Monday, December 07, 2009

MLS 2009 Plus/Minus Data

What is plus/minus? It's a hockey statistic which "...measures the team goal differential when a specific player is on the ice." It's also being used more and more in other sports, such as basketball. In soccer, we haven't seen too much of it yet. I know Dave from Sounder at Heart loves to talk about it, and it was also used on the MLS Cup broadcast. They mentioned that David Beckham had a plus/minus of +10, which was a figure that included the playoffs (he's +7 for the regular season).

I'm sure people would discuss it more, except that no one is keeping track of it (at least publicly). Well, I've put an end to that problem. I have gone through every game of the 2009 regular season and figured out the +/- for each player. In doing this, I've decided to only include the minutes played where both teams were at full strength, and only goals scored and allowed during 11 on 11 play are included below. All minutes and goals in a game after a red card has occurred aren't included. Doesn't matter whether it was 11v10, 10v10, 10v9, or 11v9. If you're interested in seeing how many minutes each team played and goals they scored in each of those situations, click here.

Why do it that way? I don't think it's fair to include the minutes where one team has a clear advantage. It's also been suggested that maybe the player who gets a red card should be punished somehow, possibly by counting the goals scored later against him. I don't think that's fair either; to me that's the realm of a player rating system, not plus/minus. I've also decided to count own goals just like any other goals. Even if they're the result of a mistake by one player, you still have to take into account the way that the attacking team got into the position to score.

Let's break it down by team. When looking at these results, obviously playing time is a huge factor. That's especially true for those teams with a large positive or negative GD. I'm not sure how useful of a stat this is, but it certainly is interesting to look at. It's probably better to compare players who play the same position, or have a similar amount of playing time (since the more you get, the closer you are to the team average).

I've also figured out the league leaders in plus/minus per 90 minutes, both in unadjusted and adjusted forms. That's at the bottom of this post.

Asterisks indicate that a player played for more than one team in 2009. The team's 11v11 goal differential is in parenthesis. Only regular season games are included below.


CHICAGO (-1)


11v11 Min GF GA
Plus/Minus Per 90
Daniel Woolard 409 8 3
5 1.100
Cuauhtemoc Blanco 1308 21 17
4 0.275
Mike Banner 893 13 9
4 0.403
Justin Mapp 941 12 10
2 0.191
Baggio Husidic 420 3 1
2 0.429
Tim Ward 1469 19 18
1 0.061
Bakary Soumare 1198 17 16
1 0.075
Dasan Robinson 903 11 10
1 0.100
Calen Carr 84 1

1 1.071
Wilman Conde 1400 20 20
0 0.000
Jon Busch 2460 31 32
-1 -0.037
Patrick Nyarko 1777 20 21
-1 -0.051
Chris Rolfe 1711 22 23
-1 -0.053
Gonzalo Segares 768 11 12
-1 -0.117
Marco Pappa 2145 30 32
-2 -0.084
Brian McBride 1765 22 24
-2 -0.102
John Thorrington 1485 16 18
-2 -0.121
Peter Lowry 654 9 11
-2 -0.275
Logan Pause 2083 25 28
-3 -0.130
Stefan Dimitrov 119
3
-3 -2.269
Brandon Prideaux 1308 15 19
-4 -0.275
C.J. Brown 1412 14 19
-5 -0.319
Austin Washington 349 1 6
-5 -1.289

To no one's surprise, Blanco comes off looking pretty good here. The Fire was +4 when he was on the field, and -5 when he wasn't.


CHIVAS USA (+3)


11v11 Min GF GA
Plus/Minus Per 90
Mariano Trujillo 1395 21 11
10 0.645
Justin Braun 1127 19 10
9 0.719
Zach Thornton 2124 29 21
8 0.339
Carey Talley 2179 31 24
7 0.289
Eduardo Lillingston 1191 17 11
6 0.453
Jesse Marsch 1613 23 18
5 0.279
Atiba Harris*** 880 12 7
5 0.511
Michael Lahoud 765 12 7
5 0.588
Sacha Kljestan 1829 24 21
3 0.148
Shavar Thomas 1270 17 14
3 0.213
Jonathan Bornstein 2028 27 25
2 0.089
Yamith Cuesta 733 10 8
2 0.246
Jorge Flores 302 3 2
1 0.298
Paulo Nagamura 2039 25 25
0 0.000
Maykel Galindo 1322 17 17
0 0.000
Gerson Mayen 361 6 6
0 0.000
Jim Curtin 127 2 2
0 0.000
Claudio Suarez 90


0 0.000
Bobby Burling*** 106
1
-1 -0.849
Jesus Padilla 686 6 8
-2 -0.262
Maicon Santos 634 8 10
-2 -0.284
Bojan Stepanovic 632 7 9
-2 -0.285
Chukwudi Chijindu 350 2 4
-2 -0.514
Alecko Eskandarian*** 177 1 3
-2 -1.017
Ante Jazic 1001 13 17
-4 -0.360
Marcelo Saragosa*** 603 6 10
-4 -0.597
Lance Parker 262 2 7
-5 -1.718
Sasha Victorine 420 1 10
-9 -1.929

Now here's a team with some extreme swings in the numbers, thanks to the great start. I'm not sure any player in this entire post comes off looking worse than Victorine. A -9 in 420 minutes of action, for a team that was +3 overall. And maybe Preki should've played the Mexicans Lillingston and Trujillo when it counted most.


COLORADO (-3)


11v11 Min GF GA
Plus/Minus Per 90
Jacob Peterson 1030 15 10
5 0.437
Jamie Smith 165 5 1
4 2.182
Facundo Diz 77 4

4 4.675
Matt Pickens 1539 25 22
3 0.175
Terry Cooke 373 6 3
3 0.724
Colin Clark 1288 21 19
2 0.140
Michael Holody 66 3 1
2 2.727
Ugo Ihemelu*** 1586 24 23
1 0.057
Gregory Richardson 11


0 0.000
Ross Schunk 4


0 0.000
Scott Palguta 1602 22 23
-1 -0.056
Cory Gibbs 1454 23 25
-2 -0.124
Ty Harden 72
2
-2 -2.500
Jordan Harvey 2380 32 35
-3 -0.113
Omar Cummings 2258 33 36
-3 -0.120
Nick LaBrocca 2201 31 34
-3 -0.123
Pablo Mastroeni 2021 29 32
-3 -0.134
Julien Baudet 769 9 12
-3 -0.351
Pat Noonan*** 757 9 12
-3 -0.357
Conor Casey 1934 28 32
-4 -0.186
Drew Moor*** 675 7 11
-4 -0.533
Kosuke Kimura 1724 23 28
-5 -0.261
Preston Burpo 928 10 16
-6 -0.582
Greg Dalby 232 3 9
-6 -2.328
Mehdi Ballouchy 1991 23 32
-9 -0.407

No Rapids fan will be shocked to see Ballouchy with the worst +/- on the team, -9. It's highly amusing to see Diz at +4 in only 77 minutes, through little or no contribution of his own. Too bad many of the top players here got injured (Peterson, Pickens, Clark) or they probably would've made the playoffs.


COLUMBUS (+11)


11v11 Min GF GA
Plus/Minus Per 90
Eric Brunner 1952 31 17
14 0.645
Danny O'Rourke 2209 36 25
11 0.448
Steven Lenhart 1020 17 7
10 0.882
Gino Padula 2213 33 24
9 0.366
Emmanuel Ekpo 1839 30 21
9 0.440
Eddie Gaven 2276 34 26
8 0.316
Brian Carroll 2184 34 26
8 0.330
William Hesmer 1628 25 18
7 0.387
Chad Marshall 1496 24 17
7 0.421
Robbie Rogers 1624 24 18
6 0.333
Alejandro Moreno 1520 24 18
6 0.355
Jed Zayner 1355 19 14
5 0.332
Frankie Hejduk 1242 18 13
5 0.362
Andy Gruenebaum 826 14 9
5 0.545
Adam Moffat 503 8 4
4 0.716
Emilio Renteria 167 3

3 1.617
G. Barros Schelotto 1706 23 21
2 0.106
Kevin Burns 340 3 2
1 0.265
Duncan Oughton 194 5 4
1 0.464
Pat Noonan*** 166 3 2
1 0.542
Alex Grendi 127 2 1
1 0.709
Jason Garey 691 10 10
0 0.000
Andy Iro 599 9 9
0 0.000
Kenny Schoeni 90
1
-1 -1.000
Cory Elenio 17
1
-1 -5.294

Schelotto has the worst +/- of all the Crew starters. They did play pretty well without him, remember that stretch in the middle of the season? Also, compare Lenhart and Moreno. Did Warzycha make the right moves in game one against RSL?


DALLAS (+1)


11v11 Min GF GA
Plus/Minus Per 90
George John 1142 26 15
11 0.867
Ugo Ihemelu*** 686 19 11
8 1.050
Dario Sala 1109 25 18
7 0.568
Jair Benitez 1079 26 19
7 0.584
Daniel Hernandez 525 13 6
7 1.200
Jeff Cunningham 1823 38 32
6 0.296
Brek Shea 656 16 10
6 0.823
Atiba Harris*** 977 23 18
5 0.461
Dax McCarty 2240 42 39
3 0.121
Heath Pearce 477 9 6
3 0.566
Eric Avila 265 7 5
2 0.679
Marvin Chavez 233 6 4
2 0.773
Alvaro Sanchez 141 3 1
2 1.277
David Ferreira 2472 44 43
1 0.036
Pablo Ricchetti 1738 34 33
1 0.052
Kyle Davies 762 12 12
0 0.000
Anthony Wallace 442 7 7
0 0.000
Peri Marosevic 47


0 0.000
Michael Dello-Russo 162 1 3
-2 -1.111
Bruno Guarda 139 1 3
-2 -1.295
Aaron Pitchkolan*** 45
2
-2 -4.000
Dave van den Bergh 2289 35 39
-4 -0.157
Marcelo Saragosa*** 551 6 10
-4 -0.653
Kenny Cooper 1190 16 21
-5 -0.378
Blake Wagner 699 8 13
-5 -0.644
Steve Purdy 338 2 7
-5 -1.331
Ray Burse 1363 19 25
-6 -0.396
Daniel Torres 722 9 16
-7 -0.873
Andre Rocha 1163 15 23
-8 -0.619
Drew Moor*** 1719 22 32
-10 -0.524

You can easily tell from this list which players featured in the first and second halves of the season. When you combine his minutes with Chivas and Dallas, Harris is at +10 in 1,857 minutes.


DC UNITED (-1)


11v11 Min GF GA
Plus/Minus Per 90
Fred 1341 23 18
5 0.336
Luciano Emilio 2001 33 30
3 0.135
Jaime Moreno 1170 22 20
2 0.154
Thabiso Khumalo 357 7 5
2 0.504
Steve Cronin 172 3 1
2 1.047
Christian Gomez 1612 25 24
1 0.056
Louis Crayton 521 9 8
1 0.173
Ange N'Silu 452 8 7
1 0.199
Devon McTavish 427 6 5
1 0.211
Clyde Simms 2368 36 36
0 0.000
Bryan Namoff 2270 38 38
0 0.000
Marc Burch 2114 36 36
0 0.000
Chris Pontius 2035 37 37
0 0.000
Greg Janicki 411 7 7
0 0.000
Brandon Barklage 98 3 3
0 0.000
Danny Szetela 90 1 1
0 0.000
Dejan Jakovic 2018 33 34
-1 -0.045
Josh Wicks 1635 25 26
-1 -0.055
Ben Olsen 1339 19 20
-1 -0.067
Julius James*** 640 8 9
-1 -0.141
John DiRaimondo 45
1
-1 -2.000
Francis Doe 40
1
-1 -2.250
Avery John 344 5 7
-2 -0.523
Rodney Wallace 2201 33 36
-3 -0.123
Milos Kocic 283 5 8
-3 -0.954
Lawson Vaughn 193 1 4
-3 -1.399
Santino Quaranta 1711 26 32
-6 -0.316
Andrew Jacobson 834 13 19
-6 -0.647

Fred and Emilio come out on top of the DC list. Many United fans seem ready to dump both of them and move on.


HOUSTON (+10)


11v11 Min GF GA
Plus/Minus Per 90
Bobby Boswell 2248 34 22
12 0.480
Geoff Cameron 2290 34 23
11 0.432
Stuart Holden 2020 32 21
11 0.490
Andrew Hainault 1425 25 14
11 0.695
Pat Onstad 2380 35 25
10 0.378
Brad Davis 1934 26 17
9 0.419
Brian Ching 1390 22 13
9 0.583
Kei Kamara*** 1605 20 12
8 0.449
Ricardo Clark 1804 25 19
6 0.299
Brian Mullan 1713 23 18
5 0.263
Richard Mulrooney 1197 14 10
4 0.301
Craig Waibel 669 10 6
4 0.538
Dominic Oduro*** 667 13 9
4 0.540
Luis Angel Landin 278 4 2
2 0.647
Cam Weaver*** 263 5 3
2 0.684
Ryan Cochrane*** 245 6 4
2 0.735
Corey Ashe 1124 16 15
1 0.080
Wade Barrett 880 12 11
1 0.102
Mike Chabala 843 13 12
1 0.107
Julius James*** 490 9 9
0 0.000
Ade Akinbiyi 229 2 2
0 0.000
Danny Cruz 107 1 1
0 0.000
Erik Ustruck 40


0 0.000
Chris Wondolowski*** 264 3 4
-1 -0.341
Abe Thompson*** 46 1 2
-1 -1.957
Eddie Robinson 30
1
-1 -3.000

Hard to look bad playing for Houston, though some are barely in the positive like Ashe/Barrett/Chabala.


KANSAS CITY (-10)


11v11 Min GF GA
Plus/Minus Per 90
Michael Kraus 98 3 1
2 1.837
Zoltan Hercegfalvi 220 5 4
1 0.409
Chance Myers 98 2 2
0 0.000
Kevin Souter 59


0 0.000
Kurt Morsink 141 2 3
-1 -0.638
Abe Thompson*** 107
1
-1 -0.841
Matt Marquess 85 1 2
-1 -1.059
Roger Espinoza 566 6 8
-2 -0.318
Jimmy Conrad 1663 21 24
-3 -0.162
Kei Kamara*** 351 3 6
-3 -0.769
Jonathan Leathers 778 8 12
-4 -0.463
Graham Zusi 445 6 10
-4 -0.809
Rauwshan McKenzie 406 1 5
-4 -0.887
Josh Wolff 1926 22 28
-6 -0.280
Santiago Hirsig 1655 17 23
-6 -0.326
Michael Harrington 1365 15 21
-6 -0.396
Adam Cristman 201
6
-6 -2.687
Claudio Lopez 2021 23 30
-7 -0.312
Herculez Gomez 1274 14 21
-7 -0.495
Lance Watson 1244 10 17
-7 -0.506
Matt Besler 1771 19 27
-8 -0.407
Aaron Hohlbein 1540 13 21
-8 -0.468
Davy Arnaud 1705 19 28
-9 -0.475
Kevin Hartman 2171 23 33
-10 -0.415
Jack Jewsbury 1991 20 30
-10 -0.452

Conrad (-3) stands out as the best of the Wizards' regulars. Meanwhile, ouch: Cristman is at -6 in only 201 minutes.


LOS ANGELES (+8)


11v11 Min GF GA
Plus/Minus Per 90
Omar Gonzalez 2354 34 26
8 0.306
Todd Dunivant 1901 28 20
8 0.379
Landon Donovan 1887 29 21
8 0.382
Eddie Lewis 1690 21 13
8 0.426
Gregg Berhalter 2053 27 20
7 0.307
A.J. DeLaGarza 1523 21 14
7 0.414
David Beckham 875 15 8
7 0.720
Donovan Ricketts 1987 27 22
5 0.226
Stefani Miglioranzi 1609 22 17
5 0.280
Edson Buddle 1177 18 13
5 0.382
Alecko Eskandarian*** 213 5

5 2.113
Sean Franklin 1043 14 10
4 0.345
Alan Gordon 975 16 12
4 0.369
Jovan Kirovski 1060 13 10
3 0.255
Chris Birchall 521 6 3
3 0.518
Josh Saunders 395 7 4
3 0.684
Josh Tudela 481 8 6
2 0.374
Leonard Griffin 16 1

1 5.625
Dema Kovalenko 1082 13 13
0 0.000
Chris Klein 961 15 15
0 0.000
Bryan Jordan 377 5 5
0 0.000
Kyle Patterson 124 4 4
0 0.000
Tristan Bowen 1


0 0.000
Mike Magee 1373 17 18
-1 -0.066
Tony Sanneh 524 8 12
-4 -0.687

Looking at this makes me wonder how a healthy Eskandarian would've changed the Galaxy's season. Magee, Klein, and Kovalenko don't come off looking good, while Sanneh was a liability on the field.


NEW ENGLAND (-1)


11v11 Min GF GA
Plus/Minus Per 90
Kevin Alston 2087 27 21
6 0.259
Matt Reis 1975 25 20
5 0.228
Emmanuel Osei 1750 23 19
4 0.206
Steve Ralston 1348 20 16
4 0.267
Taylor Twellman 109 5 1
4 3.303
Shalrie Joseph 2072 30 28
2 0.087
Kenny Mansally 1562 20 18
2 0.115
Mauricio Castro 203 3 1
2 0.887
Jeff Larentowicz 2242 27 26
1 0.040
Chris Albright 90 2 1
1 1.000
Nico Colaluca 105 1 1
0 0.000
Stephane Assengue 37


0 0.000
Darrius Barnes 2448 30 31
-1 -0.037
Kheli Dube 1608 16 17
-1 -0.056
Pat Phelan 973 13 14
-1 -0.092
Jay Heaps 1959 26 28
-2 -0.092
Edgaras Jankauskas 567 3 5
-2 -0.317
Sainey Nyassi 1740 19 23
-4 -0.207
Wells Thompson 1523 19 24
-5 -0.295
Chris Tierney 1044 10 16
-6 -0.517
Brad Knighton 473 5 11
-6 -1.142
Michael Videira 418 2 8
-6 -1.292
Amaechi Igwe 596 4 12
-8 -1.208

Playing in that 0-6 demolition at RSL certainly affected some of the bottom players here. But it didn't stop Joseph from finishing in the positive.


NEW YORK (-15)


11v11 Min GF GA
Plus/Minus Per 90
Walter Garcia 81 4

4 4.444
Carlos Johnson 871 11 8
3 0.310
Bouna Coundoul 675 8 7
1 0.133
Khano Smith 449 5 4
1 0.200
Alec Dufty 70


0 0.000
Matthew Mbuta 192 2 4
-2 -0.938
Leo Krupnik 135
2
-2 -1.333
Dominic Oduro*** 58
2
-2 -3.103
Danleigh Borman 1221 12 15
-3 -0.221
Macoumba Kandji 1577 21 25
-4 -0.228
Andrew Boyens 1072 11 15
-4 -0.336
Ernst Oebster 157
4
-4 -2.293
John Wolyniec 710 5 10
-5 -0.634
Juan Pietravallo 174
5
-5 -2.586
Albert Celades 1146 16 22
-6 -0.471
Sinisa Ubiparipovic 1103 14 20
-6 -0.490
Alfredo Pacheco 976 12 18
-6 -0.553
Carlos Mendes 693 5 11
-6 -0.779
Danny Cepero 982 11 19
-8 -0.733
Jon Conway 643 6 14
-8 -1.120
Juan Pablo Angel 1858 23 32
-9 -0.436
Kevin Goldthwaite 1206 14 23
-9 -0.672
Seth Stammler 2023 22 32
-10 -0.445
Dane Richards 1758 18 28
-10 -0.512
Jorge Rojas 1321 12 23
-11 -0.749
Jeremy Hall 1823 20 32
-12 -0.592
Mike Petke 1381 13 27
-14 -0.912
Luke Sassano 989 7 21
-14 -1.274
Nick Zimmerman 726 3 17
-14 -1.736

Yes, Khano Smith was in the positive. That's not a mistake. There's still "hard man" Pietravallo to laugh at, though. Zimmerman, just like his fellow Philadelphia Union expansion draft pick Jacobson for DC, finished last on his team.


REAL SALT LAKE (+10)


11v11 Min GF GA
Plus/Minus Per 90
Jamison Olave 1890 35 20
15 0.714
Nick Rimando 2070 38 26
12 0.522
Nat Borchers 2367 40 29
11 0.418
Andy Williams 736 19 8
11 1.345
Will Johnson 2054 36 26
10 0.438
Tony Beltran 1728 30 20
10 0.521
Chris Wingert 2169 40 31
9 0.373
Fabian Espindola 1377 24 15
9 0.588
Kyle Beckerman 1928 31 24
7 0.327
Clint Mathis 1767 28 21
7 0.357
Yura Movsisyan 1548 25 18
7 0.407
Ned Grabavoy 701 16 10
6 0.770
Robbie Findley 1559 29 25
4 0.231
Javier Morales 1833 25 22
3 0.147
Luis Miguel Escalada 94 4 1
3 2.872
Raphael Cox 177 3 1
2 1.017
Ian Joy 7


0 0.000
Kyle Reynish 6


0 0.000
Tino Nunez 1


0 0.000
Jean Alexandre 54
1
-1 -1.667
Chris Seitz 312 3 5
-2 -0.577
Rachid El Khalifi 165 2 4
-2 -1.091
Pablo Campos*** 353 5 8
-3 -0.765
David Horst 86 1 4
-3 -3.140
Robbie Russell 1286 17 22
-5 -0.350

Russell looks awful compared to the rest of the regulars. Nobody will remember that now, though. But how do you play that much time on a +10 team and up -5? On the other hand, Williams really came on at the end of the season, didn't he? That's an amazing number for that amount of playing time.


SAN JOSE (-15)


11v11 Min GF GA
Plus/Minus Per 90
Andrew Weber 180 4 3
1 0.500
Ryan Cochrane*** 171 4 3
1 0.526
Davide Somma 4


0 0.000
Antonio Ribeiro 496 5 6
-1 -0.181
Ramon Sanchez 838 11 13
-2 -0.215
Chris Wondolowski*** 823 10 12
-2 -0.219
Quincy Amarikwa 587 9 11
-2 -0.307
Cornell Glen 436 6 8
-2 -0.413
Andre Luiz 431 4 6
-2 -0.418
Bobby Burling*** 562 6 9
-3 -0.480
Michael Zaher 744 12 16
-4 -0.484
Pablo Campos*** 601 7 11
-4 -0.599
Brandon McDonald 1509 20 25
-5 -0.298
Aaron Pitchkolan*** 573 8 13
-5 -0.785
Jason Hernandez 1333 17 24
-7 -0.473
Simon Elliott 1225 15 22
-7 -0.514
Darren Huckerby 1001 12 19
-7 -0.629
Cam Weaver*** 471 5 12
-7 -1.338
Ramiro Corrales 2057 26 34
-8 -0.350
Shea Salinas 1236 13 21
-8 -0.583
Ryan Johnson 2197 32 41
-9 -0.369
Nick Garcia*** 1071 14 23
-9 -0.756
Kelly Gray 632 5 14
-9 -1.282
Eric Denton 794 8 18
-10 -1.134
Bobby Convey 2134 27 38
-11 -0.464
Chris Leitch 2458 31 44
-13 -0.476
Arturo Alvarez 1826 23 37
-14 -0.690
Joe Cannon 2441 29 45
-16 -0.590

Nobody ended up looking good here, it seems. McDonald leads the starters. Not beginning the season with Denton and Gray in the back line would've helped.


SEATTLE (11)


11v11 Min GF GA
Plus/Minus Per 90
Kasey Keller 2286 35 22
13 0.512
James Riley 2100 31 19
12 0.514
Jhon Kennedy Hurtado 2137 32 21
11 0.463
Tyrone Marshall 1898 31 20
11 0.522
Brad Evans 2006 30 20
10 0.449
Freddie Ljungberg 1691 27 17
10 0.532
Fredy Montero 2013 29 21
8 0.358
Nate Jaqua 1998 30 22
8 0.360
Osvaldo Alonso 1886 26 18
8 0.382
Sebastien Le Toux 1300 21 13
8 0.554
Zach Scott 408 8 2
6 1.324
Steve Zakuani 1805 23 18
5 0.249
Nathan Sturgis 368 6 3
3 0.734
Roger Levesque 259 5 2
3 1.042
Sanna Nyassi 247 4 1
3 1.093
Patrick Ianni 933 11 9
2 0.193
Tyson Wahl 797 12 10
2 0.226
Leo Gonzalez 887 10 9
1 0.101
Stephen King 386 6 6
0 0.000
Ben Dragavon 2


0 0.000
Peter Vagenas 640 8 9
-1 -0.141
Chris Eylander 90
2
-2 -2.000

Just imagine what Zakuani's number would be if he finished all the good chances he had.


TORONTO (-8)


11v11 Min GF GA
Plus/Minus Per 90
Nick Garcia*** 1085 17 16
1 0.083
Johann Smith 47 1

1 1.915
Danny Dichio 746 11 11
0 0.000
Amadou Sanyang 297 2 2
0 0.000
Fuad Ibrahim 204 3 3
0 0.000
Lesly Fellinga 137 3 3
0 0.000
Gabe Gala 3


0 0.000
Greg Sutton 135 1 2
-1 -0.667
Stefan Frei 2241 34 37
-3 -0.120
Chad Barrett 2055 29 32
-3 -0.131
Marvell Wynne 1836 28 31
-3 -0.147
Ali Gerba 494 6 9
-3 -0.547
O'Brian White 440 6 9
-3 -0.614
Carl Robinson 1717 25 29
-4 -0.210
Nana Attakora 1638 22 26
-4 -0.220
Kevin Harmse 805 13 17
-4 -0.447
Emmanuel Gomez 487 3 7
-4 -0.739
Rohan Ricketts 316 4 8
-4 -1.139
Brian Edwards 270 2 6
-4 -1.333
Amado Guevara 1999 30 35
-5 -0.225
Pablo Vitti 1610 22 27
-5 -0.280
Marco Velez 780 10 15
-5 -0.577
Sam Cronin 2309 31 37
-6 -0.234
Julian de Guzman 450 4 10
-6 -1.200
Jim Brennan 2428 34 41
-7 -0.259
Adrian Serioux 2187 31 38
-7 -0.288
Dwayne De Rosario 2390 35 44
-9 -0.339

Now I know Garcia was a flop for TFC. But the numbers don't lie, they actually had a positive GD with him out there. Of course, his secret was to play so bad that he didn't start in the final game in NY and have those five goals count against him.


Unadjusted League Leaders Per 90 Minutes

(1,000 minutes minimum)

Obviously these will be filled with players from the best and worst teams, but let's see the results anyway.

Top 25




Per 90
11v11 Min Plus/Minus
1 Steven Lenhart CLB 0.882
1020 10
2 George John DAL 0.867
1142 11
3 Justin Braun CHV 0.719
1127 9
4 Jamison Olave RSL 0.714
1890 15
5 Andrew Hainault HOU 0.695
1425 11
6 Eric Brunner CLB 0.645
1952 14
7 Mariano Trujillo CHV 0.645
1395 10
8 Fabian Espindola RSL 0.588
1377 9
9 Jair Benitez DAL 0.584
1079 7
10 Brian Ching HOU 0.583
1390 9
11 Dario Sala DAL 0.568
1109 7
12 Sebastien Le Toux SEA 0.554
1300 8
13 Freddie Ljungberg SEA 0.532
1691 10
14 Nick Rimando RSL 0.522
2070 12
15 Tyrone Marshall SEA 0.522
1898 11
16 Tony Beltran RSL 0.521
1728 10
17 James Riley SEA 0.514
2100 12
18 Kasey Keller SEA 0.512
2286 13
19 Stuart Holden HOU 0.490
2020 11
20 Atiba Harris CHV/DAL 0.485
1857 10
21 Bobby Boswell HOU 0.480
2248 12
22 Jhon Kennedy Hurtado SEA 0.463
2137 11
23 Eduardo Lillingston CHV 0.453
1191 6
24 Brad Evans SEA 0.449
2006 10
25 Danny O'Rourke CLB 0.448
2209 11


Bottom 25




Per 90
11v11 Min Plus/Minus
1 Mike Petke NY -0.912
1381 -14
2 Jorge Rojas NY -0.749
1321 -11
3 Arturo Alvarez SJ -0.690
1826 -14
4 Kevin Goldthwaite NY -0.672
1206 -9
5 Darren Huckerby SJ -0.629
1001 -7
6 Marcelo Saragosa DAL/CHV -0.624
1154 -8
7 Andre Rocha DAL -0.619
1163 -8
8 Jeremy Hall NY -0.592
1823 -12
9 Joe Cannon SJ -0.590
2441 -16
10 Shea Salinas SJ -0.583
1236 -8
11 Drew Moor DAL/COL -0.526
2394 -14
12 Chris Tierney NE -0.517
1044 -6
13 Simon Elliott SJ -0.514
1225 -7
14 Dane Richards NY -0.512
1758 -10
15 Lance Watson KC -0.506
1244 -7
16 Herculez Gomez KC -0.495
1274 -7
17 Sinisa Ubiparipovic NY -0.490
1103 -6
18 Chris Leitch SJ -0.476
2458 -13
19 Davy Arnaud KC -0.475
1705 -9
20 Jason Hernandez SJ -0.473
1333 -7
21 Albert Celades NY -0.471
1146 -6
22 Aaron Hohlbein KC -0.468
1540 -8
23 Bobby Convey SJ -0.464
2134 -11
24 Jack Jewsbury KC -0.452
1991 -10
25 Seth Stammler NY -0.445
2023 -10

Steven Lenhart & Mike Petke are first and last for 2009.


Now, what if I adjust the per 90 stat by comparing it to the plus/minus for each team? For example, the Crew played 2,544 full strength minutes and had a +11 differential. So the team's per 90 stat is 0.389. If you subtract that from Lenhart's 0.882, he gets an adjusted total of 0.493. This allows us to see who really stood out compared to the team's overall performance.


Adjusted League Leaders Per 90 Minutes

(1,000 minutes minimum)

See explanation in the previous paragraph.

In order to get adjusted per 90s for the eight qualifying players who played for two teams, I first got their adjusted totals with each team. Then I multiplied the adjusted per 90s by the minutes played for each of two teams, added them together, then divided by the number of total minutes. So for example, Ugo Ihemelu had the following adjusted totals:

DAL: 1.013 in 686 minutes
COL: 0.166 in 1,586 minutes

(1.013*686) + (0.166*1,586) รท (686 + 1,586) = 0.422

That explains that.


Top 25




Adjusted Per 90
Per 90 Team Per 90
1 George John DAL 0.830
0.867 0.036
2 Justin Braun CHV 0.606
0.719 0.113
3 Jair Benitez DAL 0.547
0.584 0.036
4 Jacob Peterson COL 0.546
0.437 -0.109
5 Mariano Trujillo CHV 0.532
0.645 0.113
6 Dario Sala DAL 0.532
0.568 0.036
7 Steven Lenhart CLB 0.493
0.882 0.389
8 Ugo Ihemelu COL/DAL 0.422
0.357
9 Atiba Harris CHV/DAL 0.412
0.485
10 Fred DC 0.370
0.336 -0.034
11 Danleigh Borman NY 0.348
-0.221 -0.570
12 Macoumba Kandji NY 0.341
-0.228 -0.570
13 Eduardo Lillingston CHV 0.340
0.453 0.113
14 Jamison Olave RSL 0.337
0.714 0.377
15 Andrew Hainault HOU 0.317
0.695 0.378
16 Cuauhtemoc Blanco CHI 0.312
0.275 -0.037
17 Steve Ralston NE 0.304
0.267 -0.037
18 Kevin Alston NE 0.296
0.259 -0.037
19 Matt Pickens COL 0.285
0.175 -0.109
20 Matt Reis NE 0.265
0.228 -0.037
21 Jeff Cunningham DAL 0.260
0.296 0.036
22 Eric Brunner CLB 0.256
0.645 0.389
23 Jimmy Conrad KC 0.252
-0.162 -0.415
24 Colin Clark COL 0.249
0.140 -0.109
25 Emmanuel Osei NE 0.242
0.206 -0.037


Bottom 25




Adjusted Per 90
Per 90 Team Per 90
1 Robbie Russell RSL -0.727
-0.350 0.377
2 Marcelo Saragosa DAL/CHV -0.700
-0.624
3 Andre Rocha DAL -0.655
-0.619 0.036
4 Drew Moor DAL/COL -0.522
-0.526
5 Chris Tierney NE -0.480
-0.517 -0.037
6 Ante Jazic CHV -0.473
-0.360 0.113
7 Ray Burse DAL -0.433
-0.396 0.036
8 Kenny Cooper DAL -0.415
-0.378 0.036
9 Mike Magee LA -0.368
-0.066 0.302
10 Mike Petke NY -0.343
-0.912 -0.570
11 Dema Kovalenko LA -0.302
0.000 0.302
12 Corey Ashe HOU -0.298
0.080 0.378
13 Mehdi Ballouchy COL -0.297
-0.407 -0.109
14 G. Barros Schelotto CLB -0.284
0.106 0.389
15 C.J. Brown CHI -0.282
-0.319 -0.037
16 Santino Quaranta DC -0.281
-0.316 -0.034
17 Wells Thompson NE -0.259
-0.295 -0.037
18 Brandon Prideaux CHI -0.239
-0.275 -0.037
19 Javier Morales RSL -0.230
0.147 0.377
20 Julius James HOU/DC -0.224
-0.080
21 Dave van den Bergh DAL -0.194
-0.157 0.036
22 Jorge Rojas NY -0.180
-0.749 -0.570
23 Arturo Alvarez SJ -0.175
-0.690 -0.515
24 Sainey Nyassi NE -0.170
-0.207 -0.037
25 Steve Zakuani SEA -0.167
0.249 0.417

Compared to the rest of their teammates, George John and Robbie Russell come out looking best and worst. I think it's exciting to look at this in adjusted form, because you get results like Danleigh Borman and Mac Kandji at 11th and 12th in the top 25. They may have had negative numbers themselves, but they were much better than the average Red Bull.

Labels:

Comments on "MLS 2009 Plus/Minus Data"

 

Anonymous Anonymous said ... (1:15 PM, December 07, 2009) : 

fantastic stat. very interesting. i would almost like to see what position the players play, or goals scored by player as well. As some players (strikers for instance) directly affect the numbers of the team and others don't as much. that addition could be even more telling of a teams abiltities.

one edit: You noted that "Steven Lenhart & Mike Petke are first and last for 2009." Mike Petke is -14 and joe Canon is -16.

 

Blogger scaryice said ... (4:09 PM, December 07, 2009) : 

I was referring to the per 90 stat.

 

Anonymous Anonymous said ... (4:11 PM, December 07, 2009) : 

Great info. Very interesting and, as a fan of RBNY, pretty grim reading.

 

Anonymous Eric in Baltimore said ... (5:21 PM, December 07, 2009) : 

I don't think this stat translates from hockey to soccer. Hockey is played in a much more constrained space with all 6 players combining on both offense and defense. In that context, one player has the ability to impact the +/- figure in a much more significant way.

In soccer, a defender has his +/- rating diminished by attacking players who can't find the net and goalkeeping mistakes. This stat is more indicative of a team's strengths and weaknesses rather than those of an individual.

It's a nice compilation of data, but ultimately meaningless.

 

Anonymous Anonymous said ... (9:49 PM, December 07, 2009) : 

It is not meaningless Eric, and why do Hockey dudes always make that comment? +/- is used in basketball and baseball as well.

 

Anonymous Eric in Baltimore said ... (6:55 AM, December 08, 2009) : 

I'm not a "hockey dude" as you put it. It makes sense in basketball, but not in baseball. As Americans, we are generally far too eager to invent statistical models to prove something or other.

 

Anonymous Jason said ... (1:09 PM, December 08, 2009) : 

Eric, I'm not sure how you can think that these stats don't translate from hockey to soccer. You are correct that each player has less ability to impact the +/- in soccer, but that is because there are almost twice as many players. This likely increases the cutoff of where the divergence becomes statistically significant, but above that level it is absolutely meaningful.

As for an attacking player that can't find goal or a goalkeeping mistake, I'm pretty sure that these things happen in hockey as well and you do not seem to be condemning the significance of that data.

Especially for the adjusted top and bottom 25 (which controls for the most obvious variable), unless there are statistical anomalies in the data, these results should be well above the confidence interval necessary to claim statistical significance.

 

Blogger Jason Wintz said ... (11:46 PM, December 08, 2009) : 

scaryice, this is a great stat. It would be interesting to put out a list comparing them to team average. For example, Houston as a whole had a +10 over 30 games, or a +.333 per 90 min (obviously I'm not accounting for red cards here), so a player whose +/- is ~.333 like Mulrooney is merely average for the team, whereas a player whose +/- is positive but below team average is actually dragging the team down, like Barrett or Chabala. Hmmm... interesting that both left backs have scores that are similarly below team average. Wonder what that means? When you've done all that work then you can get crazy and analyze pairings that way, such as forward pairings, central defense pairings, wide midfield/fullback pairings, etc; then you can go on to do trios, etc till you get full 11 lineups. Good thing MLS has such a long offseason :)

Eric, I don't understand why you say this stat doesn't translate. That's like saying that a team's chosen lineup doesn't particularly matter, which is of course ridiculous. Of course, the numbers can always be skewed by particular events (such as Garcia's absence in the one game where TFC allows 5 goals), but that does not render the whole statistic any more or less valid than it would be in hockey.

 

Anonymous Sancho said ... (6:52 AM, December 09, 2009) : 

I think the biggest problem here is the low-scoring characteristic of football association.

Maybe it would be better to measure it using attacking attempts, corners and/or time of possession than goals.

 

Anonymous Sancho said ... (6:54 AM, December 09, 2009) : 

...using attacking attempts, corners and/or time of possession than goals ALONE.

 

Anonymous Chest Rockwell said ... (3:13 PM, December 09, 2009) : 

I'm going to agree with Eric on this one, at least to a certain degree.

Look at Fred, for example. He routinely blew easy chances, frequently gave the ball away with good options available, and could be relied upon to kill DC attacks by dribbling into crowds of defenders. On the other side of the ball, he does work hard, but gives up way too many free kicks.

This stat doesn't point to any of that. Instead, it would tell the uninformed that Fred is more useful than Quaranta, who works just as hard but is a smarter player and less wasteful with the ball.

I'm sure you could go through here and find loads of similar examples. I see DeRosario, who spent much of the year carrying TFC on his back, has one of the worst ratings.

What I'm trying to say is that soccer, with 11 players, is a sport where you can hide a bad player in a good team (a key in MLS, with the low salary cap) or be a good player stuck on a bad team.

Soccer is structurally different than hockey, which has subs every 45 seconds, features fewer players on at a given time, and generally sees more goals scored. I think +/- will paint a more accurate picture of who's doing well on a hockey team than in soccer. I'm not saying it's entirely useless, but the potential for a good player to be poorly rated and/or a poor player to get a high positive rating is much higher.

 

Blogger Andrew said ... (6:06 AM, December 10, 2009) : 

There's one other thing to take into account: in soccer, combinations of players can be much more important than individuals. This is most obviously true in the back line. 90% of playing defense at professional level is coordinating with teammates. I'd take almost any starting back four in MLS over four world-class defenders who had never played as teammates.

 

Anonymous Howard Hamilton said ... (11:47 PM, March 09, 2010) : 

I think the plus/minus rating would be extremely useful in soccer, and it's clear that you've put in a lot of effort to generate these numbers. That being said, I do have some objections to your approach.

My first objection is to your use of solely 11v11 situations. While it definitely makes one's life easier, you're missing the 5-10% of match situations where there is a man advantage. I am aware that hockey's plus/minus does not consider goals allowed during power plays, but I don't agree with that either. I disagree with not counting goals scored against the person who committed the penalty (in hockey) or got sent off. He's responsible for putting the team at a numerical disadvantage and should be have goals counted against him, even though he is no longer on the field.

Another objection that I have is that your approach treats the 11v11 periods as one big contiguous block. It ends up treating a 40-minute period with three goals scored no differently than a 90-minute period with a goal scored (or none). I think it might be better to look at segments of 11v11 play and weight the goals allowed by the length of time of the individual segments.

I would also be interested in seeing how the plus/minus scores are affected by the personnel on the field. This would require a regression analysis that appears not be done here. Some of the players at the top of the list make sense, but others don't, and the overall rankings are skewed by the melding of the factors that I mentioned before.

I don't mean to flame you or anything like that. You've made a very good contribution here and I hope you continue refining your approach.

 

Blogger scaryice said ... (11:35 PM, March 14, 2010) : 

First of all, using only 11v11 did not make my life easier. It took a lot more work to figure it out that way. I wanted the numbers to be as fair as possible, and it's unfair to count plus/minus in a situation where one team has an advantage. All of the players in those moments clearly aren't on a level playing field. With both teams at full strength, you don't have to worry about that (BTW, in 2009 exactly 90.7% of minutes were played 11v11).

It's also unfair to count something that happened on the field against a player who isn't playing, regardless of what happens afterwards. What about the situations where the team a man down actually outscores the other team? Do you think a player should be rewarded for getting sent off then?

I don't get how you can say one 11v11 segment is different from another. Counting "a 40-minute period with three goals scored no differently than a 90-minute period with a goal scored (or none)" is exactly what I was trying to do. Why would you disagree with that? That makes no sense to me.

 

Anonymous price per head services said ... (4:55 PM, January 26, 2013) : 

It is great that Hockey came out with that kind of statics. it makes things more interesting when you know that kind of data.

 

post a comment