Barclays Premier League: Return of the armchair fans' guide to football statistics

Steven Gerrard's Liverpool look underrated considering they finished second last year

On the back of a profitable World Cup, Jack Houghton is attempting to transfer his know-nothing success with international football to the Barclays Premier League. It's all about the numbers, he reckons, and they point to an overvalued Chelsea and Manchester United...

For almost a decade I've compiled ratings for international football teams, using them to bet in major football tournaments like the World Cup and the European Championships. In the World Cup just gone, the editors at betting.betfair, in all of their beatific wisdom, allowed me to demonstrate how similar statistics and ratings - many of them freely available - can be used by the armchair football fan to improve the profitability of their punting.

The experiment was a success. Argentina and Germany were put up as pre-tournament tips for the outright market, and only England's failure to qualify from their group and a late recommendation to favour CONMEBOL teams in the closing stages blighted an otherwise perfect run of match tipping. With this in mind - and after a few weeks of badgering - the editors have agreed to let me try my hand at the Barclays Premier League.

So, over the course of this season, at monthly intervals, I'll be attempting to demonstrate that a statistical approach to football punting can, with minimal time invested, lead the armchair fan to significantly more profitable outcomes than they would otherwise enjoy.

As I've written before, I don't consider myself a football fan. I would struggle to name more than a few Premiership managers; I have no idea who the star players of the last season were; and I have no idea who has been transferred, and to where, in the off-season. Rather than viewing this ignorance as a hindrance, though, I view it as my biggest strength: it means I can more effectively block out the hype and hyperbole that leads most punters to make poor bets and instead focus on the only thing that matters - the numbers.

Take the market for the Premier League Winner for the upcoming season. Stupidly, perhaps, I would have expected Manchester City to be favourites. After all, they won it last year and, whilst not claiming that a league that awards three points for a win and one point for a draw is a perfect reflection of the relative merits of football teams, it would be difficult to argue against the assertion that, on balance, Manchester City were the best team across the 38 matches that made up last season.

(Having said that, I'm sure there are many people who would argue just that. What I'm talking about, though, is having a statistical basis from which to argue that Manchester City weren't the best team. My own ratings - of which more in subsequent weeks - and various other robust approaches to assessing relative footballing form, make it hard to come to any other conclusion.)

Why, then, are Chelsea, who finished four points behind Manchester City in third place, the current 3.02/1 favourites, with last year's winners only second-best at 3.65n/a? Why are Manchester United, who finished 7th last year, the current third-best at 5.79/2. Perhaps most surprisingly, why are Liverpool, who finished second last year, only fifth in the market at a whopping-looking 13.5n/a?

No doubt there are reasons. Managerial changes, player transfers and relative performances in pre-season friendlies will all have been factored into those odds. How, though, are these relatively subjective factors accurately assessed?

In an admittedly very crude piece of analysis, looking at the entire history of the Premier League, the winning team in each year had finished either first or second the previous year on 17 occasions, or over 80% of the time. Although certainly not definitive - there could be a plethora of reasons why this was the case, and none of those historical reasons need necessarily to apply to the upcoming season - it suggests that the placings in this year's Barclays Premier League might not differ as much from last season as the market predicts.

And there are various others indicators of a lack of volatility in the results of the Premier League: only five teams have ever won the competition; only eight teams have ever filled either of the top two spots; and even a cursory glance at recent results shows that the League is dominated by the same seven or eight teams year on year.

It might be ignorantly simplistic, but this quick-and-dirty statistical approach of using last year's league results to make predictions for the forthcoming season suggests that the following teams are currently overvalued by the market:

Chelsea 3.02/1
Manchester United 5.79/2

And the following teams are undervalued:

Manchester City 3.65n/a
Liverpool 13.5n/a
Tottenham 75.074/1
Everton 160.0159/1
Southampton 1000.0n/a

So, as a starting point, but with more technical statistical analysis promised as the season progresses, my opening recommendations are:

Back Manchester City at 3.65n/a to win Barclays Premier League
Back Liverpool at 13.5n/a to win Barclays Premier League
Back Tottenham at 75.074/1 to win Barclays Premier League
Back Everton at 160.0159/1 to win Barclays Premier League
Back Liverpool to finish in the Top 4 at 2.0421/20
Back Everton to finish in the Top 6 at 2.68/5
Back Southampton to finish in the Top 6 at 20.019/1