Can Excluding Over-Bet Factors Make a Difference?

[Brett V.02]

Hi All,

Having read a lot of the previous threads, one aspect that has always intrigued me is the concept of filters.

My intention of this thread is to look at when a filter becomes 'over-bet' and then as such, becomes a detriment to our profit.

My first contribution is Barriers - I have read on here, and believe that the inside barriers are over-bet. So with this in mind I took a system that had a little bit of success, and applied the filter, barrier 6 or wider.

The S/R dropped slightly but the POT lifted at a lot higher level.

Now I will preface this by saying, I do not have a database and have used a free source to look at the figures and I doubt if it is overly reliable, but all I did was run the system, then add the filter and the results were positive.

I am interested in 2 things:-

1 If the database guys out there have ever run this filter and does it return similar results?

2 Are there any other 'over-bet' factors people believe can have a similar effect.

Cheers

Brett

[UselessBettor]

Hi All,

1 If the database guys out there have ever run this filter and does it return similar results?

2 Are there any other 'over-bet' factors people believe can have a similar effect.
1. Yes inside barriers are a well known to be overbet. The further out you get the more the profit it likely to rise.

2. Yes there are many. Some depend on the situation. For example first uppers who have won first up before are generally overbet, but resumers in general are underbet. There are many more situations like that as well.

[beton]

They are generally overbet for a reason. exploit the reason, don't go for next layer of low hanging fruit

[Brett V.02]

Hi Benton and UB,

I fear I have tried to feast on Low Hanging fruit for years............. ;)

UB, I know you used to have a great site to test systems on, what happened to it if you don't mind me asking?

[darkydog2002]

I found under $3 for me were overbet and unprofitable and since I now wont bet anything under that the profit has increased considerably..
This applies to both wet and dry conditions.

Cheers

[beton]

Beware that by saying "I will only bet above my breakeven point" only moves the breakeven higher and greatly reduces the bets. It does not improve the bottom line. Moving away from the low prices may bring it into profit, but get too greedy and you run the risk of losing out. You are better to ask why the bets are overbet in a similar manner that RP did for longshots. Remove some obvious deadwood to improve the bottomline.

[Brett V.02]

Benton,

What if I fear I am the DEADWOOD!!!!!! :)

[Lord Greystoke]

Benton,

What if I fear I am the DEADWOOD!!!!!! :)LOL you might be if you keep stuffing up his name?!

LG

[Brett V.02]

Wow,

Sorry LG and Beton............I have misread your name since the start of time

hahahahaha

[SpeedyBen]

Wow,

Sorry LG and Beton............I have misread your name since the start of time

hahahahahaYou can beton it, Brett.

[ianian]

By adding apprentice riders only, gains 2% in POT, place >70% adds over 2% in POT, not won at distance adds 2% in POT from this site somewhere.

I must say i thought it was benton as well - just goes to show how our mind views things, racing included.

[UselessBettor]

Brett,

I put it up just to help people out ( a few friends). The server guys said I was using too much CPU on the server and told me they wanted me to pay stupidly high for the service so it got pulled down.

Unfortunately when I looked at others based on the cpu they all said it was too high and I needed my own server. Too costly for me (as I already have access to the data) so I didn't put it back up.

The good thing was it exposed a few peope to system testing and they moved onto products like Chrome Princes which I have heard is a great product.

[Lord Greystoke]

Always wondered if there were any really good systems tested UB??

Cheers LG

[UselessBettor]

Always wondered if there were any really good systems tested UB??

Cheers LG
LG there were lots of great systems tested. It collected hundreds of unique systems that made a profit. There were probably thousands collected but some had minor differences like barrier >= 4 and then the next system was barrier >= 5.

A lot of them revolved around CP and CF in the neurals and the top Don scott rating for the Year Rating and Base Rating (rating before adjustments).

[Lord Greystoke]

UB I remember remember coming up with a few good ones for favorites over different price ranges & SRs but lost the settings in an old laptop which died in the **********. Ce la vie!

LG

[Brett V.02]

Thanks UB.

Will do some research on CP's database.

Note - To avoid a repeat of the Benton/Beton debacle, I will now be referring to everyone on the forum as champ, mate or cobber.

Thanks for your understanding.

Brett

[beton]

I don't know where the Benton came from. I had never heard of it before this forum and there has been a few instances. Just don't call me late for breakfast.
Fortunately I have a thick skin and I have realized that I have made more than one mistake (2 at last count)

[Brett V.02]

Well said COBBER...........

What a winner!!!!!! ;) If only I can have as much success on the punt tomorrow

[Brett V.02]

Now getting back to my original topic, I am left wondering as to whether combining 'over-bet' factors will produce greater success.

I know that this is not the case with many of the filters that can be used at the other end of this argument. While they can be successful individually, when combined they are no magic bullet.

So lets take these 3 'over-bet factors'

So a system that reads,

Barrier 6-24
First Up - Never won First Up
Ridden by a Claiming Apprentice

After 15 years - I give you the Holy Grail !?!?!?!?!?!?

If only it was this easy

[UselessBettor]

So lets take these 3 'over-bet factors'

So a system that reads,

Barrier 6-24
First Up - Never won First Up
Ridden by a Claiming Apprentice


contrary too most peoples belief if you add "must be fav" as a rule you will likely be on a winning system at top fluc.

[Rinconpaul]

contrary too most peoples belief if you add "must be fav" as a rule you will likely be on a winning system at top fluc.
Scanning the fields today, about the only selection that comes close is No 3 "Dragon Street" R2 Gold Coast B Wallace (A) Form 3x but it's got barrier 1 ?

RP

[Lord Greystoke]

Scanning the fields today, about the only selection that comes close is No 3 "Dragon Street" R2 Gold Coast B Wallace (A) Form 3x but it's got barrier 1 ?

RPLooks like a lottery RP = 5 newbies(technical term). Better still a field with only 1 FU or FS runner which starts favorite and you get on at top fluc?

LG

[Rinconpaul]

Do you think management should seek sponsorship from the suppliers of "No-Doz" ??

I mean, have a look at some of the post times:
UB 12:33am, Brett V.02 12:04 am, LG 6:47 am, Chrome Prince 5:54 am, Vortech 6:44 am...........??

I mean, you're going to need a supplement of some kind, to get through your betting day?.....lol

So long as it's ASADA approved boys!

RP

[beton]

Do you think management should seek sponsorship from the suppliers of "No-Doz" ??

I mean, have a look at some of the post times:
UB 12:33am, Brett V.02 12:04 am, LG 6:47 am, Chrome Prince 5:54 am, Vortech 6:44 am...........??

I mean, you're going to need a supplement of some kind, to get through your betting day?.....lol

So long as it's ASADA approved boys!

RPMore like a lump of 4 x 2 to make them sleep sometime. Mind you, go through the week and it looks like a lot of sleeping was caught up. I thought it was a ghost forum.

[Brett V.02]

Does that mean Late Night Flight is the omen bet of the year???

AR 6 no 1 - Late Night Flight - $15

Days since last run - 24 (1-21 days over-bet)
Days since last win - 889 (won in last year over-bet)
Barrier - 9 (Wider the better)
Age - 7 (Horses 6 or under over-bet)

Have a sneaking suspicion that this type of approach will take nerves of steal, veins of ice and deep pockets of plentiful cash (which may become less plentiful at a great rate!!!!)

[mattio]

Brett,

I put it up just to help people out ( a few friends). The server guys said I was using too much CPU on the server and told me they wanted me to pay stupidly high for the service so it got pulled down.

Unfortunately when I looked at others based on the cpu they all said it was too high and I needed my own server. Too costly for me (as I already have access to the data) so I didn't put it back up.

The good thing was it exposed a few peope to system testing and they moved onto products like Chrome Princes which I have heard is a great product.Hi UB,

If you are interested in getting it back online I would be able to host it for you at no cost.

Cheers,

Matt.

[The Ocho]

Hi UB,

If you are interested in getting it back online I would be able to host it for you at no cost.

Cheers,

Matt.
Now there's an offer too good to refuse. Good O mattio. :D

[Chrome Prince]

I mean, have a look at some of the post times:
UB 12:33am, Brett V.02 12:04 am, LG 6:47 am, Chrome Prince 5:54 am, Vortech 6:44 am...........??


Some live in Perth, so the time difference.
Some are bakers.
Some are early risers.

Me, I actually suffer from insomnia :D

[Magister Ludi]

1 If the database guys out there have ever run this filter and does it return similar results?

2 Are there any other 'over-bet' factors people believe can have a similar effect.
1. Barriers are a track-specific factor and not a universal factor.

2. You need to use metrics that are far-removed from what the public is using. The metrics that the public is using are already reflected in the tote odds.

[beton]

2. You need to use metrics that are far-removed from what the public is using. The metrics that the public is using are already reflected in the tote odds.
Pauline had a famous quote which I will pinch "Please explain."

[Magister Ludi]

Pauline had a famous quote which I will pinch "Please explain."Racetrack betting markets are like any other financial market. The price of a horse reflects all of the hopes, fears, dreams, WAG's, SWAG's, careful analysis, and silly guesses of all of the bettors on that horse and is reflected in the horse's price. The more popular metrics are more thoroughly incorporated into that price and provide no overlay opportunity. You need to find out what everyone else is doing and go the opposite direction.

[Lord Greystoke]

You need to find out what everyone else is doing and go the opposite direction.OR Find out what makes everyone else go in the opposite direction and stay put?*

One thing I have learnt about price of late is that are many versions of price to consider, other than SP. e.g. if one gets on early enough onn a fixed basis, LOT can be turned into POT and a low return can become acceptable or even 'considerable'

LG

*Assumption being that this doesn't affect a horse's true chances of winning too much!

[darkydog2002]

Eliminating under $3 on wet tracks made a major difference for me.

[Magister Ludi]

One thing I have learnt about price of late is that are many versions of price to consider, other than SP. e.g. if one gets on early enough onn a fixed basis, LOT can be turned into POT and a low return can become acceptable or even 'considerable'

LG
Astute observation, milord. Overlays have a way of doing just that!

[beton]

Racetrack betting markets are like any other financial market. The price of a horse reflects all of the hopes, fears, dreams, WAG's, SWAG's, careful analysis, and silly guesses of all of the bettors on that horse and is reflected in the horse's price. The more popular metrics are more thoroughly incorporated into that price and provide no overlay opportunity. You need to find out what everyone else is doing and go the opposite direction. Thanks Magister Ludi. My quest was that you proceed and nominate some of these obscure metrics to ponder.

[shifty]

I found under $3 for me were overbet and unprofitable and since I now wont bet anything under that the profit has increased considerably..
This applies to both wet and dry conditions.

Cheers
i follow your posts darky, and i remember very well your longshot winner at last years Melbourne carnival. and well done on getting rid of barney, the thread closed down and my post congratulating you got lost i think. its a better place now.

[Magister Ludi]

Thanks Magister Ludi. My quest was that you proceed and nominate some of these obscure metrics to ponder.One very useful metric is a quantitative measurement of race entropy.

[beton]

One very useful metric is a quantitative measurement of race entropy.
I, for one looked at this in wonder. I am sure that a majority here did the same. I was going to the standard Pauline response but instead I will simply request "Please quantify" One of my attributes is that I can lift heavy things.

[Magister Ludi]

A race with an infinite number of horses with an equal amount of money bet on each horse is a perfectly competitive race. If all of the money were bet on a single horse, it is a perfectly uncompetitive race. Most races, of course, are somewhere in between these two extremes.

The following is a formula that quantifies race entropy or competitiveness:

sum ([1/O(i)]^2)/n

where
O(i) = odds of the ith horse
n = number of entries

The larger the value, the more uncompetitive the race.

1. Convert the odds to probabilities (1/o = p)

#***o***p
1***2***1/2=.50
2***3***1/3=.33
3***4***1/4=.25
Total prob = 1.08

2. Normalize the probabilities

1/total prob = 1/1.08 = .926

#***o***p******normalized
1***2***1/2=.50***x.926=.46
2***3***1/3=.33***x.926=.31
3***4***1/4=.25***x.926=.23

Total normalized prob = 1.00

3. Square the normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

4. Average the squared normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

Average squared normalized prob = (.21+.10+.05)/3=.36/3=.12

The competitiveness index of this race is .12.

[Vortech]

My cat's breath smells like cat food


A race with an infinite number of horses with an equal amount of money bet on each horse is a perfectly competitive race. If all of the money were bet on a single horse, it is a perfectly uncompetitive race. Most races, of course, are somewhere in between these two extremes.

The following is a formula that quantifies race entropy or competitiveness:

sum ([1/O(i)]^2)/n

where
O(i) = odds of the ith horse
n = number of entries

The larger the value, the more uncompetitive the race.

1. Convert the odds to probabilities (1/o = p)

#***o***p
1***2***1/2=.50
2***3***1/3=.33
3***4***1/4=.25
Total prob = 1.08

2. Normalize the probabilities

1/total prob = 1/1.08 = .926

#***o***p******normalized
1***2***1/2=.50***x.926=.46
2***3***1/3=.33***x.926=.31
3***4***1/4=.25***x.926=.23

Total normalized prob = 1.00

3. Square the normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

4. Average the squared normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

Average squared normalized prob = (.21+.10+.05)/3=.36/3=.12

The competitiveness index of this race is .12.

[syllabus23]

Originally Posted by Magister Ludi
A race with an infinite number of horses with an equal amount of money bet on each horse is a perfectly competitive race. If all of the money were bet on a single horse, it is a perfectly uncompetitive race. Most races, of course, are somewhere in between these two extremes.

From there we morph into 1st year psychology statistics.

A few tips from "The Master Of The Game" would be appropriate at this point.

[Lord Greystoke]

My cat's breath smells like cat foodI find that very probable. The question is, have your normalized the outcome in this instance?

LG

[UselessBettor]

A race with an infinite number of horses with an equal amount of money bet on each horse is a perfectly competitive race. If all of the money were bet on a single horse, it is a perfectly uncompetitive race. Most races, of course, are somewhere in between these two extremes.

The following is a formula that quantifies race entropy or competitiveness:

sum ([1/O(i)]^2)/n

where
O(i) = odds of the ith horse
n = number of entries

The larger the value, the more uncompetitive the race.

1. Convert the odds to probabilities (1/o = p)

#***o***p
1***2***1/2=.50
2***3***1/3=.33
3***4***1/4=.25
Total prob = 1.08

2. Normalize the probabilities

1/total prob = 1/1.08 = .926

#***o***p******normalized
1***2***1/2=.50***x.926=.46
2***3***1/3=.33***x.926=.31
3***4***1/4=.25***x.926=.23

Total normalized prob = 1.00

3. Square the normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

4. Average the squared normalized probabilities

#***o***p******normalized**norm squared
1***2***1/2=.50***x.926=.46***x.46=.21
2***3***1/3=.33***x.926=.31***x.31=.10
3***4***1/4=.25***x.926=.23***x.23=.05

Average squared normalized prob = (.21+.10+.05)/3=.36/3=.12

The competitiveness index of this race is .12.

That's an interesting idea on classifying races. I think this could be extremely useful as its another input to look at to find a profitable angle.

[Magister Ludi]

As you can see, the competitiveness index (ci) is a race metric. This is a highly competitive race:

#***o
1***4
2***4
3***4
4***4
ci = .0625

This is a highly uncompetitive race:

#***o
1***1.1
2***21
3***21
4***21
5***21
ci = .1382

Question: does a 7:1 horse (for example) perform differently in a highly competitive race as opposed to a highly uncompetitive race?

[Magister Ludi]

1. Calculate the ci for several races.
2. Populate a table with your results. Use the ci for the rows and odds for the columns. The body of the table will be the roi for each combination of ci and odds. I guarantee you that you will be amazed at the results. It should look something like this (several columns omitted):

Return on Investment

****odds range***
****1-2*****8-9
1**********-14.32%
2*-13.70%**-16.89%
3*-14.24%**-19.69%
4*-13.96%**-19.88%
5*-13.39%**-19.57%
6*-13.19%**-25.16%
7*-12.65%**-18.55%
8*-13.39%**-27.85%
9*-10.32%**-37.51%

ci x 100
<ci x="" 100<ci="" 100<="" b=""></ci>

[beton]

As you can see, the competitiveness index (ci) is a race metric.

Question: does a 7:1 horse (for example) perform differently in a highly competitive race as opposed to a highly uncompetitive race?Thanks for your explanation Magister Ludi. This raises more questions.
Firstly we are adjusting to a 100% market. What odds are we starting with? Our ratings? PP? Open? T Fluc? SP? The key here is to square the normalised odds, hence the more lower odds, along with the sharper they are, load one side of the equation. This is then balanced or mellowed by the number runners.

So what is the effect of a one horse race (Black Caviar) in a field of 15, vs a 2 horse race (say $2.50 and $3) vs a 3 horse race or an even field?

Are we not comparing the value of our selection in comparison to all the other runners?

It is just as well that I can still go outside, pick up the tractor and put it on my shoulders and carry it up to the top paddock.

I look forward to the next instalment. It makes a welcome change from the naysayers and the "I have the holy grail but you have to find yourself" posters that we have had the pleasure (sic) to endure of late.

[Magister Ludi]

You can use any public odds that you like to calculate the race's ci.

To reiterate, a maximum entropy/minimum ci race is one where all horses have identical parimutuel probabilities. A minimum entropy/maximum ci race is one which has a uniform distribution of parimutuel probabilities over the interval 0 to 1.

Ceteris paribus, a longer field will yield a lower ci. Conversely, a shorter field will yield a higher ci.

The ci can used to determine the state of the tote board as well as a metric to index the races after the race has been run. It is a metric measuring competitiveness as determined by the bettors. It is also a concentration or dispersion of the odds in a horse race. The ci does not measure the strength of field. It measures the public's perception of the competitiveness of the field.

It is a way to measure more than just the favorite to determine the public's perception of the race's competitiveness.

[Lord Greystoke]

It is a way to measure more than just the favorite to determine the public's perception of the race's competitiveness.For lack of a better contribution here; I see that the 'handle' fits, Magister Ludi !

Cheers LG

[Magister Ludi]

For lack of a better contribution here; I see that the 'handle' fits, Magister Ludi !

Cheers LGThank you for the kind words, milord.

[beton]

You can use any public odds that you like to calculate the race's ci.

To reiterate, a maximum entropy/minimum ci race is one where all horses have identical parimutuel probabilities. A minimum entropy/maximum ci race is one which has a uniform distribution of parimutuel probabilities over the interval 0 to 1.

Ceteris paribus, a longer field will yield a lower ci. Conversely, a shorter field will yield a higher ci.

The ci can used to determine the state of the tote board as well as a metric to index the races after the race has been run. It is a metric measuring competitiveness as determined by the bettors. It is also a concentration or dispersion of the odds in a horse race. The ci does not measure the strength of field. It measures the public's perception of the competitiveness of the field.

It is a way to measure more than just the favorite to determine the public's perception of the race's competitiveness.
So once the handicapper has finished (please note the handicapping is done before the barrier draw) all horses are equal. However being smarter than the handicapper and with the benefit of the barrier draw, the average punter knows that whilst all horses are equal some horses are more equal than others.

This new found smartness (punters being smarter than handicappers) starts with the PP or punters own (or acquired) odds and goes through to the race's finish. This CI provides a method of measuring the degree of smartness, in each set of odds, vs the handicapper, in each race.

So you can have a CI on the PP, A/odds, open and SP and any other set of odds.

Anyone assessing a race can see that there is one, two or three horses favoured to win a race. This is easily seen in the price or price grouping (1, 2 or 3 horse represent most of the money in a race)

Thus I am assuming here that this CI measures how much a 1 horse race is so that it can be used to compare similar races. I am trying to see where this could be a benefit or is it a distraction.

[evajb001]

On the face of it the CI measure seems to me as being a numerical way of identifying how 'open' a race is in terms of the odds on offer.

For example a 12 horse race with the lowest priced fave at $5.00 vs a 12 horse race with a lowest priced fave at $2.50, if all other horses odds are the same you'd say the first race is more 'open' and hence would have a lower CI value.

I obviously haven't looked into this CI calculation or its use much at all, but on the face of it as i've said it seems like simply an indicator as to how open a field is. The benefit of this is that by having it as a numerical value you can base systems or selections around it like, only bet if the CI is below 0.15 as this represents an open race where I can achieve value on my selections.

By looking at a race yourself you should be able to determine this anyway, but by having it numerically represented in your face it could act as a more effect mental trigger possibly?

Anyway thats just my take on it based on the last 10-15 posts in this thread. Interested in further discussion though.