Race Number Analysis

Early races...

I mentioned recently that we seem to do better in the earlier races every day.

I thought that it would be a nightmare to do the analysis but,

Actually, it was easier than I thought.

I wrote a program to add an extra couple of columns to the results table and then ran a query on it.

However, when it came to doing the graphs...

That's another matter!

An explanation

I have done a graph for each Number of Meetings per Day.

I have ignored the one day in early 2013 when there was only one meeting - less work for me and, after all, how often does that happen.

And, if there is only one meeting it's probably because of extraordinary weather conditions and you'll probably have more on your mind (flooding, snow and hurricane force winds, for example) than risking your children's inheritance on the horses.

Each chart shows 2 data series. The jiggly red line shows the ROI and the jiggly Blue line shows the strike rate.

There are also sloping lines which are a sort of mathematically smoothed out jiggly lines and,

There are flat lines which show the average (or Mean.)

My strategy is to assume a cut off point where the sloping line hits the Mean line.

In most cases it's the same for ROI and Strike Rate but, I shall be using the Strike Rate Line (in Blue) to base my decisions on as it's slightly more mathematically sound to do it that way.

However, where the crossover point is ambiguous, I shall let the ROI influence me.

I shall produce a PDF for you to print out and stick on the wall (or put in your Racing Notes Folder) but I can't really add them to the daily ratings as, if I do it for one format, then I'll have to do it for all of them (23 different formats as of today.)

You should bear in mind that I am only using the Top Rated data for this analysis.

If you want to see the graphs in more detail, there are stretched copies of them on the Graphs tab of the spreadsheet.

The spreadsheet may be found at... http://focusratings.com/data/meetings.xls

It contains just over two years of data.

So, here goes.

2 meetings

The cutoff point here seems pretty clear.

For days when there are only 2 meetings, the first 8 races are the best from a strike rate point of view.

3 meetings

Once again, a clear cut off point.

For days when there are only 3 meetings, the first 12 races are the best from a strike rate point of view.

4 meetings

Although there is a divergence between the Strike Rate and the ROI, I am only interested in the Strike Rate.

For days when there are 4 meetings, the first 15 races are the best from a strike rate point of view.

5 meetings

As the number of meetings increase the lines on the graph get flatter.

For days when there are 5 meetings, the first 19 races are the best opportunities.

However, there is very little in it.

6 meetings

For days when there are 6 meetings, the first 24 races are the best opportunities.

7 meetings

For days when there are 7 meetings, the trend line exactly matches the mean line and, thus, all races present equal opportunities.

8 meetings

For days when there are 8 meetings, the first 24 races are the best opportunities but there is next to nothing in it.

9 meetings

For days when there are 9 meetings, the first 33 races are the best opportunities.

10 meetings

For days when there are 10 meetings, we see something strange happening and the last 28 races are the best opportunities.

However, in the scheme of things, the trend line so nearly matches the mean line that there is nothing in it.

11 meetings

For days when there are 11 meetings, the first 34 races are the best opportunities.

With this many races it would be foolish to ignore the later races, though.

12 meetings

For days when there are 12 meetings, the trend line exactly matches the mean line..

Rules...

2 meetings - the first 8 races are the best from a Strike Rate point of view.

3 meetings - the first 12 races are the best from a Strike Rate point of view.

4 meetings - the first 15 races are the best from a Strike Rate point of view.

5 meetings - the first 19 races are the best from a Strike Rate point of view.

6 meetings - the first 24 races are the best from a Strike Rate point of view.

7 meetings and more - in the scheme of things, where there are 7 or more meetings on the same day, there is no real determining factor.

Conclusion

So, I need to come back and re-visit this but, the observation that we do better in the early races is proven but...

Only for days when there are six or fewer meetings.

The data is available for all to play with.

The spreadsheet may be found at... http://focusratings.com/data/meetings.xls

Now, this is just my interpretation of the data and you may well come up with ideas of your own.

I look forward to your ideas as to how to better analyse the data.

Logically, from my viewpoint, the earlier races should be more accurate as I generate the forecast prices at 9am Uk time and the forecast price does have an effect on the ratings.

Also, there is more opportunities for changes in going and non-runners as the day goes on.

Both of these elements also affect the ratings.

It has also been explained to me that the first races of the day are generally non-handicaps and these are far easier for me to rate.

As always...

My kindest regards