NWSL- 2016 Portland Thorns FC – Player Statistical Review

Appendix

“SNAFU” Statistical Reminders

Introduction

For those of you unfamiliar with my work, “SNAFU” is a “homemade” statistical system that is called  “Statistically Normalized- All Fudged Up”. There are 10 categories ranging from overall passing to goal scoring to goalkeeping to try to evaluate all players and to also compare teams statistically.

Data retrieved from a site called WoSo  Stats.

Table description

For the Grand Total in the table on page 1, an “average player” should be worth about 50 points. The scale is from 1 to 100 but a player sitting on the bench will have 40.25 points. In theory, all the actions a players can lead to a goal or an error can cause the other team to score. Therefore, a point difference higher in your average would mean that you gave your team a goal scoring opportunity per game as compared to another player. A point higher in team average should mean that the team won by 1 goal, or should have won by a goal.

Revisions from original calculations.

The original formulation had a couple errors. These are changes that are different from the original article, linked below (as “Blame it on Rio Olympics- Part 3”)

The original data was set to “Per 90 Minutes” for most categories. However, this was changed to “Per Game Played”.

For most players this may not matter. But in a theoretical case, let’s call her “Lucky”. She is a substitute who played 1 minute the entire season and scored 1 lucky goal. Therefore, if we were to translate that statistic to “per 90 minutes”, the player would receive “90 goals”.  Her appearance score would still be miniscule. But if this player simply completed 1 pass,  successfully dribbled around an opponent, and had 1  defensive play in that minute with that goal,  she would have the enormously inflated SNAFU score of about 338, not 33.8. She would be the greatest player ever.

So now, with the “Per Game Played” statistic, that goal would set to the correct “1 goal” as she now played in 1 game. With all of the other statistics, such as the completed pass and tackle now just simple plays in an entire game. “Lucky’s” new SNAFU score is now 43.6, good enough to be ranked about 166th in league of 205. Thanks to 1 goal, she jumped roughly 40 places over other players with limited playing time, but she no longer has an obscene score of 338.

The “team punishment” section was found NOT to be adjusted per 90 minutes, and was cumulative. This is  now corrected to “Per Games Played” status.

However, with all other categories being adjusted, it was not fair to players like Tobin Heath, or Lindsey Horan who had several yellow cards over the course of the season. We now have corrected to a per game average to match all other categories to show their “normal game average”.

Minor calculation changes.

For “Defensive Skills” category, the first change is that the overall “fudge factor” is now 0.067 from 0.134.

The original calculation had a factor of 0.134, which is the same as the “possession” categories. However, this inflated all the player averages to nearly a score of 6 out of 10, where the other categories were “normalized” to closer to 5. Lowering the factor to 0.067, which is the same factor as the “passing completion” category, reduced that category average to near 5 out of 10.

The smaller factor makes sense for passing.

The number 0.067 is 1 divided by 15, because from what I could tell from my original data at least 15 possessions are needed for a decent goal scoring opportunity . So, when I doubled everything, such as a goal being worth 2 points and not 1 point, the “possession calculation factor” became 0.134. But with passing, I found that players could pass “randomly” around “to waste time”. Therefore, the original 0.067 factor makes sense for passing as not all passes are to advance the ball toward goal.But with defense, there is sometimes the same effect, just because a player disrupts a play, it does not necessarily mean that the team regains a “clean possession”. The ball could be bouncing around, and a clearance means that you kick it out of trouble,. The same goes for a tackle, you may stop their progress, but someone else will have to recover the ball. Therefore, reducing the factor also makes sense as far as defensive skills.

The other “Defensive skill” change is now “Ball Shields” are added to the other parameters in that category.

A “Ball shield” is where a defender gets between the offensive player and the ball, and lets the ball go out-of-bounds, so that the defensive team earns possession of the ball.

 In the “Team Punishment” category, some scores were doubled to match the “Goal scoring” category.

The scores of Red Cards, Yellow Cards, and Net Penalties were doubled. In my opinion, a Red Card means that you lose a goal. With a goal being worth 2 points in this system, the Red Card became 2 points from 1 point. Yellow cards were doubled to 1 point from a half-point. The same goes for “Net penalties”. If you win a penalty kick, you just earned the team a theoretical goal, therefore it is worth 2 points, and therefore if you concede a penalty, you likely lost the team a goal, and it is now minus 2 points.