Welcome to the Alternative Premier League

Photo by Duncan

Have you ever watched the Premier League and felt a sense of injustice? Have you ever wondered what the league would look like if points were awarded on the basis of winning margins, away wins and strength of opponent? Do you think Manchester United’s 7-1 obliteration of Blackburn, Blackpool’s shock win at Anfield and Wolves’ win against Chelsea deserved more than the standard three points? If you answered ‘yes’ to all, or any of the above questions then the Alternative Premier League (APL) might be right up your street.

As the name suggests, the APL offers an alternative, more detailed look at order in England’s top flight, and you can see the league evolve right here on EPL Talk in a new, regular column throughout the season.

To put it simply, the APL rewards points to teams based on the following criteria:

  • Result of the game (win, draw or loss)
  • Score of the game (the goal difference from the individual match)
  • Strength of opponent
  • Comparative APL position of the two teams
  • Away win bonus

Because of the amount of variables that contribute to the winning of points each game is unique and the teams who entertain and score goals are the teams who get the big rewards. It is also worth noting that because of the way scores are calculated, losing teams are also awarded points. While it seems strange to award points to a losing team it is a useful tool for rewarding a good, albeit unsuccessful, performance.

Let the season commence!

18 thoughts on “Welcome to the Alternative Premier League”

  1. This is an awesome idea, though emphatically subjective. I look forward to seeing how the EPL season turns out under these criteria.

  2. Now who wants to go back and apply these rules to last year’s table? Seems like a good place to start if we’re to determine their usefulness.


  3. I do believe a more in depth look at results is a great thing and the APL is a good start, my main complaint towards this type of analysis is the disregard of a strong defence. I continue to hear EPL fans talk of the ‘right’ way or the English way of playing and entertaining play but a win is a win. The most ‘entertaining'(Blabkpool) and least entertaining (Birmingham) were both relegated. Why then reward a goal scored buy not a goal prevented? The worst a team can do of they prevent the opponent from scoring is a draw.
    I know I am being picky here but I have seen this analysis fail on other sports.
    That being said, I’m excited to see a different approach t football study. I can’t wait to see how t?is unfolds.

  4. Are you trying to suggest that Manchester United didn’t deserve the win the PL last year or justifying that the 3-1-0 point system is obsolete?

    Goal difference already adds up throughout the season.

    Strength of opponent varies so much week to week, thus; form is temporary, class is permanent.

    Also, if you’re in a winning position after 80+ minutes and you lose or draw, shouldn’t you be penalized? Or how about being 4 goals up to nil….

    If your system gives, it also must taketh away.

  5. Put it unsimply – what’s the actual model you’re using here, or at least a vague sense of how it would evaluate a specific game.

    1. Not exactly. NCAA standings are completely poll driven. There is no table based on actual won-loss-tie points like the EPL. I actually like the fact that the BCS considers more than one poll/statistic to rank teams. For me, its major defect is that it does not include all teams, such as Boise State. Other than that I think it is a better system than, say, just accepting whatever the USA Today poll says.

      Sorry for the digression. :-)

  6. That’s exactly what I thought when I saw it. Which is why i am not fond of it at all. A system like this is already ruining one of the best american leagues out there.

    1. Easy, now. This isn’t intended to be a replacement for anything. It’s a subjective alternate scoring system FOR FUN, not an endorsement of differential scoring or a critique on the 3-1-0 scoring method.

      Nobody’s suggesting that the EPL adopt this system.

  7. Nice idea. I look forward to following the APL. “Strength of opponent” is always a sticky issue, but let’s see how you work it out.

    Rugby fans are already used to differential scoring. Scoring four tries gets you a bonus point. Losing by seven points or less gets you a bonus point. As a result you seldom see a rugby match merely being played out. Both teams usually have something they are trying to achieve right up until the very last second.

    Not saying the EPL should go that way, but it will be fun to see how your league standings differ from the actual scoring. I’m assuming you’re prepared for the slings and arrows of outrage. 😉

    1. With a decent model, strength of opponent shouldn’t be too hard – other than the three newly promoted teams, you’ve got a whole year of still pretty good data about performance in the Prem, and you’ll have rapidly developing data for this year – within a week or two you ought to have enough fixtures having been played to start relying primarily on this season’s data.

      A fairly easy way to model it is to take last year’s league, and treat every team’s two games as a two-legged tie, adding the scores up. Then you can compare teams first by head to head records (i.e. Arsenal is treated as a better team than Aston Villa having beaten them 5-4 on aggregate) and by records against the same team (Arsenal beat Birmingham 5-1, but Aston Villa drew 1-1 against Birmingham). From there it’s a matter of weighting, but you should be able to get a fairly good weighting by using the same process on the 09-10 season to get comparative strength rankings for 10-11, and seeing how good a predictor of outcomes you’ve got.

      But for the most part, within a league enough games are played to get quite good data. 380 games per season is a lot of data. I mean, you can certainly come up with even better models – one that figures out the strength of a squad based on its players instead of on past results directly would be even better. But it’s not too hard to come up with a pretty good model.

  8. Okay, on the exterior the idea is wonderful, but do we really want to alienate an entire footballing style? Attacking football is brilliant, but there is also a certain beauty in the defensive game – see under Inter Milan vs Barcelona, Champions League 2009/10.

    Sure, goal counts would go up, but you could argue that it is the presence of defensive football which makes the attacking game so fantastic.

  9. Disagree with the entire premise, but my main complaint is that the opponent strength will balance out by end of season. Each team plays all others twice, once home once away. The only reason opponent strength is needed in NCAA football is that only 12 games are played each year for each of the 120 teams. This doesn’t allow sufficient saturation of opponents, so strength of schedule varies wildly between two different teams.

    As a side note, I currently run an NCAA football computer ranking at lawlersranking.blogspot.com, and I’ve considered on several occasions using the same basis for EPL games. At the end of the season, the rankings would directly reflect overall goal difference, but it would be fun to follow over the course of the year. If there is interest, I could add the EPL rankings for this season to see how it plays out.

  10. Hey guys, thanks for the feedback. For those of you who are worried about the emphasis being on attacking football, perhaps I was a little misleading with my piece. Yes attacking football is rewarded but, for example, a 3-0 victory would warrant a higher point reward than a 3-1 victory – so a strong defence and clean sheet are still important

  11. Interesting idea. I assume this would make up for games like Spurs/Chelsea last season with certain goals NOT actually counting, no matter what the ref thought.

    I thought I’d suggest adding a Wealth Factor modifier. Going along with the rest of the idea this would be very subjective. If Man City plays Swansea, for example, a MUFC win would be multiplied by .5 as a handicap, because anyone sitting on their bench probably makes a good deal more than the entire Swansea team combines. Swans winning would be a 2* multiplier. Liverpool and Arsenal would be at parity so no multiplier needed and Stoke would always have a negative multiplier because who roots for Stoke?

    Just a thought.

  12. A few random thoughts:
    (1) you didn’t actually explain how the point scoring system would work. I know you said what the criteria are, but go further – how many points are awarded for a win? What is the multiplier factor for beating a stronger opponent? How is the strength of the opponent determined? How is the “strength of the opponent” factor distinguishable from the “relative league position” factor? etc etc

    (2) As others have said, it would be nice to see how last season’s table would have panned out with this scoring system.

    (3) How/why would a losing team get any points, as you mention? Your critera specifies that one of the criteria is “goal difference from the individual match” – in which case, a 5-3 loss would be the same as a 2-0 loss, so “goals scored” obviously don’t earn any points. So how would a losing team get points?

Leave a Reply

Your email address will not be published. Required fields are marked *