Place Betting Strategy – Improvements

Some logical improvements

Something was worrying me a bit about my new Place Betting Strategy (of which I'm very excited, I have to say)...

So, yesterday, I decided to test the strategy on all the data that I've got.

This is from the 8th of July 2013 to the 9th of June 2015.

In other words, almost 2 years of data.

And I realised that I'd been making a really big mistake!

You see, from my view of the numbers, we are looking at two different types of race;

We have the bigger races (16 runners and upwards) for which you can blindly back the top three rated horse which is top on yield (of the top three rated horses) and make a profit.

However, for the smaller races (8 to 15 runners) we need to add more filters.

We could just ignore the smaller races but they make up the majority of races and there are ways to make money from them.

Thus, I've decided to come up with two separate strategies; one for the smaller races and another for the larger ones.

Don't worry, this won't mean any more work for you (although it does mean a couple of extra days work for me); I shall automate the selections so that you don't have to do anything and...

I'll share the raw data in case you want to come up with a better way of doing things.

The Spreadsheets

I ran a query on the database to select every top on yield horse (from the top three rated horses) in every non-handicap race where there were 16 or more runners.

This data is available at PBS01.csv

The final (and most useful) spreadsheet is available in the following formats...

Excel 2007 format - PBS01.xlsx

Open Office/Libre Office format - PBS01.ods

Excel 97 format (looks OK to me but it did report some missing rows) - PBS01.xls

This is what I shall be working with.

You'll see that, at the bottom of the spreadsheet, there are a number of tabs.

The first tab (Data) contains the raw Data.

I split the runners into runner bands.

I did some analysis on this and have displayed the results on the Basic Analysis tab.

So far, the rules of the strategy are...

1). The horse which is top on Yield (Yld) from the top three rated horses in

2). Non-handicap races where

3) There are 16 or more runners.

As you can see, for the 23 month period, there were 684 selections.

That equates to 29.74 selections a month.

We had a strike rate of 40.35%

The ROI was 147.97% (after deducting 5% Betfair commission.)

Not a bad start!

R4_status Analysis

I then decided to look at how the results varied according to R4 Status.

The results are displayed on the R4 Status Analysis Tab.

A quick look at this shows that we'd do well to discard 2nd rated horses in races of between 20 to 24 (inclusive) runners and to discard top rated horses in races of 25 runners and more.

However, I'm going to look at some other factors first.

After all, this is a pretty simple strategy so far and I'd like to keep it that way.

In any case, there's no reason why I shouldn't come up with a simple strategy and let you add additional filters to it manually in order to improve the strike rate and profitability.

RC0 Analysis

I then took a look at the effect of RC0 on the results.

To my mind, this is a key filter and so I've decided to remove any RC0=3 races from races with 20 or more runners.

I've cut some new data to show this and this may been seen on the Data01 tab.

I did some analysis on this and displayed the results on the Data01 Analysis tab.

So, the new rules of the strategy are...

1). The horse which is top on Yield (Yld) from the top three rated horses in

2). Non-handicap races where

3). There are 16 or more runners.

4). For races with 20 or more runners, races where the RC0 is 3 are not allowed.

As you can see, for the 23 month period, there are now 603 selections (down from 684.)

That equates to 26.22 selections a month.

We had a strike rate of 43.12% (up from 40.35%.)

The ROI was 158.32% (after deducting 5% Betfair commission), up from 147.97%.

Getting better!

A pause for a deep breath...

Now, so far I like these numbers.

They're based upon the best part of the last two years data (which is a significant amount of data) and, whilst the strategy won't give me the 3 races a day that I was looking for, don't forget that the majority of all races are between 8 and 15 runners (a strategy for which will follow.)

Last Time Out.

I did take a quick visual check on the LTO number (days since last run) and I colour coded the paying results (isn't conditional formatting great?)

I'm not going to use this as a filter as there are plenty of high paying winners/placers who haven't run for over a year; I guess that because these horses haven't run for a while, they are underbet, to our advantage.

Country Analysis

Just to keep Ang (one of our members) happy, I did do some analysis of the effect of race country on the results.

I've detailed this on the Country Analysis tab.

To be honest, this hasn't shown anything that I didn't expect; the 3 UK 20 to 24 runner races that aren't profitable make up only 0.5% of all races - you could discard those if you felt like it but I'm keeping my strategy simple.

In fact, when I publish the strategy, I'll also publish a cheat sheet of other factors that you may wish to use in order to improve the strike rate and profitability.

Conclusion

And there, Ladies and Gentlemen, I'm calling it a day, for today.

I'm happy with a simple system which has the following rules...

1). The horse which is top on Yield (Yld) from the top three rated horses in

2). Non-handicap races where

3). There are 16 or more runners.

4). For races with 20 or more runners, races where the RC0 is 3 are not allowed.

Now, tomorrow I shall be banging my head against my desk working out what to do with the majority of races - those with 8 to 15 runners.

I'm expecting to be able to get another 60 selections a month from those races (albeit with more complex filters) and that should get me the 3 races and 5 winning days each week that I was after.

In the meantime, I'm counting the minutes (and seconds) until Wine O'Clock but, as I've been up since 1am this morning, I rather suspect that it may be arriving early!

Our first Place Betting Strategy selection today (Maybe Tomorrow in the 14:50 at Salisbury) was a non-runner.

From the All Horses PDF (option 3 on the ratings email)... http://focusratings.com/data/focusratings-complete-2015-06-14-602295.pdf you can see that Althania would be the new selection, however her Forecast SP would make her a non-contender.

Our second Place Betting Strategy selection today (She is no Lady in the 15:05 at Doncaster) came 2nd in a 17 runner race so, we're one for one so far (which is far better than being zero for one!)

Betfair paid 2.46 for the place so, so far today, we're running at a 100% strike rate and have had 2.337 points returned for 1 point staked.

Have a great afternoon!

As always...

My kindest regards