Betting stats part 3

In our previous articles we have looked at types of betting stats, general terminology and how to define a population and a sample. You find those articles here:

• Introduction to betting stats
• Betting stats part 2

The scope of this betting stats article is to explain what frequency distributions are and how you can use them in order to quantify betting related data. Let’s get to it.

What is a frequency distribution?

A frequency distribution in betting stats terms is a grouping of betting related data into categories or intervals showing the number of observations in each mutually exclusive category. Absolute frequencies are the number of observations in each interval. Relative frequency measures the proportion of each interval or category. This is done simply by dividing the absolute frequency of the interval by the total number of observations.

Frequency distribution example

Say you have a sample consisting of 100 football games of teams where you consider the teams in every fixture to be of the same quality more or less. Let’s assume you find the following distribution:

The absolute frequency for interval Home would be 37, for Draw 34 and 29 for Away. The relative frequency in percent would be 37% for Home, 34% for the Draw and 29% for the Away win.

Application of frequency distribution

In general one can say that this summarization of data is done to simplify the data collection process as well as make the end product more easily measurable and in order to see outcomes or trends more clearly. For instance; if you are looking to break down soccer stats to percentages, and then further to odds, this is the way to go about.

You can apply this way to pretty much any of the available soccer stats or other sports stats or financial stats as well as a wide range of other statistics available. Only your imagination and access to data limits the possibilities. If you spend some time thinking about how to define your data set, setting correct categories and intervals you can really dig deep into the heart of the data available. If you can define groups of data, say soccer teams for instance that are likely to perform similar results over time versus other defined groups then you can be on to something interesting, hint hint! This all sounds very easy when you read it, but try going at it in real life. It will require a lot of time and energy.

Also, we could for instance use the data example above to isolate the cumulative relative frequency. If you wonder what that means; it is simply the running total of the relative frequencies. Say we are figuring out a HU bet scenario and wonder how often a team would meet our criteria. Mathematically this would simply be found this way: (37%+34%)/100% = 71% chance for home win or draw.  37% and 34% would point to the relative frequencies of home wins and draws. We divide the sum of those two relative frequencies on the total, and voila.

Frequency distribution construction

We have worked through it, and many of you will already have grasped it. However, it doesn’t hurt to drill a good concept further in with a little bit of repetition and figurative speech.

First you got to set up the intervals or categories. If you work with percentage based intervals (could be you are analyzing stock performance for instance) you can make things a little easier for yourself spending some time looking at your data and set the lowest interval a little lower than the lowest you find, and the highest a little above the highest you find.

Second you must make sure the intervals or categories are mutually exclusive. This doesn’t mean much else than that each data point must fit one, and only one interval. The endpoints of your intervals should not overlap each other. To get this right you should just make sure the upper limits are considered to be “up to” but not including an upper limit that is equal to the lower limit on the next interval.

Final thoughts

We hope you enjoyed this article on betting stats. Next up is a presentation on different ways of stating averages in statistical theory. It helps you understand your data better and it will enable you to get to more correct assumptions relating to your data set. In other words; read on!