• 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

Chi, Exp and other statistical methods for evaluating systems

Thank you runnerjp runnerjp,
So Chi is king along with A/E? I have been thinking along the same lines but have tried to use Exp rather than instances (if I get your meaning). I struggle to get to a good threshold using Exp though, even though there may have been plenty of runs and other excellent stats. Can I ignore it after a reasonable incubation time?
 
Couple more checks I use also!

What is the strike rate? (How often do you find the winner?)
What is your Return on Investment (How much are you staking or placing on your bets relative to the amount of monies which are being returned in to your betting account?)
How many bets a day do you need to place?
Does the method work at level stakes, if not, the chances are it will not be a winner in the long run whatever staking plan you adopt.
How big a “bank” do you need? REMEMBER only bet what you can afford to lose.  What is the longest losing run?
 
I consider A/E the most important, but for my methods I am finding that strike rate is very important too so these two things are what I consider most, A/E and strike rate. I look at LSP but for me it’s not that important as it can be affected by difference in price ranges, so you can have good A/E but poor LSP, it will be because the system works better with shorter prices so I would look to introduce a price filter in that case.
 
Frontrunner said:
I consider A/E the most important, but for my methods I am finding that strike rate is very important too so these two things are what I consider most, A/E and strike rate. I look at LSP but for me it’s not that important as it can be affected by difference in price ranges, so you can have good A/E but poor LSP, it will be because the system works better with shorter prices so I would look to introduce a price filter in that case.
Click to expand...
Chi Score uses A/E as part of its formula.

You can have a decent A/E but no way of knowing if that is down to luck. Chi helps to determine the likelihood of your system being down to luck or not.
 
runnerjp said:
Couple more checks I use also!

What is the strike rate? (How often do you find the winner?)
What is your Return on Investment (How much are you staking or placing on your bets relative to the amount of monies which are being returned in to your betting account?)
How many bets a day do you need to place?
Does the method work at level stakes, if not, the chances are it will not be a winner in the long run whatever staking plan you adopt.
How big a “bank” do you need? REMEMBER only bet what you can afford to lose.  What is the longest losing run?
Click to expand...
Thanks runnerjp runnerjp. Yes, strike rate is mega important to me. I am one of those people who doesn't cope well with long losing runs either so although I don't fiddle around too much with my systems for fear of backfitting, forewarned is forearmed. ROI too, I ditch systems which don't have a good ROI and for me 'good' is above 100%.
I'm interested in your point "How many bets a day do you need to place?" - I tend to place far too many. I make a rod for my own back because when I get a lot of qualifiers, I don't then have time to qualify the qualifiers so to speak. Which brings me back to my original reason for posting. I want to prune my systems, just keeping the best but I want to explore how I might use the various statistics/metrics to do it. I always use level stakes to 1% of my bank.
I'm in despair at the moment though because only about 2 of my 30 systems are performing well according to the stats. It may be simply a difficult time of year so I am just sitting it out right now.
 
Chesham said:
Chi Score uses A/E as part of its formula.

You can have a decent A/E but no way of knowing if that is down to luck. Chi helps to determine the likelihood of your system being down to luck or not.
Click to expand...
True Chesham Chesham, but A/E is very useful to me when I am researching particular ideas. Not sure that I would disregard it at that stage if Chi was low.
 
Chesham said:
Your choice

Hope things work out for you with your systems
Click to expand...
Chesham Chesham, I can't thank you and all involved enough for the link to the 'hrb for novices thread'. Although I have been using it for a year, the scope of it is so huge that it is impossible to do more than scratch the surface in such a short time. I didn't really start using it in anger until June when racing restarted.
A lot to get through but really worth spending the time
 
Chesham said:
Chi Score uses A/E as part of its formula.

You can have a decent A/E but no way of knowing if that is down to luck. Chi helps to determine the likelihood of your system being down to luck or not.
Click to expand...
Good point Chesham Chesham, and you are correct. I do understand Chi score and the formula behind it, but I don’t know it instinctively in same way as I do with A/E, so I wasn’t suggesting that it is not important. It is very important for any system for the reasons you mentioned.

I don’t use the system builder so much for creating systems though, I use it for considering the effect of different factors for a race I am looking it, so that’s where I was coming from when I said A/E is most important for me. One example would be the draw bias, let’s look at Chester 5f races, low draw has A/E of above 1 but Chi score is rightly very low as it would not be profitable to blindly back every low drawn runner at that track. But for me A/E is good enough in this case as I am using the draw as only one part of my selection process. I hope that clarifies my position.
 
If you have used the Ratings Tool to develop your own Ratings they can be useful to see how they perform with your systems

This one has been going for some time now

4EA22732-379E-4B6D-AA50-E2EAE66057F2.jpeg
These ratings were posted by me for awhile on the OS Part of the Forum

If I apply them to the system above

234C743A-EBEB-4EC3-9737-086A8463BD82.jpeg
The number of bets reduce and the P/L improves
566/270. P/L 211 .81
 
Frontrunner said:
when I said A/E is most important for me. One example would be the draw bias, let’s look at Chester 5f races, low draw has A/E of above 1 but Chi score is rightly very low as it would not be profitable to blindly back every low drawn runner at that track.
Click to expand...
A low Chi does not mean a system would not be profitable. Chi is simply an indication of whether or not the results to date are just down to luck, good or bad, and can indicate, if high, that the current performance is more likely to continue.
 
ArkRoyal said:
A low Chi does not mean a system would not be profitable. Chi is simply an indication of whether or not the results to date are just down to luck, good or bad, and can indicate, if high, that the current performance is more likely to continue.
Click to expand...
I stand corrected.

A low chi would then indicate that a good A/E is down to luck and less likely to continue, which in turn would mean that a system is less likely to be profitable, that is what I was trying to get at. Chi uses A/E as part of its formula, A/E in turn uses SP as part of its formula, so it all comes down to profitability in the end. In theory, even a system with high Chi can become worthless if everyone including the bookies become aware of it as they would simply slash the prices.
 
Frontrunner said:
I stand corrected.

A low chi would then indicate that a good A/E is down to luck and less likely to continue, which in turn would mean that a system is less likely to be profitable, that is what I was trying to get at. Chi uses A/E as part of its formula, A/E in turn uses SP as part of its formula, so it all comes down to profitability in the end. In theory, even a system with high Chi can become worthless if everyone including the bookies become aware of it as they would simply slash the prices.
Click to expand...
If,you read the article it later goes on to explain after finding a system that has a good chi score how you can test the System to see if it is failing or just a blip . It helps to know when to abandon a system
 
Chesham said:
If,you read the article it later goes on to explain after finding a system that has a good chi score how you can test the System to see if it is failing or just a blip . It helps to know when to abandon a system
Click to expand...

Which article is that Chesham Chesham ? Btw, great numbers for your system above, well done.
 
You must log in or register to reply here.
Back
Top