There is something wonderfully pure and noble about a shot on target; whether it’s a 30-yard thunderbolt or a six-inch mishit, the player in question has done all they can to put the ball in the back of the net.
Theoretically at least, if you have more shots on target than your opponent then you *should* win the match. Of course, football isn’t as simple as that as we know, but from a mathematical standpoint you have a greater chance of scoring if you shoot on target more readily, and so consequently you have a greater chance of winning if you are shooting more accurately than your opponents.
The advanced analytical world, which is growing in number and popularity, has devised Expected Goals as it’s go-to for predicting the outcome of football matches. This takes shots on target data to the next level by quantifying just how likely those shots are to be goals (e.g. a close range effort from the middle of the penalty area is more likely to result in a goal than a shot from the touchline).
That is slightly too complex for our limited brainpower, and so instead we’re happy to have devised a much simpler metric: Shots on Target Ratio. This, quite simply, is the number of shots on target that a team has divided by the number that they concede. The logic is the higher the number, the more dangerous you are in the final third while limiting the damage at the other end of the pitch.
Does this theory hold any weight? We did some number crunching for the English Championship, and this is displayed in the graphic below:
On the left we have the actual league table as it stands, and on the right how it would look based purely on Shots on Target data. The coloured numbers dictate how much a side is over/under achieving compared to their league position (green for places up, red for down, yellow for stayed the same).
What early conclusions can we draw? Well, certainly that QPR, Wolves and Sheffield Wednesday deserve to be higher up in the league table, and that perhaps Huddersfield, Birmingham and Reading are punching above their weight.
Think about it: a ratio less than one dictates that the team in question is conceding more shots on target than their opponents, and so they must rely on good goalkeeping to maintain their vanity defensively and hope that their attackers are suitably skilled at finishing to make the most of the limited chances that come their way.
Now is the time to test the theory: we have a set of midweek Championship fixtures, so let’s see if we can pick out some value by throwing traditional bookmaking logic out of the window and backing our SOTR heroes instead:
- Brighton vs Huddersfield: Brighton to Win (19/20)
- Preston vs Cardiff: Cardiff to Win (23/10)
- Sheffield Wednesday vs Bristol City: Sheffield Wednesday to Win (Evens)
- Wolves vs Barnsley: Wolves to Win (Evens)
- Aston Villa vs Brentford: Aston Villa to Win (4/5)
- Rotherham vs Notts Forest: Forest to Win (17/11)
The teams we have backed are the ones that are at least three places higher in our SOTR than their opponents, and their price represents good value. With six fixtures fitting the bill, clearly there is a good sample size to see just how effective the theory is.
And yes, we understand the nuances of football can disrupt a clean data analysis: the deflected goal, the ‘park the bus’, the home bias. The SOTR numbers are taken to offer long-term value, and so across a whole season they should, ultimately, pay dividends.
But then you could argue that Leicester City’s title win of 2015/16 makes a complete mockery of this system. Nobody said it was perfect….