• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Types of system

decastro

decastro

I am a relative newcomer to horse racing and horse racing systems. When I started looking into it I had assumed that what everyone did was look at past data and then for future races, determine a probability of winning for each horse. What you then bet on would depend entirely on the value of the odds being offered. It could be that there are no "value" odds being offered for your highest probability horse, but good value for your second highest probability one, in which case you would bet on that. Lets call this system 'A'. By the way, what I have been developing over the past few months is exactly this.

Looking around it appears that most "systems" appear to be doing something quite different. They seem to be just choosing an individual horse to win. There appears to be no probability attached to each horse in the race. Let's call this type 'B'.

So my first question is about nomenclature, what do you call systems A and B?

My second question is - am I the only one attempting a type 'A' system? and do they exist commercially? I have never seen a website offering this.

And of course I am interested in any general observations anyone has about A vs B.
 
Last edited:
Hi decastro decastro , I think if you look deeper into it both types of system are looking for value.
System A is looking for value per race, System B for value over a number of races.

There are plenty of examples of both approaches on this forum , some people use a hybrid, it's what suits the individual I guess
 
decastro said:
I am a relative newcomer to horse racing and horse racing systems. When I started looking into it I had assumed that what everyone did was look at past data and then for future races, determine a probability of winning for each horse. What you then bet on would depend entirely on the value of the odds being offered. It could be that there are no "value" odds being offered for your highest probability horse, but good value for your second highest probability one, in which case you would bet on that. Lets call this system 'A'. By the way, what I have been developing over the past few months is exactly this.

Looking around it appears that most "systems" appear to be doing something quite different. They seem to be just choosing an individual horse to win. There appears to be no probability attached to each horse in the race. Let's call this type 'B'.

So my first question is about nomenclature, what do you call systems A and B?

My second question is - am I the only one attempting a type 'A' system? and do they exist commercially? I have never seen a website offering this.

And of course I am interested in any general observations anyone has about A vs B.
Click to expand...

Welcome to the forum decastro decastro with the first one 'A' i would say is more of a method outlook, i use this type of idea on my C/Ds or similar course type method in sprints, depending on the amount of races won with a given weight before plus other elements and then weigh up the value accordingly.

Most here with a system just check back past results for many years, and if have shown a good constant profit to a level stake then hope for the same in the future, no real guarantee mind you because circumstances change sometimes as we all know.

But as you state is a good idea to check on the prices back also as to maximise ones profits which ever approach is being used.
.
 
So how is your type A approach panning out for you, i.e. what is the strike rate for your top 3 probables. Something around one of the top 3 winning 50% of the time is about standard for most ratings
 
ArkRoyal said:
So how is your type A approach panning out for you, i.e. what is the strike rate for your top 3 probables. Something around one of the top 3 winning 50% of the time is about standard for most ratings
Click to expand...

I never considered strike rates before. So I added some code to calculate them and ran a test on a total of 1200 races spread over two separate periods (June&July 2014 + June&July 2015) - BTW, I had never analysed these periods before. My strike rate (by your definition) was 45%. I should say that my profit in both those periods was positive (assuming Betfair SP).
 
dave58 said:
There are plenty of examples of both approaches on this forum
Click to expand...

I've scarcely seen any discussion of a type 'A' system anywhere - so a link to a good thread about that would be appreciated... and info about any commercial service offering probabilities of winning for all horses in a race would be great too.
Thanks,
 
Chesham said:
Have you been applying Chi Scores to your systems and also checking the A/E it could be that a system fails to make a profit because it is over bet
Click to expand...
I think you may have misread my post - my system is making a profit already - at least the simulations are.
 
decastro said:
I've scarcely seen any discussion of a type 'A' system anywhere - so a link to a good thread about that would be appreciated... and info about any commercial service offering probabilities of winning for all horses in a race would be great too.
Thanks,
Click to expand...

Look under horse racing, there's ratings and utilities and also speed figures where you may find what you're looking for
 
Chesham said:
the Chi Score will tell you if it is down to luck or not.
Click to expand...
I do many things to ensure the statistical significance of my results. Chi alone can be misleading - if you randomly try enough different systems then you will eventually find one that looks like it wasn't a fluke (even though it is). I am acutely aware of this phenomena.
 
decastro said:
I think you may have misread my post - my system is making a profit already - at least the simulations are.
Click to expand...

I would not worry to much about the type then because the practical part it is working, have you checked back many years with it and not backfitted to come right from your orginal ideas or thoughts i think that is an important aspect to any systematic idea unless checking the best odds scales etc over time.
 
OnTheNod said:
have you checked back many years with it and not backfitted to come right from your orginal ideas or thoughts
Click to expand...

If I understand you correctly... On a daily basis I am training and testing my system on around 500 races in June and July. I am tweaking the system to be as accurate as possible for those two months. But I would never assume that the results in those two months represented the future performance of my system. Instead I would much less frequently (indeed so far I have only done this once - it takes many hours even on my super fast PC) test by running my system over 11,000 bets over 2.5 years starting from august 2013. My single big test showed a profit of 3.6% per bet - this is all assuming the betting odds were the Betfair SP.
 
decastro said:
On a daily basis I am training and testing my system on around 500 races in June and July. I am tweaking the system to be as accurate as possible for those two months.
Click to expand...

If I am understanding you correctly I believe you have a system 'tweaked' to show the best results over 500 races in June and July.
And when using these settings you can generate a profit of 3.6% against 11,000 races from 2013.

How much profit does your system show against the 500 races it has been trained against?
 
ArkRoyal said:
If I am understanding you correctly I believe you have a system 'tweaked' to show the best results over 500 races in June and July.
Click to expand...

Yes... but "best" does not mean "most profitable", instead it means something more like "the positions of all the horses in all the races is most in keeping with what I predicted".

ArkRoyal said:
And when using these settings you can generate a profit of 3.6% against 11,000 races from 2013.
Click to expand...

Yes.

ArkRoyal said:
How much profit does your system show against the 500 races it has been trained against?
Click to expand...
that jumps up and down like a yoyo. Tiny adjustments in my system can have huge effects on the profit. I mostly ignore the profit. If I had the choice between A) predicting the winner in a race and having a badly wrong rankings for the rest of the horses in the race and B) getting the winner wrong but having almost correct rankings for every single horse, then I choose B.
 
Chesham said:
It reads like you are using AI Program (s) Probably William Benter is more in line with what you are looking for
Click to expand...

Yes indeed. Thank you - that looks very interesting.

Chesham said:
If you are using BF SP, then you will not know before the race is run if your predicted odds are value, as the SP is not declared until after the race has been run
Click to expand...

Very good point. But the Betfair SP is the only odds data I have - its included in the proform.com data. BTW, I am assuming I can get significantly better odds on the betfair exchange. If you have any suggestions as to how I can more rigorously test my profitability than I am all ears.
 
You must log in or register to reply here.
Back
Top