Hello…

I actually run another ratings agency, Focus Ratings (which you may have heard of.)

However, I never stop looking at ways of improving Focus Ratings and realised that I need another layer of ratings to sit on top of the current Focus Ratings strategy (called R4) to create a new and improved strategy (to be called R5.)

So, to create this extra layer I decided to start again from scratch loosely inspired by Clive Holt’s Fine Form strategy.

What I didn’t expect was to find that the ratings I produced would be “out of the box” profitable!

Using 2016 as a data set, the top rated horse has a strike rate of 25.76% and returned 13.77% level stakes profit (to ISP) on the stakes.

I am in the process of working out how to add the new ratings to Focus Ratings (to improve the accuracy of that service) but…

With results this good I decided to publish Advance Ratings as a stand alone service (whilst also adding it to Focus Ratings.)

Starting with FineForm…

Clive Holt’s Fine Form ratings system is very simple. The rules of it are below…

Last 2 Races in the Current Season
1st Place 5 points
2nd Place 3 points
3rd Place 2 points
4th Place 1 point
Course and Distance win 3 points
Distance win 3 points
Course win 1 point

Now, I had some questions about some of those numbers.

They just look to me (and my devious mathematical mind) just a bit too rounded.

So, I decided to test all the numbers to see if I could improve on them.

I did this by firing up Wagner; Wagner is a 12 year old Intel server with a raid 10 array of fast SCSI hard drives, twin hotswap power supplies and dual Xeon processors. She would have cost about £10,000 when she was new; I paid about £300 for her – second hand servers have pretty much no resale value.

I don’t switch Wagner on very often as, when I do, the lights dim throughout Western Brittany (which is where I live) and she does make a bit of noise.

She does warm the office up nicely though, on a cold winter day.

Wagner is specifically configured to do number crunching. She has no internet access, no “Windows” type screen (only what some of you might remember as something like a Dos prompt) and she only runs 2 programs (both of which I wrote myself.) Sherlock does high speed number crunching on large data sets and Holmes turns the output into something a human might be able to understand.

I fed Wagner with 4 years of data and started off by looking at the number of races to use.

Number of Races

Now, Clive Holt published FineForm in 1988 and, presumably, it took some time to research and write. A lot has changed in the last 30 years and I’m sure that most people now, when looking at a horse’s form, look at more than the last 2 races.

I kicked off the analysis and went out for a long walk – Wagner is irritatingly loud.

When I got back the results were in; these days you need to analyse a lot more than two races in order to get accurate results.

Obviously I can’t tell you the exact number as that would be giving away all my hard work but I can tell you that Advance Ratings looks at far more than the last 2 races.

So, on to the points awarded – the interesting bit.

Number of Points

For the next ten days I ran Wagner for about 10 hours a day to work out the right number of points awarded for a win or a place.

Once again, I can’t tell you the exact numbers but, if we start by giving a win 5 points then you’d get something like the following numbers…

Win – 5.0
2nd – 4.6
3rd – 3.8
4th – 2.9

Obviously those are made up numbers but you’ll see that there’s far less distance between the numbers (especially between 1st and 2nd.)

One thing that I did add was that I only gave points where the bookmakers payout for places.

So, up to 4 runners, only the winner gets awarded points.

5 to 7 runners and the winner and the horse that came second get awarded points.

8 to 15 runners and the winner and the horses that came second and third get awarded points.

16 or more runners and the first four horses get awarded points.

I moved on to looking at the points awards for Course winners, Distance winners, Course and Distance Winners and Beaten Favourites.

The actual numbers aren’t so important but I can say that being a Course winner is far less important than being a Distance winner.

I put together a script to analysis what I had got so far and ended up with a reasonable ratings system. Not good enough to publish but maybe something that I could add to Focus Ratings to improve the strike rate of that service.

Then I had my Eureka moment; the key that really makes the difference. I decided to add some value…

Adding Value

The one key thing that I believe that Clive Holt missed was the quality of the race.

Imagine the following scenario…

Dennis the Donkey has run 3 races in all his racing life. The first one was a class 7 and the next two were class 6 races. He won all three races.

Speedy Simon has also only run 3 races and came second, fourth and third. His race were all class 3.

So, Dennis the Donkey gets awarded loads more points than Speedy Simon but, in the real world, if the two horses met in a class 4 race, Speedy Simon might not win that race but he’d certainly beat Dennis the Donkey.

To resolve this issue I initially thought about adjusting the number of points awarded according to the class of the race that those points relate to.

However, that gave me a problem with Irish races (which don’t use the class system.)

Then I realised that the Class system is based upon the minimum prize awarded to the winner of the race.

Bingo! I just multiplied the number of points awarded by the prize money of the race to get a better rating.

Imagine Black Beauty has run 4 races and came 5th, 5th, 2nd and 1st.

Those races had prize money of £4000, £5000, £3000 and £6000.

In the race where Black Beauty came 2nd I multiply the points awarded to him by 3000; in the race where Black Beauty came 1st I multiply the points awarded to him by 6000. I then add those two numbers together and divide the result by four.

This gives me a number that I call Value (or horse_value to be more precise.)

To get a rating I then divide the horse_value by the prize money for today’s race.

That gives me a number that I call horse_rating and that’s basically it.

Almost…

ADR

Just by looking at the horse rating numbers I could see that…

1). I’d never be able to find a way of using them for Focus Ratings and…

2). They’d confuse the hell out me.

So, I decided to make the numbers more relevant by dividing horse_rating by the average horse_rating for the rated horses in the race.

That gives me a number that I call ADR and that is a far more relevant and useful number.

Results

I ran the new formula through Wagner once more and was truly shocked by the results.

So shocked, in fact, that I spent a week looking for bugs because I knew that something had to be wrong.

I even spent 2 days manually calculating the ratings using pen and paper (and a pocket calculator.)

A waste of time really as the numbers weren’t wrong.

You see…

What I’d discovered that, had we had the ratings in 2016 we would have had a strike rate (for our top rated horse) of 25.76% (which isn’t bad) and…

Wait for it…

We would have made a profit (to level stakes to ISP) of 13.77% on our stakes!

And that’s amazing – no ratings service claims to be blindly profitable for its top rated horse (Timeform did use to, I think, but they sometimes have 2 or 3 top rated horses in the same race; it’s easy to have a top rated winner if you have 3 top rated horses in a 8 horse race.)

Conclusion

Fundamentally I have come up with a very simple ratings system.

So simple that I can actually rate an 8 horse race just using a pen and paper (and a pocket calculator) as long as I’m sat in front of the Racing Post or the Sporting Life (or similar website) in about 5 minutes.

In a way, it is a modern version (or interpretation of) Clive Holt’s FineForm.

It is blindly profitable to level stakes although I do have to say that you’d have to be mad to blindly back the top rated horse in every race.

I hope that you’ve found this helpful and informative. Although I can’t give away the specific numbers for the points awarded, I do like to be as honest and open as I can.

And, after all, there’s nothing to stop you from using my steps to come up with a similar ratings system of your own.

Or, you could just save yourself a lot of effort, masses of head scratching and hours and hours sat in front of 6 computer screens and…

Subscribe to Advance Ratings and just use my ratings.

Have a great day’s racing!

As always…

My kindest regards

keith-eckstein1

For a monthly payment of just £27.00 (plus VAT) you can use the service to narrow the field so that you can concentrate on the real contenders….

ClickBank®

For a quarterly payment of just £67.00 (plus VAT) you can use the service to narrow the field so that you can concentrate on the real contenders….

ClickBank®

ClickBank is the retailer of products on this site. CLICKBANK® is a registered trademark of Click Sales Inc., a Delaware corporation located at 1444 S. Entertainment Ave., Suite 410 Boise, ID 83709, USA and used by permission.

ClickBank’s role as retailer does not constitute an endorsement, approval or review of these products or any claim, statement or opinion used in promotion of these products.

 

Leave a Reply