the contenders reduced to 4 in a race such as with PH.

And dav4463,I understand your excitement. I started modeling 3 years ago and have never looked back. By far the best form of handicapping.

[toetoe]

Light,

In these troubled times, please refrain from use of the word "bomb" to describe a horse.

[Red Knave]

... I see 1 big flaw. ... Can't see lumping all the different races on a card and getting a handle on anything.
Light, I would suggest re-reading the article. There is nothing said about lumping all races together. The example shown is for races at 6F at HOL and is just that, an example. I think you can extrapolate it to whatever track/surface/distance etc. that you want.

[dav4463]

Still winning ! I hope it continues. Looking at the cards I've handicapped, usually 3 or 4 races on the card have a Legitimate contender that is either the Best Last Race Beyer or scores 10 or less in my ranking. By passing those races, the profit is better. I have to total up some results, but so far so good, I will continue to ride this streak until reality bites me !

[andicap]

IMHO, discipline is one of the reasons Dave is doing well.

We've all totaled up our annual records and said, "IF only I hadn't played those cheap action bets, I would have been in the black."
By demanding to be paid well and staying away from races that don't offer profitable opportunities at a fair price, Dave is doing the right things.
Of course his handicapping has to be decent as well.

[Light]

Red Knave

I'd like to see your point but I can't find it. It still look like DS's example lumps all the races together on 1 card as he says this:

If we were to describe the profile of a winner in these 8 races, we could say:

1.Must have a 4 or less in EP.

2.Must have a 4 or less in SP.

3.Must have an SPxCL of 5 or less.

4.Must have an EpxSP of 6 or less.

This is for an entire 8 race card at HP that obviously has different classes,ages,sexes,distances and surfaces. I don't see anything refering to these categories.

I'm not trying to argue,but if people here do get interested in "modeling",I can assure you it would be a waste of time and money if you went by that example which clearly lumps all those races and therefore categories together.

My personal experience in modeling shows if I take all the races in my database,there is chaos.This is similar to taking 8 races at Hol.There are many more variables that haven't shown up yet. If I narrow it down to something specific,say Males 6250NW2L 3yo&up
6f on the dirt ,suddenly,you find a revealing pattern you can work with. That's where you find the power of modeling.

What first sold me on modeling was when I hit 2 longshots on 1 card when I tried it at MNR. I had never did that before.The model said that longshots would win these races and narrowed it down to 2 horses in each case. Again the model was specific to the race,not the entire race card which would not have given me these winners.

[Light]

Yeah,I know the article says 6f races,but that's still way too general.

[Light]

One other point. You don't make pars based on lumping all 6f races together. Why would you model this way?

[Tom]

I break my model down by age, winners and non winners. Anything more than that, I think, defeats the purpose of a short term model.

Depends on how you use your model, I guess.

[Red Knave]

... You don't make pars based on lumping all 6f races together. Why would you model this way?
Nowhere does the article say that. You're taking it too literally. The article really isn't so much about modelling as it is about how powerful compounding the rankings can be.

[bobsbet]

Dave:

Sounds like you have something good going on.

Would you please clarify the area in your post where you say "If a horse does not rank top 4 in a category, I call it a six etc.

thanks

bob

[classhandicapper]

I love the idea of modeling races because I think most handicappers think in more general handicapping terms and not the specifics often required for certain types of races and/or circumstances.

Speaking of the Breeder's Cup races.

IMO, the probability of a fast pace is much higher in the juvenile races than in the typical race. Look at the PPs of the typical horses every year. The races are almost always loaded with sprinters that are trying to stretch out. Many of them are pace pressers and a few often actually have enough early speed to lead in sprints. The combination of a large field and all that potential speed usually leads to enough of a pace to eliminate plenty of 2nd string speed horses that look like contenders.

[46zilzal]

Tom babeis, still learning how to run, USUALLY run EARLIER their first few..don't you think?

[formula_2002]

In terms of ranking horses, you should read Dave Schwartz's stuff on Reynolds rankings -- of course don't read it unless things go south!! Don't change a thing now.
Its a different way of rating horses once you have final rankings. Instead of adding the rankings you multiply them, I believe.

Another thing to consider is if you have 1 low odds horse and two longshots you may want to play the race. By low I mean 2-1 or higher depending on field size, etc.
Check out what a 2-1/10-1 can pay in a 10-horse field.....

Wow, had no idea you were talking about “The Reynolds Number”.
First time I studied it was in high school over 50 yeas ago.
And to think what I have been missing.

RN= Dvp/u

Where D= diameter of pipe in feet(I don’t recall if it’s inside diameter or outside diameter. Probably ID)
v= velocity in feet per second
p=density in lb per cu. Ft
u= abs viscosity in lb per ft sec

Ofcourse I had to look that up. I didn’t remember it.

The one I remember it was the coefficient of lift for an airfoil.

It was a great high school.

However the most import formula to me is standard deviation (sqrt(n*p+(1-p))
I never appreciated it at the time.
Where n =number of trials
P=probability of winning

[dav4463]

Dave:

Sounds like you have something good going on.

Would you please clarify the area in your post where you say "If a horse does not rank top 4 in a category, I call it a six etc.

thanks

bob


If for example I am ranking the Best Beyer in last three races, I look at each horse's personal best of his last three races and mark it. Then, I rank the top 4, including ties. All those who rank top 4 would score their ranking, 1,2,3,or 4......all others would score a 6. So, in my four categories that I use for a basic ranking....if the horse is the best in each category he gets a 1-1-1-1=4.....I've found those that rank 10 or less usually show no value, so I try to beat them.
The best horses rank 1 or 2, in a category or two and usually not top 4 in one of the categories.....example a 2-1-4-6 scores 13 and approaches being a value bet. Those that score 19 or higher, but have at least one 1 or 2, are my best longshots.

You still have to look at the horse's that score 10 or less points and decide if you can beat them...I also look at the top last race Beyer to see if I can beat him. If I feel that they are the best....they are listed as LEGIT on my form and I will pass the race because my records show that I am losing money when I take fliers against horses who I have classified as LEGIT. Sometimes, I will bet the LEGIT horse to win if I feel strongly enough about him.

I normally find 2-4 LEGIT horses on an average racing card. 5 or 6 in California !

[formula_2002]

Dave, this is how I keep my “reality” in check.

Example:
Testing for exacta and win pool profits using top bris power figure and some basic handicapping concepts I call “x13”

Back fitted results using 2005 data;
394 races
814 plays
58 wins
$2263 in exacta pay-outs ($2 exacta base)
expected # of wins= 61 (a dutch bet based on the final win pool probabilities of the exacta horses)
flat bet exacta profit =39%
dutch bet loss =5%
win pool profit on the top horse = 2%.

Previously untested results 2002, 2003 and 2004 data

472 races
831 plays
53 wins
$1585 in exacta pay-outs ($2 exacta base)
expected # of wins 69 (a dutch bet on the final win pool probabilities of the exacta horses)
flat bet exacta loss =5%
dutch bet loss =23%
win pool loss on the top horse = 1%.

The top horse is always the favorite.
In the 2005 data the top ranked Bris power figure , running as a favorite lost 15%
In the 2002-2004 data the top ranked Bris power figure , running as a favorite lost 12%

What could one “normally” expect to win in the 2002-2004 data.
If I allow my self to average the exacta pool odds ( I hate to average odds), I find that the average exacta probability is;

Probability sum = 37
37/831 = .0444
Where the exacta pool odds are based on normalizing the win pool odds and calculating the exacta probability.

Calculating the standard deviation,
sqrt(831*.044*(1-.044))=6

2 ½ deviations =6X2.5=9
37+9= 46
I won 53 races, well beyond the normal
In a fair game(no take-out), that would promise a large reward.
But because of the track take out, this is an unfair game.


I’m not too sure what this all means, so I keep my serious money in the rubber band.