We can take this article down two different routes.
The first is simple, but you’ll have to take what I’m saying as true, and you’ll have to understand one statistical term: regression to the mean. Regression to the mean means that it is basically impossible to consistently upset the odds. If you play the lottery for years on end and eventually win it, the chances are that you’re not going to win it the next week as well. What goes up must come down, and vice versa.
Good? Now onto the football.
Winning a game of football has a lot to do with luck.
Whether that last minute shot crashed off the bar to deny your team a victory, or your reserve centre back came the bench to score a worldie, predicting what’s going to happen in any given game is difficult.
How many times have you seen your team dominate the game only to lose to a flukey counter attack?
Bearing that in mind, Rangers are actually playing pretty well. They are creating more good chances than any other team in the league and giving their opponents less chances than any other team. They have probably been unlucky to not take maximum points from the four matches, and the only real area for concern is the rate at which opponents are converting their (few) chances into goals.
Meanwhile, Celtic have been over-performing. They’re scoring more and conceding less than their stats say they should. Both teams are actually fairly close.
So to summarise the short version of this article: Rangers’ level of performance suggests that they will start winning more consistently soon.
Those who are sceptical as to how we’ve reached these conclusions, read on.
You now need to understand a favourite term of football analysts: expected goals.
A penalty is scored roughly 80% of the time, so the expected goals for a penalty is 0.8. Meanwhile that worldie scoring centre back I was talking about earlier? He actually only had a 3% chance of scoring that 30 yard effort, giving that shot an expected goals (xG) total of 0.03. If that penalty and 30 yard stonker were the only two shots your team had in the game, they would finish on an expected goals total of 0.83.
Expected goals doesn’t pay attention to whether the shot scored, hit the post, or went 20 yards wide; it is purely a measure of the quality of chances a team creates.
When calculating an expected goal, you have to look at a few things: distance from goal, centrality to the goal, numbers of defenders between shooter and goal, and the game state (whether the team is winning, losing or drawing). All of these things are then compared to every other shot taken in similar circumstances and then what the outcome was. That gives you your expected goals figure for that shot.
Analysis of predictors of future performance found that expected goals were a far better indicator of how a team was going to do than any other conventional measure (specifically goals, shots, and shots on target). As you can see in the following tables, Rangers have the best, most consistent expected goals totals in the league, while they also have the second lowest expected goals conceded.
To again translate this into simpler terms: Rangers create the best opportunities for scoring more consistently than anyone else, and only Hamilton have been better at preventing their opponents from creating good chances. Things are looking good for Warburton’s men.
Dougie Wright volunteers by providing analytical data to a professional Scottish club. You can follow him on Twitter here.