individual statistics on creating success for soccer players

Skip to first unread message

Chris Gluck

Apr 5, 2013, 1:38:48 PM4/5/13
From another tack - working off the NFL Passer Rating concept :)
IF (a big IF) an assumption is made that a right midfielders objective is to generate crosses for strikers to get hold of to put in the goal then an individual statistic like the passer rating can be developed but this works off the assumption that a right midfielder is only on the pitch to deliver crosses 'for attacking'...
your denominator becuase crosses attempted and then the derivatives work from successful versus unsuccessful and then from the successful ones whether or not they generated a shot taken and then a shot on target (a bit dubious here though since a shot on target is a reflection of a good striker as much if not more than a reflection of a good crosser...) but on to the next part of that equation... it becomes then shots on target and then goals scored all relative to the individual player who crossed the ball... but as mentioned this gets on very shaky ground since it is the individual striker who is responsible for putting the ball into the back of the net and to grade the efficiency of a crosser of the ball based upon the success of that other player remains in question...
if I were to critique the NFL Passer rating I would look to quantify somewhere in that equation the ability of the reciever to 'catch the ball'... so a factor in the NFL Passer Rating that is not there is the ratio of reciever catches to quaterback throws... there can be as many as 20? different recievers on a team at any one time... that 'could be on the field of play' so the new NFL Passer rating that is more statistically valid is the equation of equations that adds up all the individual passer rating stats for each reciever first (who caught a ball in that game - this includes starters and replacements as the game is played - bench guys) and then calcuates the 'summation' of those passer ratings.... this accounts for recievers as well as passers...  so even to a guy like me who doesn't follow football the NFL Passer rating is flawed ;) (statistically speaking) unless one of those constants in their equation represents an analysis of 'recievers'... perhaps others know better than I?
so back to soccer - teams track minutes played - touches - etc... so what about OPTA data... there is lots of it available to the public but what value does it bring?  fouls are good but do the public really care about 'flick-ons' or 'headers' - probably not...
in the military we always ask ourselves this question - what is the value of the statistic and how does it help me better understand the 'end-state'?  the end-state in soccer is a goal scored and the accumulation of more goals scored than conceded drives who gets three points...
for strikers this is easy - individually it's how many shots have they taken versus how many shots have been on goal (accuracy) and goals scored (greater accuracy than the reaction of a keeper)
i can see merit in 'adding up' some OPTA data to create a rating of some sort - you see what I offered up in my comment on your article and I am more than willing to work with others to etch that approach in stone - I would ask that you consider my PWP analysis and give me direct feedback on it... 
All the best,
Reply all
Reply to author
0 new messages