Another post today looking at some useful football stats and how different leagues can have different characteristics. Today I’m looking at draws which are an important part of football trading with many football trading strategies revolving around them.
As usual looking at statistics over the course of a few games or even a couple of seasons inst always the best idea given the fluctuations involved with football. Personally I like to look for characteristics which help identify good value for my trading systems over the longer term.
In the post below I am using the last 10 seasons worth of data from the main European leagues for my research (2008/2009 to 2017/2018). I am also using abbreviated league names, below is the key table for these:
The Best Leagues for Draws
Here I am looking for the average number of draws per game for each of the main European leagues over the past 10 seasons:
As can be seen above the main European leagues all have similar draw percentages ranging from 23% up to 32% at each extreme. We can also see the Premier league is around 25% which is a commonly cited statistic whenever the premier league is concerned.
How Consistent are Draws Over the Years
Whilst we are looking at long term averages we need to be aware we are also averaging out any short term extreme values. This means whilst, for example the Premier league might average a 25% draw rate over the long term this might vary wildly on a year-by-year basis. If this does turn out to be the case it might be much harder to predict a draw in such circumstances?
The table above shows the draw percentage for each league on a year-by-year basis and thankfully they are, for the most part fairly consistent. To help highlight this I have produced a graph below showing the year-by-year draw percentage for each league over the 10 seasons:
The graph above is a little hard to read since most leagues have a very similar draw percentage over this time frame. Despite been hard to read this graph can still be useful as it highlights how looking at how several leagues across multiple seasons we still find the draw rate is not only similar but also consistent year after year across all leagues!
The table above shows the highest and lowest yearly draw percentages for each league over the 10 seasons along with the percentage differences between the two.
Again, no wild swings here and interestingly all leagues seem to vary between having a “low” figure around 20ish and a high figure of around 30ish. Given our all games and all leagues average is around 25% this is proving to be a fairly steady metric overall. A steady metric in this case however might suggest the market will be more likely to price such an event (the draw) correctly so potentially leave less value for the trader?
The Best Leagues for 0-0 Draws
A big part of conventional football trading is about waiting for an event to happen (usually a goal) and then trading out at a profit or a loss, as such many traders seek games which are likely to see goals. Much less common is the practice of seeking games with a lower probability of goals and with lower final scores overall. As such been able to identify games more likely to have zero goals (I.e. a nil-nil) can help form the basis of many a profitable trading strategy.
Often some of the highest (yet still realistic) correct score odds for any game are that of a nil-nil and conventional football wisdom would have us believe a 0-0 draw happens roughly 10% of the time. Below I’m going to look into this further using the same data as above and see what insights (if any) show through.
The table above show each league’s long term nil-nil percentage for the 10 seasons, interestingly an average of 10% which was mentioned earlier seems to be more at the top end of the range than the norm!
Also it is interesting to note that some of the smaller leagues including the Italian and French 2nd divisions head the table with the highest number of nil-nils. Bigger leagues including the English Premier League, English Championship and the La Liga are more towards middle and bottom meaning they see less 0-0 games on average. This is interesting as I’ve written before about looking for lower scoring games and that the value (and frequency) of these games seems always to be skewed towards the lower leagues, this is again showing to be true here!
In the table above we see the highest and lowest 0-0 percentages for each league over the 10 seasons.
Interestingly we again see little difference in the nil-nil maximum and minimum values although due the relatively low nil-nil rate to begin with this represents a fairly large percentage change. This is interesting as it potentially means such a large percentage variance makes it harder for markets to price a nil-nil result as accurately versus a regular draw. Such metrics could be a good for starting to develop a low scoring games strategy?
Not a long post today but hopefully you have found this useful!