EPL - 18 Years Of The Draw

I mentioned in a recent post that I would have more on the English Premier League (EPL) Draw, and as a man of my word, here it is.

Thanks to Joseph Buchdahl’s Football-Data.co.uk website, it is possible to look at prices for matches going back to the 2000-01 season, i.e. we have 18 full seasons of price data. Unfortunately back in the olden days, the over-rounds were a lot higher than they are since Pinnacle arrived on the scene, so making comparisons of earlier seasons with more recent ones isn't fair.

It's worth mentioning that Pinnacle's over-round appears to be creeping back up this season, currently at 102.24%. Perhaps not significant to most readers, but if you are taking your betting seriously, it's something to be aware of.

Many readers will be familiar with David Sumpter's book “Soccermatics: Mathematical Adventures in the Beautiful Game” and in other articles he has promoted the idea that backing the Draw in Big 6 games is a profitable strategy:

When two big-six teams meet then all the focus is on one of the two teams winning and the draw is forgotten by the punters and the bookies. These circumstances are somewhat unique to the Premier League, because there are so many ‘just for fun’ gamblers involved in the market and six teams with a worldwide following.

There hasn't always been a Big 6 of course. The EPL began competition with the 1992-93 season, and is now in its 27th season.

In the early seasons, Manchester United dominated, winning four of the first five Championships, before Arsenal, Chelsea and Liverpool joined United in forming a ‘Big 4’.

Of the twelve seasons from 1997 to 2009, these four teams filled the top four places six times, with three of the four in the top four the other six seasons. 

Only Leeds United, Newcastle United and Everton were able to break into the top four places in the final table during this time.

The nine seasons since have seen the emergence of a Big 6, with Manchester City and Tottenham Hotspur being added. These six have taken the top six spots in four of those seasons, five of the top six four times, and only in 2015-16 were two of the six displaced, famously by Leicester City who won the Championship, and less famously by Southampton who finished sixth.

As an aside here, since the formation of the EPL, only three non-Big 6 teams have won the FA Cup, namely Everton, Portsmouth and Wigan Athletic, and Tottenham are the only Big 6 team with no Cup wins in this time.

Back to the League, and I thought it might be interesting to compare the returns from the Big 4 matches (2000-09) and Big 6 matches (2009-date) with the over-round adjusted to the 102% of recent seasons. 

First, here are the results of backing the Draw in every EPL game from the 2000-01 season: 

As I've written before, there are many matches where backing the Draw is akin to throwing your money away.

For Big 4/6 matches over the same period:

Clearly, blindly backing the Draw in all Big games isn't a profitable strategy, and why should it be? Big 4 teams have finished as low as 7th, while Big 6 teams have finished as low as 10th. A 'Big' match isn't necessarily a 'close' one. 

For the unwashed masses, "Back the Draw in Big 6 matches" is a nice and simple system. However, David Sumpter has also written:

It turns out that when two well-matched teams meet (i.e. the probability of a home win is only slightly bigger than the probability of away win) then draws are under-priced. When matches are skewed so there is a strong a favourite (i.e. the probability of one team or the other winning is larger than the other) then draws are over-priced.

While the definition of well-matched teams is vague, there are a couple of measures I like to use. One is to exclude any matches where a team has a 'true' win probability greater than 0.5, and where the 'true' probability of the Draw is 0.25 or less.

Using these filters, backing the Draw in 'Big' matches from 2000 would have resulted in this:

The other method is to calculate the difference between the 'true' win probabilities for both teams. For teams within 40% of each other the profit is 22.81 points from 232 matches, an ROI of  9.8%. A more selective rule of less than 25% and this climbs to 41.78 points from 155 matches, an ROI of 27%. 

Expanding this criteria to include all EPL matches, and the profit is 137.67 points from 1,431 matches, an ROI of 9.6%. 

ROIs around 10% over 18 years should catch everyone's attention.