Season Reviews

End of season review: Castrol Index vs. PSP Player Ratings

Photo: Nicolae Stoian

Toward the end of the 2012 season, comments from readers led to a post comparing the ranking of Philadelphia Union players in the Castrol Index to how players would be ranked according to the average of their PSP player ratings. We returned to the idea at the end of the 2012 season and also at the midway point of this season. Now that the 2013 season is at an end, we do so again.

The Castrol Index is designed to be an objective measure of player performance in which “every move on the field” is assessed based on “whether it has a positive or negative impact on a team’s ability to score or concede a goal.” Player ratings, while informed by objective measures such as match statistics, are an inherently subjective assessment of how a player performed in a game. Nonetheless, it is interesting to see where the two ratings systems converge or diverge.

The chart below lists Union players according to their ranking within the team via the Castrol Index and PSP’s postgame player ratings. The Castrol Index information additionally includes each player’s ranking within MLS. Also included are Castrol Index and PSP player rating ranking within the team for the end of 2012 and the midway point of the 2013 season. The 2012 Castrol Index numbers include the league ranking for Conor Casey, Sebastien Le Toux, and Jeff Parke, who were not with the Union that season. Excluded from the list due to lack of playing time are Matt Kassel (434th in the league, 88 minutes), Chris Albright (454th in the league, 3 minutes), Roger Torres (460th, 72 minutes), and Don Anding (470th, 30 minutes).

Castrol v PSP end of season 2013Castrol Index

With the exception of Brian Carroll and Danny Cruz swapping the 10th and 11th spots, the top 10 Union players in the Castrol Index remained the same from the midseason point. Zac MacMath’s movement from 6th at midseason to 3rd at season’s end was the most significant, which is understandable when one considers that 7 of his 12 shutouts came in the second half of the season.

Jack McInerney and Conor Casey were No. 1 and No. 2 in the midseason Castrol Index Union rankings and remain so with the conclusion of the season. Both players fell a few spots in the league rankings from their positions in the beginning of July — McInerney was being discussed as a Golden Boot contender and Casey was not far from coming off of scoring 3 goals in two games when we posted the midseason rankings — but they still finished high. McInerney climbed 7 spots from his 2012 finish and Casey, scoring 10 goals for the first time since 2010, moved up a whopping 212 spots from an injury marked 2012 season in Colorado. Casey’s 212 spot improvement was by far the most significant when we look at Union players’ rankings in the league, followed by Keon Daniel moving up 118 spots, Michael Farfan moving up 75 spots, MacMath climbing 55 spots from his 2012 rankings, Danny Cruz moving up 51 spots, Amobi Okugo 44 spots, and Sheanon Williams 42 spots.

In fact, every player in the the top 10 improved their 2013 league-wide ranking compared to 2012 except for Jeff Parke, who fell 59 spots from his 2012 ranking, and Brian Carroll, who dropped 10 spots. Comparing 2013 to 2012, Antoine Hoppenot (-30) and Michael Lahoud (-143), who missed much of the season with injury, also saw a drop in their rankings within the league.

PSP Player Ratings

The Top 5 Union players based upon the averages of their PSP player ratings at the end of the season features 4 defensive players and only one offensive player, Conor Casey. Amobi Okugo repeated his No. 1 2012 finish in 2013. Only 3 midfield players, Farfan, Kleberson, and Le Toux, made the top 10 and Kleberson saw relatively few minutes. I’ll leave it to others to argue whether Le Toux was a midfielder or a forward.

There was a fair bit of movement within the Top 10 Union players in the PSP player ratings between the midseason and the end of the season, perhaps the most notable — and understandable — change being McInerney’s drop from the No. 1 spot in July to No. 6 after the last game of the season. Perhaps equally understandable was MacMath’s 4 spot rise from No. 9 to No. 5. Also notable was Farfan’s 5 spot improvement from No. 13 to No. 8. While Raymon Gaddis managed to finish in the top 10 after placing at No. 12 at the midseason point, Hoppenot fell from No. 6 to No. 11 while Carroll went from No. 8 to No. 12.

In case you’re wondering, the overall player rating average in the chart above for the Union in 2013 is 4.79. Including players not on the chart above because of lack of minutes and players no longer with the team, the average was 4.89. In 2011, the overall average was 4.99.

Castrol Index vs. PSP Player Ratings

As described above, comparing an “objective” ranking system such as the Castrol Index with a subjective system such as player ratings can be difficult. At best, doing so is a good conversation starter. For example, would most Union fans agree with Keon Daniel coming in at No. 9 on the Castrol Index? Is Daniel coming in at No. 16 in the PSP player rating averages unfair? Was Jack McInerney the Union’s best performer in 2013, as shown in the Castrol Index? Or was Amobi Okugo the best performer, as shown in the PSP player ratings averages, or was it someone else?

In the end, you can decide. Feel free to share your own rankings below.

6 Comments

  1. Jeremy Lane says:

    Love this stuff, Ed. Good work.

  2. Maybe I’m mistaken, but isn’t the Castrol Index a combined measure of quantity and quality as opposed to PSP’s strict adherence to quality? I would expect that is why Keon Daniel is high despite his lack of quality.
    ———–
    If so, you can’t really compare the two lists without combining PSP ratings with minutes-played and adding them up, right?
    ———–
    It’s kind of like comparing a list of Phillies’ players sorted by number of Hits and comparing it to Phillies’ players sorted by Batting Average (while eliminating those who don’t have enough hits to qualify)

    • Ed Farnsworth says:

      Chris, thanks for reading and for your comment.

      Regarding the question of how minutes played is factored in, the Castrol Index FAQ page says:

      “6) Are there a minimum number of minutes a player has to play to be ranked in the Castrol Index?

      “All players are ranked irrespective of how many minutes they have played in the season.

      “Those who have not played a specified amount of minutes each month (equal to approx 60% of season game time) will have their Castrol Index score divided by that amount of minutes, therefore penalising players who are either injured, not selected or suspended for a considerable period of time.

      “7) Surely the player who plays the most games will top the Castrol Index?

      “No, the Castrol Index is calculated on a per 90 minute score, measuring the quality of each player performance rather than the number of games played.”

      The most minutes that can played in a 34 game season is 3060 minutes. 60% of that is 1836 minutes. Daniel is the only player in the top 12 of the Castrol list who played fewer than that at 1719 minutes.

      The bit about not playing “a specified amount of minutes each month (equal to approx 60% of season game time)” kind of throws me. Is it minutes per month or per season that is the important factor?

      Anyway, the post is meant to be a conversation starter more than anything else. Like I said in the post, comparing an objective and a subjective measure of performance can only go so far, but it is interesting to see where the two rankings converge and diverge.

  3. I have to think the Castrol overweights goals, assists, and saves- goals are often a product of team effort. I always struggle against numerical assessments of the game. Some goals are wonder goals, others are expected. Some saves are everyday, others are phenomenal. Same goes for assists. I love LeToux but it’s hard to argue he’s our fourth biggest contributor this year. You can say this for passes too. I agree with the PSP rankings that Okugo and Parke were this season’s best, and MacMath, though I admit this is a late season thought and that I’m forgetting his early season bad days when I say that. Basically, soccer isn’t baseball and performances are most usefully evaluated qualitatively, good job PSP.

  4. I’m not a huge fan of the Castrol Index, just because I think it favors attacking players, and goalkeepers more than other positions. I don’t know how you guys feel about Whoscored, but I wouldn’t mind seeing how you guys rated the Union Players based off of the ratings Whoscored gave them.

    I think it would be nice to see since theres is purely based on stats, and yours include a little more objectiveness.

Leave a Reply

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

*