Although my statistical based model has been a fan of Liverpool for most of the season, I’ve gone on record the last couple of weeks to state that I think the market (and probably the wider public) are now over estimating the chances of Liverpool winning the league.

What is the Market saying?

Right now, the Betfair Premier League outright market looks like this:

Betfair

For those that aren’t familar or comfortable with betting odds, the above Betfair market suggest that the chances of winning the league are as follows:

Man City 54%
Liverpool 34%
Chelsea 12%

How do I view the Title race?

Looking at the predictions generated by my stats based betting model which have been ran through 10,000 simulations my current probabilities are:

EPL

Someone asked me on Twitter why I was so bullish on City, and correspondingly why I was giving Liverpool less of a chance of winning the league than most others.
I’ve had a look at the engine of my model and it looks like there are two main reasons for this:

1 – I rate City as marginally the better team; and

2 – City have 5 of their remaining games at Home.  They are an impressive team at home, and given the quality (or lack of) of the opposition that they will face at home it wouldn’t be unreasonable to think that City will win all 5 of those games.

Of course, I am aware of the fact that the title is in Liverpool’s hands and if they win all of their remaining games then they will most definitely have been deserving winners, but as a gambler I work on probabilities – and I see Liverpool winning the league just slightly more often than once in every 4 runnings of the 2013/14 Premier League.

 

Predictor

Anyway, that’s enough about what I think.  So who do you think will win the league?

For those who would like to make a logical estimate of who should win the league I am attaching a small Excel spreadsheet in this zipped file: EPLPredictor
(It is probably best to choose “Save As” when the dialogue box opens.)

The file is easy to use, all the remaining league fixtures for Liverpool and Man City are listed.  All you need to do is enter your estimate of the Home and Draw probabilities for the remaining games and the sheet will do the rest.

Hopefully the act of inputting your probabilities and seeing your predicted final league table based on the expected league points will let you see the size of the task that Liverpool have ahead of themselves if they are to win their very first Premier League crown.
It is possible, but I believe there is just a 1 in 4 chance of 2013/14 being the season that this occurs in.

 

 

  • beige

    nice post… just missing disclosure of any financial interest! 😉

  • http://www.StatHunting.com Steve Fenn @StatHunting

    Good stuff, Colin, but stepping away from the betting perspective, the striking thing here is just how much single outcomes (especially one outcome in particular) will swing the race. So, even if the key match’s odds really are 45% LFC win, 22% draw, and 33% City win (and I like those odds), the actual result will be enormously important.

    While your model on its own gives a final expectation of 84.94 City points, and 82.61 LFC points, if Liverpool wins that one match, it swings to 83.71 City xPts, and 84.04 LFC xPts. City win would shift it to 86.73 / 81.04, while a draw keeps xPts close to final expectation at 84.73 / 82.04.

    Also worth noting that the tiebreakers are really close, with City at goal differential of +52, and Liverpool at +49. If LFC win the big match by multiple goals (and I admit that’s a huge if), they’ll be in good position for the tiebreaker, too.

    • http://www.StatHunting.com Steve Fenn @StatHunting

      Side note: when you run Monte Carlo sims of the remainder of the season, what’s the title equity at stake in that May 13th fixture alone?

    • Colin Trainor

      Steve,
      As I would expect, you’ve raised some very valid points. And because each game is only played once there is a real possibility (1 in 4) that the results could fall in such a way that would see Liverpool win the league.

      I had a thought about the GD too, but personally I think that City will shade the GD (even if they are beaten at Anfield) due to the 5 pretty easy Home games they have remaining.

      I’ve obviously had to vastly simplify all the remaining nuances in the title race to publish this article, but the reason I did is that I think it shows just how tough a task Liverpool have if they are to win the title. And that’s regardless of whether they achieve it or not.

  • Errorr

    Those percentages seem about right.

    I doubt betting lines are in anyway efficient enough to correct for recency bias inherent in how most people see the league. Although, favorites are almost always overvalued.

    The question for me would be if you could call the betting lines rational if you adjust for things outside your model. These would mostly be health effects and effects of playing “safe” teams who may field reduced quality squads as winning is no longer the only primary motivation.