If you have any queries, questions, or debates you’d like me to try and support, or if you want to use some of my data or advertise, please contact me on lpcorbett at hotmail dot com




4 Responses to “Contact”

  1. Johan December 5, 2011 at 13:17 #

    Love it mate, these are the kind of stats that offer proper comparables!

    Look forward to seeing how this develops.

    • lpcorbett December 5, 2011 at 13:46 #

      Cheers mate, I thought you might enjoy the Dutch angle with RvP!

  2. Rónán December 6, 2011 at 15:03 #

    Site and method looks very interesting Liam. Definitely addresses some of the points of contention when measuring players on, for example, Opta stats alone. Lots of players get accused of never scoring in the big games. I’ll be interested to see how it develops.

    I’m sure you’ve got your work cut out already, implementing the index you’ve proposed, but I can’t help but offer suggestions for subsequent development, being a coder and a bit of a stats nerd myself 🙂

    1. The ranking of a team, at the time of the goal, offers a good weighting for a goal’s value, but it doesn’t address form, which is often included in argument e.g. “So what, he scored a brace against a side who haven’t won for weeks”. A team in 11th might have the worst form in the league.

    So I’d propose a calculation of rank that’s weighted on:
    – league position of opposition at the time of the game
    – form of the opposition at the time of the game.
    Keep a separate form table, and allow joint positions: e.g. if top 3 all won last 5, they are all in position 1.

    Now you have a table of teams in your database with 4 columns:
    Team name, League Table Position, Form Table Position, Rank Score

    Then you just need to weight form against league position to calculate Rank Score. Lets say that form is 1/5 and league position is 4/5.

    At the end of each week, populate RankAndFormScore with:
    (4/5 * League Table Position) + (1/5 * Form Table Position)

    You’ll end up with non-integer values for Rank Score, but if you sort by ascending Rank Score, the position of each team’s row will give you your new (weighted) rank.

    Different criteria can be used to temper the absoluteness of scoring based on the opposition’s league position but you’ll quickly balloon the amount of data you need for your calculations, down to storing a full list of results to date in a season.

    2. You might consider goals against in the weighting too. The Newcastle sides of old used to go out and aim to simply score more than the opposition. A good striker might stand a great chance of scoring against a side that is ‘decent’ on paper, but actually lacks a decent defence (though this levels out for everyone, and would be difficult to score – you’d need to weigh up goals conceded by a team against total goals scored in league or something)

    3. It would be interesting to calculate the probability of a goal in next game, as a measure of scoring record against strength of next opposition. This would create a good test of your index, as you can score probabilities of scoring a goal against actual goals in some interesting striker head-to-heads. When you’ve calculated your odds on scoring a goal, it might be fun to measure them against the bookies.

  3. lpcorbett December 6, 2011 at 17:44 #

    Cheers Rónán.

    Yeah, I’ve got a few ideas to implement in future weeks, also some stuff to add as features, though I must admit that i’d never thought of incorporating form as well. That would definitely give a clearer picture of the opposition at the time – maybe on a 3 game or 6 game basis. I’ll have to include that one down the line, if I ever finish building the historical and foreign leagues in! I’ll give you a nod in the commentary.

    With regards to point 2, I’m traking goals against in the database and plan to use it at some point. Though was thinking more about a defence analysis rather than tying it in with the goalscorers. There is some benefit in seeing the importance of the goal though – whether it be an equaliser, winner, or just one of many in a rout.

    Probability index might be have to be a long term measure, and I may have to reach out on that one!

    Thanks for the input – this is exactly the type of thing I was hoping for.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: