Extreme Favorites

[sligg]

I'm back to square one on money management. Apparently, if the win % is high, there is no enhancement using money management. I subscribed to both HMI manuals and the CD. And I was not able to improve the results of a test sample.

I used the data files from the TSN Advantage $59.95 for the month of March 2007. This was my sample betting multiple tracks. Horse had to qualify from two of the data files I used and had to go off as favorite.

Post time favorite, morning line any value:

Bets: 131
Won: 76
Win %: 58.0
$ Won: 273.1
ROI %: 4.2

From this I extracted a subset where the post time favorite had to also be the morning line favorite with odds less than 2/1: These results are impressive:

Bets: 74
Won: 54
Win %: 73.0
$ Won: 177.7
ROI %: 20.1


I put these subset numbers into the HMI spreadsheet and found that the results were the same as flat betting. There was no improvement using HMI.

It seems if the win % is high enough, flat betting is the way to go. No spreadsheet, no calculations, and no money management. As Dave Schwartz said, "Mortgage the house, start betting and get rich."

If it was only that simple. My sample was for one month. Will the next sample be equally as good? I don't know and I will probably never find out. I'm a $2 bettor and paying $59.95 for the TSN data files I would still lose money. If I increased my bet to $4, I would break even. Above $4, I would profit.

On to the next system.

[Dave Schwartz]

Ah, so what you sent me was the result of back-fitting.

I would say that if you have a 73% hit rate and a 20% advantage you do not need money management. You need a CPA.



Why don't you run your original sample against HMI? This one:

Bets: 131
Won: 76
Win %: 58.0
$ Won: 273.1
ROI %: 4.2

That is something HMI can help with.

[mhrussell]

...is way too small to draw any statistically significant conclusions. You need thousands of races.

But I have favored and used flat betting for the last 4 years after tinkering with different variations.

I recommend just make it easy on yourself...and I've found psychological advantages too. I am not second guessing myself out of bets with flat betting... if your analysis shows you have strong 'numbers' AND price.... JUST BET IT.

Relax. Do good work. Make good bets. In a hundred years, who's going to care?

Some economist had a famous qoute: "In the long run, we are all dead."
So true.

[sligg]

Ah, so what you sent me was the result of back-fitting.

I would say that if you have a 73% hit rate and a 20% advantage you do not need money management. You need a CPA.



Why don't you run your original sample against HMI? This one:



That is something HMI can help with.

Dave,

There was no BACK FITTING. These were actual bets that I made at $2 for the 31 days in the sample.

I will run the 131 race sample through the HMI spreadsheet and will post the results.

[sligg]

...is way too small to draw any statistically significant conclusions. You need thousands of races.

But I have favored and used flat betting for the last 4 years after tinkering with different variations.

I recommend just make it easy on yourself...and I've found psychological advantages too. I am not second guessing myself out of bets with flat betting... if your analysis shows you have strong 'numbers' AND price.... JUST BET IT.

Relax. Do good work. Make good bets. In a hundred years, who's going to care?

Some economist had a famous qoute: "In the long run, we are all dead."
So true.

Matt,

I agree with you but I don't have enough time left in life to run thousands of races. For me the fun in racing is doing the research and betting my $2, especially that I can bet online.

[Dave Schwartz]

After looking at the second sample (of which the first seems to be a subset), I have concluded that it is just too consistent.

There are no ups or downs, just a smoth upward trend.

You need to just become wealthy from this.

[Broken]

You need to just become wealthy from this.

I'm sorry, I just fund that amusing.

Seriously though, I know sharing secrets is not done in this kind of business, but would someone care to enlighten me on my the above numbers are not considered fantastic?

A 58% win percentage seems to be incredible to me???

[sligg]

I'm sorry, I just fund that amusing.

Seriously though, I know sharing secrets is not done in this kind of business, but would someone care to enlighten me on my the above numbers are not considered fantastic?

A 58% win percentage seems to be incredible to me???


Remember, these results were for one month only and betting extreme favorites.

The 58% win rate was for the original sample. The subset win rate was 73%.

[Edward DeVere]

"Burton P. Fabricand studied 10,035 races and . . . found that bets on favorites paying extremely short prices (1-2 or less) resulted in a small profit."

Tom Ainslie, Ainslie's Complete Guide to Thoroughbred Racing.

[Robert Fischer]

ML FAV(<2-1) ----> POST TIME FAV

It is a tad inflated, but you are on the right track to maximum win percentage. This is very similar to what I did when playing around with some variations of a martingale system where win percentage is the limiting factor.

To add another level to your filter you can use a rating system. Any power rating or speed rating or pace figure will do. From that point you can even do more- for example; say that the horse must have the best power rating by 10pts (or best last beyer by 5pts) or must have best prime by xpts and best latepace fig by x etc... you get the idea.

You then eliminate factors chaos factors. No fields with more than 8 horses, No maiden special weight...
You can get the win percentage pretty high, although you may have had a better than expected sample in your 74 tested wagers, and probably should expect somewhere around 55%.

TopBrisPrime AND ML FAV(<2-1) ----> POST TIME FAV

[Kelso]

Bets: 131
Won: 76
Win %: 58.0
$ Won: 273.1
ROI %: 4.2

<snip>

Bets: 74
Won: 54
Win %: 73.0
$ Won: 177.7
ROI %: 20.1

Sligg,
How many cards ... and from how many tracks ... did you review over the 31 days to find the above-referenced bets?

Thank you.

[sligg]

Sligg,
How many cards ... and from how many tracks ... did you review over the 31 days to find the above-referenced bets?

Thank you.


I selected from at least 15 tracks listed in Brisbet for the 31 day test. I did not fine tune as far as class or distance or size of field. The key between both samples was the morning line odds which had to be less than 2/1 and then go off as the post favorite. The average mutual of the winners was $3.29 which means you need better than 62 % winners to show a profit. Place and show bets both showed losses in my samples.

I will run another test starting May 1, 2007. Do I expect to maintain more than 70% winners? No. Horse racing can never be that simple. This in effect is a black box with no handicapping involved. Maybe I just got lucky for the one month.

[Kelso]

I selected from at least 15 tracks listed in Brisbet for the 31 day test. I did not fine tune as far as class or distance or size of field. The key between both samples was the morning line odds which had to be less than 2/1 and then go off as the post favorite.



Thank you, Sal. I'm trying to get a handle on how common this angle might be. Am I in the ballpark that you looked at something around 2400 races (average 4 cards/week/track .... so average 8.5 cards/day at maybe 9 races/card .... 8.5 x 9 x 31) to find the 74 wagers you placed?

Thanks again.

[porchy44]

Over a larger sample of races, YOU WILL "regress to the mean".

[Robert Fischer]

I suggest being very selective in your filter process. Use things like field size. Plan it out as you would actually carry it out. Less plays is actually a good thing if you are more profitable and spend less time watching the odds to see if a horse goes off as post time fav.
The goal is not win percentage by itself, it is ROI for the method, so you may want to consider things like a minimum odds threshold and see if that improves ROI - then you will have to be realistic - you cant just say I am not betting anything less than or equal to 0.6-1. You have to estimate a real post time odds that you would see and allow enough additional odds for an expected odds drop after post. Hes 4-5 on the tote the race goes off and you read in the charts that he was 0.6-1 or .5-1...

[racefinder2]

Is somebody advocating betting to win on heavy favorites??????

Try to clear your head and figure out exactly what it is you're hoping to accomplish with this madness.........before you do it with actual $......

[sligg]

74 bets in 31 days scanning 15 tracks would indicate that this is not a common angle. There were days with no bets. It sure is selective.

I'm disappointed that Dave Schwartz's HMI did not work with my two samples.
It did no better than flat betting.

For my next sample which will start May 1, I need to devise a betting plan using flat bets. In the 74 race sample there was a 3 bet runout, twice and there were two 13 race consecutive wins. At 73% winners, that is not surprising.

[SMOO]

I actually think this will do better the larger the field. 2/5's in 4 horse fields seem to lose more than 2/5's in 10 horse fields.

[Robert Fischer]

I actually think this will do better the larger the field. 2/5's in 4 horse fields seem to lose more than 2/5's in 10 horse fields.

When looking at past performance i will love a horse that was 2/5 in a 10 horse field and did well.
I think the horse has to be REALLY well liked by the public, but my guess is that they win less. The average favorite is a higher price in larger fields so simply looking at field size and favorite win% will not answer this. The last time I recall seeing this was Le Dauphin in a 12 horse field at gyulfstream going off at 4/5. An allowance n1x and LE DAUPHIN had horses like Sedgefield and Twilight Meteor? in his running lines, He finished fourth and was one helpful key in opposing Sedgefield later at Keeneland.
Maybe someone has hard data on on win% for a given odds at variable filed size ?

[Dave Schwartz]

I'm disappointed that Dave Schwartz's HMI did not work with my two samples. It did no better than flat betting.

LOL - You keep posting this "HMI disappointed you" message as if one would expect to do better with a sample that got 73% winners and a $2.40.

Why would you need to improve on that?


HMI is designed for players who need improvement. It depends upon streaks, both good and bad, to do its work. You have no streaks... just straight up growth.

HMI was not designed for anyone as successful at racing as you.

[SMOO]

When looking at past performance i will love a horse that was 2/5 in a 10 horse field and did well.
I think the horse has to be REALLY well liked by the public, but my guess is that they win less. The average favorite is a higher price in larger fields so simply looking at field size and favorite win% will not answer this. The last time I recall seeing this was Le Dauphin in a 12 horse field at gyulfstream going off at 4/5. An allowance n1x and LE DAUPHIN had horses like Sedgefield and Twilight Meteor? in his running lines, He finished fourth and was one helpful key in opposing Sedgefield later at Keeneland.
Maybe someone has hard data on on win% for a given odds at variable filed size ?
I have no hard data to support my theory, one would have to compare all horses that went off at exact subsets (.40 or .30 for example). I'm just assuming more "dead money" on extra longshots in the bigger fields, while the 4 horse race example would have fewer dead horses.

[jma]

I have no hard data to support my theory, one would have to compare all horses that went off at exact subsets (.40 or .30 for example). I'm just assuming more "dead money" on extra longshots in the bigger fields, while the 4 horse race example would have fewer dead horses.

But a lot more things can go wrong in a 10-horse field than a 4-horse field. Of course if you wait for races with 2-5 favorites in 10-horse fields, you won't be doing a lot of wagering anyway, so there's time to do research and find out. :)

[SMOO]

But a lot more things can go wrong in a 10-horse field than a 4-horse field.
Yes, but my point is that there is a bigger talent difference in the bigger fields. It does not take much to make a horse odds on in a small field.

[Bill Olmsted]

74 bets in 31 days scanning 15 tracks would indicate that this is not a common angle. There were days with no bets. It sure is selective.

I'm disappointed that Dave Schwartz's HMI did not work with my two samples.
It did no better than flat betting.

For my next sample which will start May 1, I need to devise a betting plan using flat bets. In the 74 race sample there was a 3 bet runout, twice and there were two 13 race consecutive wins. At 73% winners, that is not surprising.

Sal, Baby! Sweetheart!

[pandy]

You did.

[AwolAtPA]

74 bets in 31 days scanning 15 tracks-----
---
For my next sample which will start May 1,
I need to devise a betting plan using flat bets.
In the 74 race sample there was a 3 bet runout, twice
-------


great result Sligg,

may I suggest that you ignore all the suggestions about how to change your system that is working!! I agree with David and say just decide what your biggest bet you are comfortable with now. Maybe just double (yup Four dollars is a good test number) or up to ten or twenty or whatever you feel ok betting.

now my opinion about a plan. Use your history which says your longest string of loses was three. Ok, set your bankroll at four or maybe five times the unit bet and start with two bankrolls. If something is really different, then you will lose your first bankroll. If you lose the second bankroll before the end of the month then you have confirmed the history has changed. However, with another month like the FIRST month, you can have fun betting and fine tuning your system for betting solid/extreme favorites.

for example, Bet unit 10, Betting Bankroll 50 betting, Reserve Bankroll 50

oh, yes the plan does not include the longest string of collections because that is just neat bragging material which you expect will be repeated!!

so, there is my opinion about a plan for you.

have fun betting your new system.

AwolAtPA/AWOL/duane
A--Winner--Or--Loser

[andicap]

I just wish I had a dollar for every spot play I developed (low-priced horses, not one bomb that made it profitable) that worked for a month --

and then proceeded to lose the take and more over the course of a year.

[Topcat]

Sal,

I hope this is not beating a dead horse but your sample size is so small as to be almost meaningless-and then to backfit a subset of it makes it more so.

Try this reality check: If the numbers you submitted held up you wouldn't need any money management system-just a Brinks truck. With a win rate like that you could easily bet $1000/race (your only limitation would be the size of the pools-after all you will only lose 1 out 4 bets and and only 3 in row) so even taking lengthy vacations and not bothering with rebates you would win over $100k.

Oh and I would say that HMI does work better with a lower win% or a least % and payoffs that vary.

[sligg]

LOL - You keep posting this "HMI disappointed you" message as if one would expect to do better with a sample that got 73% winners and a $2.40.

Why would you need to improve on that?


HMI is designed for players who need improvement. It depends upon streaks, both good and bad, to do its work. You have no streaks... just straight up growth.

HMI was not designed for anyone as successful at racing as you.

Dave,

A one month sample is not my idea of success at racing. Nowhere in the hype for HMI was it stated that it would not work if you had a high win %. I'm not going to do the research but it's my feeling that HMI will not be effective with a win % of 60 or better.

[sligg]

I just wish I had a dollar for every spot play I developed (low-priced horses, not one bomb that made it profitable) that worked for a month --

and then proceeded to lose the take and more over the course of a year.


Andicap,

You are so right. I have experienced this scenario many times myself both in horse racing and sports betting. But what motivates me is that "Maybe, this time it's different." It's never different.

[Dave Schwartz]

Sligg,

A one month sample is not my idea of success at racing. Nowhere in the hype for HMI was it stated that it would not work if you had a high win %. I'm not going to do the research but it's my feeling that HMI will not be effective with a win % of 60 or better.

HMI depends upon ups and downs. The sample you sent me simply has no ups and downs to speak of. With a profitable 73% hit rate (and I keep being as polite as I can here to not challenge your ability to maintain such a hit rate with profit), you will need a very large sample to find ups and downs.


You do seem a bit obsessed here with HMI when you should really be obsessed with wagering your "magic beans" system.

Perhaps you need to read the HMI Promo Page (http://www.horsestreet.com/products/hmi/index.html) again.

Permit me to quote:

You become the fund selector, deciding what to invest in, while HMI becomes the fund advisor, deciding, how much to invest.

Let me make it perfectly clear. You do not need a fund manager with a 20% advantage and a 73% hit rate. You simply need to put all your money into play.


As far as I am concerned, this sample is not meaningful but hypothetical. Permit me to give you another sample that HMI cannot outperform.


Bet# Payoff
1 $12.00
2 $4.00
3. $10.00

See, when you win them all you don't need no stinkin' money management. You also don't need no MoMan when you almost win them all at a price which would make us all envious.


My guess would be that if you actually play this system for real ove rthe next 6 months one of two things will happen: You will either make so much money from it that you will become wealthier than you already are or you will come to appreciate how important extracting growth from the ups and downs really is.


My Best to You.

[badcompany]

Do you research how those 10% and up jockey do on big favorites? Certain good jockey and drivers seem to dislike winning on big favorites.