Thursday, October 25, 2012

Preliminary Evaluation, Crew

Only one game left for the Crew but before it all ends I wanted to take a look at different ways to evaluate how the players did this year and, by extension, how team management did.

A tad long on this post and it goes on about stats. Fair warning. If that's what you're looking for - then settle in with your favorite scotch or pour yourself a barley pop. Let's dig in.


Earlier this week Craig Merz brought up Robert Warzycha's poor September/October record with Technical Director Brian Bliss. I can't figure how Merz got his winning percentages there no matter how I cut it.

In his article Merz states Warzycha is "...10-16-5 for a winning percentage of .403" in Sept and October since 2009. That's a total of 31 games and 10 wins in those two months. That comes up as a .323 winning %, not a .403 as Merz states. I also ran it as percent of possible pts earned (a true soccer metric) and that came up 35 pts earned out of a possible 93, or 37.6% earned (.376). I haven't a clue where he got .403 for those months.

[EDIT/NOTE: Thanks to Mr. Steve Sirk for informing me on how Merz got his winning percentage.]
Steve Sirk ‏@stevesirk
@HelltownBeer MLS calculates winning % in the traditional US sense. (W + (T*0.5))/GP

What's funny is that Brian Bliss goes right along with Merz though, saying: “It is an alarming statistic, for sure. We’ve got to delve into this.”

That said, calling out Craig Merz isn't the point of this post. The reason I bring it up is because I need some baseline numbers. The point Merz is trying to make (Warzycha struggles late in seasons) is valid.

Here are the % of Points Won by Warzycha (RW) coached teams:
51.2% : RW Overall
55.6% : March thru August
37.6% : September and October

Note the measurable dip in Sept and Oct. A similar thing happened this year in that the Crew struggled in those months going 3 wins, 1 draw and 4 loss.

So... QUESTION: Is this RW or is it players? No sense in rewinding the clock back to 2009 right now. 2012 saw RW's team earn 41.7% possible points in Sept and Oct, pretty much inline with his normal 37.6% over his coaching career... so lets focus on this year.

49.5% : Possible Pts Earned, 2012
52.0% : Possible Pts Earned, Mar-Aug 2012
41.7% : Possible Pts Earned, Sept-Oct 2012
-10.3 Pt dip in Sep/Oct (bad, but amusingly better than his atrocious career -18 pt dip)

Lets focus on the 49.5% for all of 2012. Below is how each player did as a starter in comparison to the team average (above or below 49.5, minimum of 5 starts):

+23.8 : Cole Grossman
+14.1 : Carlos Mendes
+13.8 : Federico Higuaín
+12.4 : Dilly Duka
+10.5 : Bernardo Anor

+06.1 : Danny O´Rourke
+03.4 : Jairo Obando Arrieta
+01.8 : Joshua Williams
+01.5 : Sebastián Miranda
+00.5 : Shaun Francis
+00.5 : Chris Birchall
+00.5 : Emilio Renteria
+00.0 : Eddie Gaven

+0.0 : Replacement (Team)

-01.6 : Andy Gruenebaum
-01.9 : Julius James
-02.3 : Chad Marshall
-05.1 : Nemanja Vukovic
-05.5 : Milovan Mirosevic

-10.1 : Justin Meram
-11.4 : Olman Vargas
-11.4 : Eric Gehrig
-16.2 : Tony Tchani

Some interesting players doing much better than the team average, but it doesn't really tell the full story. What you see up there is how it looks over the entire year. The Crew finished smack dab in the middle of the table; 10th out of 19 teams.

How about we tease out results versus the top half of the table only. The Crew's average against the Top Half of the table in 2012 is 39.2% Total Possible Pts Earned. Here is how Crew players did against that number (minimum of 5 starts).

+34.1 : Cole Grossman
+19.1 : Carlos Mendes
+19.1 : Dilly Duka
+14.1 : Nemanja Vukovic

+02.5 : Joshua Williams
+02.5 : Sebastián Miranda
+02.5 : Emilio Renteria
+00.8 : Justin Meram
+00.8 : Chris Birchall
+00.0 : Andy Gruenebaum
+00.0 : Eddie Gaven

+0.0 : TEAM

-01.1 : Tony Tchani
-01.1 : Danny O´Rourke
-08.9 : Chad Marshall
-09.2 : Jairo Obando Arrieta

-12.5 : Julius James
-24.4 : Milovan Mirosevic

Whoa... What happened to Milovan Mirosevic? And Federico Higuain didn't accumulate more than 5 starts against good teams??

There are some notable players whose possible points earned drastically changes versus good competition. Here is the difference if you compare the full season to how they did against top clubs.

+15.1 Tony Tchani
+10.9 Justin Meram
+10.3 Cole Grossman
+09.0 Nemanja Vukovic
+06.7 Dilly Duka
+05.0 Carlos Mendes
+02.0 Emilio Renteria
+01.6 Andy Gruenebaum
+01.0 Sebastián Miranda
+00.7 Joshua Williams
+00.3 Chris Birchall

+00.0 Eddie Gaven

-06.6 Chad Marshall
-07.2 Danny O´Rourke
-10.6 Julius James
-12.6 Jairo Obando Arrieta
-18.9 Milovan Mirosevic

Tchani, Meram, Grossman and Vukovic really stepped up against tougher opponents this year. When I ran these numbers I struck me as odd to see these players rising to the challenge of good opponents, so I decided to look at the percentage of games each player played against top clubs. For players with over 5 starts, it looks like this:

73% : Carlos Mendes
70% : Cole Grossman
67% : Justin Meram
65% : Dilly Duka
65% : Chris Birchall
59% : Jairo Obando Arrieta
58% : Julius James
57% : Tony Tchani
55% : Joshua Williams
53% : Andy Gruenebaum
52% : Eddie Gaven
50% : Sebastián Miranda
47% : Nemanja Vukovic
46% : Emilio Renteria
46% : Chad Marshall
43% : Danny O´Rourke
38% : Milovan Mirosevic

Again, Federico Higuain didn't even make five starts against the top half of the table. What the hell gives with Milovan Mirosevic?


Only 9 of Mirosevic's 26 starts were against good teams. He won a grand total of 0 in those 9. You read that right. ZERO. He was 0 wins, 4 draw, 5 loss when starting against top half teams. Not only did he face the least amount of tough talent, he had the worst record against them. It's terrible and might be the single most damning stat against any individual player.

To further drive home the point on Mirosevic against top talent; Here's how each player looks as far as average goal difference when "on the pitch" versus teams 1-9 on the table (this includes when a player comes on as a sub)...

Goal Diff : Name

+1.00 : Cole Grossman
+0.80 : Nemanja Vukovic
+0.63 : Dilly Duka
+0.50 : Carlos Mendes
+0.00 : Justin Meram

-0.06 : Joshua Williams
-0.06 : Sebastián Miranda
-0.10 : Chris Birchall
-0.14 : Tony Tchani
-0.14 : Danny O´Rourke

-0.17 : TEAM

-0.18 : Andy Gruenebaum
-0.24 : Eddie Gaven
-0.25 : Emilio Renteria
-0.40 : Jairo Obando Arrieta
-0.60 : Julius James
-0.64 : Chad Marshall
-0.67 : Milovan Mirosevic

The same players keep showing up at the top and at the bottom of these metrics but put weight in the Goal Difference when on the pitch against top half of the table, if you put any value in any of it.

The more and more I shape the numbers of the past year the more I see that too much trust might have been put in players like Mirosevic, Marshall, Higuain and even Arrieta. As talented as they may be, they couldn't get it done against better talent (or in Higuain's case, not even faced it).

My hope is that Brian Bliss puts value into results and not gut feelings. His "We’ve got to delve into this" comment is telling. Either he is playing dumb or he really doesn't know or track valuable metrics that will make his team better.

Look, I'm no expert on statistics... I'm an Ops Manager who does work in central Ohio. What I'm pulling here isn't over the top for anyone with access to Excel and a little bit of time.

I can see the missteps in the productivity of certain players but doesn't it come down to who puts these guys out there match in and match out? Of course the responsibility falls on the head coach but doesn't it also fall on the team administration? They are the folks that negotiate salary, which in turn results in playing time. Should they be held most accountable?

Does the Crew hold anyone accountable?

No comments: