Ten Year Trends: Sorting Out the Wheat from the Chaff

A week’s holiday gives you a chance to refresh and recharge the batteries. In my case, it also gives me the chance to relax on the beach, drink cocktails and to evaluate betting methodologies.

Just before Christmas, I had a week in Mexico. Not only did I have a great time but away from the daily grind of horse racing and betting I had an opportunity to think. That thinking time isn’t normally available when it’s ‘wall to wall’, seven-day racing on a screen in front of you.

Last year I wrote an article on the “less is more approach”. In short, it highlighted how studies had shown that more filters or variables that you use, doesn’t make you a more successful punter. Some people say, of course, that the more information that you have the better it is but more information isn’t necessarily good for you. Too much of the wrong type of information can indeed have a detrimental effect on your betting bank that’s for sure.

The key is to find the right information or sort out the “wheat from the chaff” if you like. Sort out the useful stuff from the downright unhelpful. If you can get rid of the useless information not only does it clear your thought process but it can also prevent you from heading in the wrong direction.

This month’s article is this first in a three-part series on using ten-year trends in your betting. What you should be looking for and what you should be avoiding and most importantly how the “less is more approach” can be used to good advantage when looking at trends. But let’s begin by taking a very quick introduction to the use of statistics in general and in regard to betting on horses.

“There are three kinds of lies: lies, damned lies, and statistics.

It was Mark Twain, writing his autobiography, who first attributed to British Statesmen Benjamin Disraeli the quote “There are three kinds of lies: lies, damned lies, and statistics.”

Why did Twain use these words?

It’s because statistics can be manipulated to “prove” anything someone wants them to “prove” if the person being subject to the statistic doesn’t know how the stat was created, what sample size was used…How can they see flaws in the figures?

Here’s a non-racing example of what I mean. If someone tells you that 80% of people prefer dark chocolate to white chocolate. As someone who prefers the latter. How do you feel about that? Do you feel the odd one out because you prefer white chocolate? What if the people who answered the survey were just one person’s immediate family! How meaningful would that figure be?

It’s all about the sample size, if it’s too small it will skew the results. Likewise, if data is left out because it doesn’t conform to the desired result it will also skew the results. It’s all too easy to mislead yourself or of course others even when using “true” statistical information.

If you have a scale from 1 to 100, a change from 44% to 46% doesn’t look a big one. However, if you have a scale from 40 to 50. The same percentage change would be much more significant.

Looking at a racing example: The Champion Bumper at Cheltenham. Looking at the last 11 renewals of the race:

“91% of Winners Were Aged 5 or 6”

That sounds a cracking stat/trend, whatever you want to call it until you realise that 88% of the total runners were aged 5 or 6.

You will often see or hear a pundit say Jockey A has a 50%-win strike rate when teaming up with Trainer B at a particular course. If they have combined just four times how reliable is the statistic? Not very reliable.
Firstly, the sample size is too small and more significantly the winners might have all come with the same horse.

The value of any statistic lies purely in the sample size. I mentioned in last months article on backing favourites that “the proportion of races won by the favourite is running at a healthy 33%”. That one-third figure has been about the same since betting on horses began and will never really change. What It doesn’t mean is that today’s six race card at Fakenham will see two favourites win. No favourites may win that day or all six races could be won by the market leader.

Another statistic that’s proven is that a horse’s chance of winning a race are reflected in its odds. If you bet on every even money shot you will lose but you will lose less than if you backed every horse at 3/1 which in turn would lose you less money than if you backed every horse at 5/1.

Those are stats that will always be with us.

It was never my intention to delve too deeply into use of statistics here. If you want to learn more about stats, I recommend this site: http://onlinestatbook.com/ and this old but still useful book by Darrell Huff, “How To Lie With Statistics” which you can download as a PDF.

https://www.horace.org/blog/wp-content/uploads/2012/05/How-to-Lie-With-Statistics-1954-Huff.pdf

The Ten-Year Trends

I have a confession to make I am big fan of the ten-year trends analysis when it comes to big races. I like the idea of predicting future performance based on historical data. Big race trends have found me some great priced winners in the past but they haven’t been the golden key to profitable betting.

Now you will see plenty of good 10-year-trends analysis out there. Simon Rowlands in his insightful Rowley File on the Timeform website, will look at some of the key ones when he’s analysing a big race. You will find big race trends from Andy Newton on the www.geegeez.co.uk website. Kevin Morley looks a big race trends in the Racing Post. And of course, there’s my colleague Dr Nick Hardman who uses such trends for his excellent big race profiles.

There are many systems based on following the trends of ‘key races'. From personal experience, I can say that following big race trends does work for some but not all races. The very fact that it does seem to work on certain races means it’s worthy of further research.

Many big races have been run for 20-years or more, so it is possible to look back at past winners to get an idea of the type of horse capable of winning. They are normally run at the same point of the year each season. Trainers will prepare horses in the same way as previous years and indeed their prep races could be the same as for previous winners.

As Nick Hardman does, you can build up a profile of each race to find horses running with a similar profile that is worth further study. At the very least this trends approach can help reduce a 30-runner field down to manageable shortlist of contenders. It’s that list of contenders that you can further form study on.

In Part two. I will give some examples of races where the use of the trends can be a powerful tool that can be used to profitable effect and some races where the trends can lead you in the opposite direction.

Until next month
John Burke

Read part 2 here https://www.oncourseprofits.com/horse-racing-trends/