The numbers are in, the calculations have been done, and it's time to announce the BigSoccer Manchester United board player of the month for August as voted by you in the player ratings threads! The top 3 are: *************************************************** 1st Darren Fletcher - 3 appearances @ 7.43 *************************************************** 2nd Wayne Rooney - 4 appearances @ 7.01 ************************************************** 3rd Patrice Evra – 4 appearances @ 6.97 ************************************************** Full list: Player/Appearances this month/Average rating out of 10 Darren Fletcher - 3 - 7.43 Wayne Rooney - 4 - 7.01 Patrice Evra - 4 - 6.97 Ben Foster - 4 - 6.76 Dimitar Berbatov - 4 - 6.44 Nani - 3 - 6.43 Jonny Evans - 3 - 5.75 John O'Shea - 4 - 5.71 Michael Owen - 3 - 5.47 Ryan Giggs - 3 - 5.36 Antonio Valencia - 4 - 5.29 Wes Brown - 3 - 4.89 ------------------------------------------- Nemanja Vidic - 2 - 7.15 Paul Scholes - 2 - 6.76 Darron Gibson - 1 - 6.33 Fabio - 1 - 5.88 Gary Neville - 2 - 5.59 Park Ji-Sung - 2 - 4.61 Michael Carrick - 2 - 4.25 Anderson - 1 - 3.27 Players below the line didn't get enough games to qualify for player of the month. To do so, they have to feature in more than 50% of the matches. When I asked the question the other week, some people said that they would prefer it if substitute appearances didn't count. Here's a seperate list that only includes starts. I'm not giving a pseudo-award out for this one so have just listed it by highest rating. Player/Starts this month/Average out of 10 for each start Darren Fletcher 3 - 7.43 Dimitar Berbatov - 2 - 7.30 Nemanja Vidic - 2 - 7.15 Wayne Rooney - 4 - 7.01 Patrice Evra - 4 - 6.97 Ben Foster - 4 - 6.76 Paul Scholes - 2 - 6.76 Nani - 3 - 6.43 Gary Neville - 1 - 6.02 Fabio - 1 - 5.88 Jonny Evans - 3 - 5.75 John O'Shea - 3 - 5.70 Antonio Valencia - 3 - 5.45 Ryan Giggs - 2 - 5.14 Wes Brown - 2 - 4.78 Michael Carrick - 2 - 4.25 Michael Owen - 1 - 3.43 Park Ji Sung - 1 - 3.33 Anderson - 1 - 3.27 Man of the match awards: Birmingham City (H) - Wayne Rooney - 7.60 Burnley (A) - Patrice Evra - 6.60 Wigan Athletic (A) - Dimitar Berbatov - 8.37 Arsenal (H) - Darren Fletcher - 8.67 Top 3 Indivual performances of the month: Darren Fletcher vs Arsenal (H) 8.67 Dimitar Berbatov vs Wigan (A) 8.37 Wayne Rooney vs Wigan (A) 7.87 Bottom 3 stinkers of the month: Anderson vs Burnley (A) 3.27 Park Ji Sung vs Burnley (A) 3.33 Michael Owen vs Burnley (A) 3.43 That's your lot for now, thanks for all the votes.
WOW awesome stats! Me likes it. It's also great that you don't make a summary after each match and make waiting for the end of the month more exciting.
Thanks. Erm, now that they're up does anyone understand electoral procedure or have a knowledge of averages and numbers? Strange question, I know, but there's a possibility I might not have used the fairest averaging criteria. For working out averages for the month, I took every vote received by that player over said period of time and simply divided it by the total number of votes cast. I felt this would make every vote count which is good. On the downside, it means more weight is given to the matches where more people voted. At one end of the scale, 15 people gave ratings against Burnley, 24 against Arsenal. Not sure if that's the right way to go about things now and whether or not I should give each match, rather than each vote across the month equal footing. I could do that by calculating each player's average for each match (which I did anyway), adding the averages together and then dividing that by the number of games they played that month. That'd mean taking away some of the 'power of the people', I suppose as it would reduce the importance of each vote when it came to producing the results at the end of the month. Am I arguing proportional represenation vs some other thing? First past the post, maybe? If I'm just nuts and nobody has a clue what I'm on about we'll continue to do it the same way. Edit: Thought of an example that might make it clearer. Under the current system if Joe Bloggs plays 2 matches in a month, and in: Match 1, 10 people vote, all giving him 10/10 Match 2, 2 people vote, both giving him 0/10 I would take 100 (his total score) and divide by 12 (total votes) and get 8.33 as his monthly score. The other method would be to take 10 (his average from match 1) and 0 (avg. from match 2) and he'd get a 5 over the course of the two games. I'm taking this a bit too seriously, aren't I? I'm going back to bed.
The second method is better, I think. It's probably more objective and realistic, because each game speaks for itself and it's difficult to just combine all votes and get an average. bt that is just me.
Personally I think that makes sense. Although it is important to have consistency (and this is also reflected in the ratings too), showing up in the big games against the likes of Arsenal should be weighed more heavily than when facing Burnley, Wigan, etc. It is far more likely that more people will watch and weigh in on the bigger and 'more important' matches (e.g. Arsenal) that the less glamorous week-to-week ones (e.g. Wigan, Burnley) and as such I think it leads to fair representation. I can see what you mean ultimately but I think unless there is under a certain amount of votes on a particular match (which is doubtful given the coverage we get as club) that the current system works best.
Great stuff ROM, appreciate the work you are putting into this. I think its a great rating season for the board to try out this season. As for what form of statistical analysis you use, I think the system you are using is fine. It will attribute more weight to games where there are more votes, but as Billy points out, this may indicate "bigger games". Although now that I think about it, the number of votes cast could also be influenced by a number of things, and not just the 'size' of a game. A game in a bad timeslot for example may have lower views / lower votes, but might still be a big game (European games for example). Ultimately I doubt it will be a big factor however. Most of the board watches most / all of the games. I suspect we will have a pretty similar group of ppl voting for most games, and the numbers of voters wont change that much.
Think it should be an average based on a per game rating personally. Relying on people to turn up and vote leaves too much room for reactionary voting to skew the results. In either case, it seems like you have per-game ratings anyway so there's no reason why we can't have both Good job with separating the subs; it makes far more sense to me to isolate those in starting roles from those with 5 second cameos. Are you thinking of uploading the spreadsheet somewhere every once in a while? It'd be nice to play around with the data around the January transfer window or so.