The consistent teams tend to be the best teams in the league.  Therefore, I would consider the traditional top 6 in the premier league to be consistent.  This would include Man City, Liverpool, Chelsea, Man United, Tottenham and Arsenal.  

At the end of a football season, the top 6 teams tend to end up in or close to the top 6 positions in the Premier League.  This suggests that these teams are consistent.

Although these teams can have good and bad runs, they are generally consistent.  Nevertheless, I would still be aware when a team is going through a difficult time.  For example, I wouldn’t trust Man United to perform as consistently as Chelsea.  

I recommend following consistent teams for several reasons.  

The main reason for following consistent teams is that I can put a lot of year’s data into the Poisson calculator.  

For some fixtures, you may need to use a small amount of data.  This is usually when the opposition to the consistent team is an inconsistent team.  

In addition, if necessary, you can put less data into the calculator and still have reasonable trust in the result.

If one of the teams in the match is consistent, you only need to adjust the amount of data for your analysis in accordance with the inconsistent team.

With inconsistent teams (or teams that have had an unusually bad season), you have to figure out which seasons are most representative of their current form.

When a team has had an unusually poor previous season, it is difficult to obtain clear information, concerning whether the team is back to form.

In the 2021/22 season, Man United had a poor seasons, relative to past performances.   If you put 1 year’s data into the calculator, you will get very different results, compared to a long term calculation.

I would avoid trading or betting on such clubs until they show whether they are back to their better form.  

Even when you see good results, you need to do further analysis.  

In my lesson, How to Objectively Analyse a Football Team’s Last 6 Matches, I explained how to determine whether a team is playing well or whether the team has just been lucky.  This is the type of analysis that you will need to do.

When a new manager moves to a struggling team, he may get better results and the form of the team might also improve.  However, the improvement is often short-term.  An example of this was when Ole Gunnar Solskjær became manager of Man United.  The team started off with some good results, but this was short-lived.

Generally, it is best to avoid getting involved with such teams for, at least the first half of the season.  

That said, there may be an angle if the 6 match analysis suggests that an improvement in results is due to luck and not an improvement in form.  In this case, I might try to find an opportunity to bet against the team and use a 1 year analysis as my guide to finding value odds.

As an example of a team that has performed well last season, and poorly in previous season, let’s look at West Ham.

West Ham’s last 10 year’s finishing positions in the Premier League include:

7th in 2021/22

6th in 2020/21

16th in 2019/20

10th in 2018/19

13th in 2017/18

11th in 2016/2017

7th in 2015/2016

12th in 2014/2015

13th in 2013/2014

10th in 2012/2013

In terms of market value, West Ham are currently ranked 10th in the Premier League.  As such, West Ham have over-achieved over the last 2 seasons.  

However, apart from 2019/20, West Ham have finished the league, within 4 positions of its market value position.

So, how many year’s data should we use to analyse West Ham?  

If you think that West Ham are still over-achieving, you might use 2 years of data.  

However, if you think that improved performance over the past 2 years is about as far as West Ham can go, you might use a lot of data.

The worst amount of data, that you can use is 3 or 4 years.  This is because the 16th place finish in 2019/20 doesn’t fit with the rest of their finishes.

Therefore, if West Ham starts the season performing well, it may be best to use 2 years of data.  Otherwise, it may be best would use as much data as possible.  

The idea of using a lot of data is to dilute the effects of the 2019/20 season when West Ham came 16th.

Although the Poisson model is powerful, we still need to do some analysis.  This isn’t an exact science.  However, if you are analysing data in more detail than the crowd, you have an advantage.