I agree with Serie Zed that fouls committed and offsides should count against the player (would be even better to consider only offensive fouls, but unfortunately the data aren't split out that way). I also think shots on goal is a more fundamental measure of a player's "danger" than goals, since it is less dependent on the goalkeeper. With this in mind, I calculated an Adjusted Shot on Goal and Adjusted Foul Suffered statistic to yield an Adjusted Danger Rating (ADR). And whadya know, Segroves favorite player tops the list... <br /> </font><blockquote><pre><font face="verdana,arial,helvetica" size="1" >code:</font><hr><font size="2"> RANK POS PLAYER ADR 1 FWD Brian McBride 72.32 2 FWD Carlos Ruiz 68.58 3 MID DaMarcus Beasley 67.30 4 MID Kyle Martino 60.89 5 MID Jose Cancela 59.10 6 MID Amado Guevara 57.77 7 MID Preki 57.10 8 FWD Ante Razov 54.84 9 MID Ben Olsen 54.69 10 FWD Taylor Twellman 54.15 11 MID Cobi Jones 48.30 12 FWD Damani Ralph 44.46 13 MID Mark Chung 42.74 14 MID Dema Kovalenko 42.40 15 MID Marco Etcheverry 38.49 16 MID Ricardo Clark 37.84 17 MID Richard Mulrooney 37.52 18 FWD Landon Donovan 37.26 19 FWD Mike Magee 36.20 20 FWD Pat Noonan 35.20 21 FWD Chris Carrieri 34.25 22 FWD John Spencer 33.59 23 M/F Brian Mullan 33.32 24 FWD Igor Simutenkov 32.72 25 F/M Clint Mathis 32.45 31 D/M Frankie Hejduk 29.59 <hr></pre></blockquote><FONT face="arial, helvetica" size="2"></font></font></p> <p><font face="verdana,arial,helvetica" size="1" color="#666666"> <br>
This sounds like you're trying to come up with something like Voros's idea about baseball pitchers (where they are fundamentally responsible for defense-independent events: walks, strikeouts, and home runs). That's a commendable pursuit, but for goodness sakes, have you looked at the data? Prior to 2003 Chris Albright 56 shots on goal Taylor Twellman 58 shots on goal Your measure says that Albright and Twellman were fundamentally equal, even though Albright had 4 goals and Twellman had 23!
A couple of things: - Clearly fouls are a second order concern in terms of offensive contribution, but they shouldnt be ignored. Unfortunately the tools for figuring out how important are not available. Additionally, while individually PKs drawn may not match up with fouls, as a group players who draw more fouls near the box are probably more likely to draw a PKs. Again, numbers arent available, but that also would increase a players production. Although a simpler solution is to just measure PKs drawn. - Shots, SOG and Goals all probably have their place in evaluating marksmanship, but how much variation is due to luck and how much due to a players accuracy can be hard to say in a short period. Clearly Pat Noonan is an accurate shooter, but is he as good as his statistics would suggest? Id be inclined to say probably not quite as good as he has shown, but that would be pure speculation. Really, soccer statistics arent really available, MLS as a whole has done as good a job as any league, but they arent ideal. There is also a lot of number crunching that can be done to try to identify signifigant statistical patterns, to understand the ones we do have. Like how much variation is there in shooting percentages between players?
Not necessarily. My "danger" measure is a function of Shots, Shots on Goal, Fouls Suffered, Fouls Committed, and Offsides. I haven't looked at Chris Albright's numbers, but it could be that differences in the other parameters would yield a significantly different rating for him compared to Twellman. Lurking has it absolutely right. Clearly shots and fouls by themselves are totally inadequate as danger measures, but those are the only individual offensive statistics MLS deigns to collect. In my view, passing, trapping, and dribbling are likely to be important components of "danger" that our current stats completely ignore. Still, it's interesting to see what falls out of the meager statistics we do have - note that according to this metric the top four "most dangerous" American forwards in MLS are McBride, Razov, Twellman, and Donovan. Edit: FWIW, for the 2003 season, Twellman ranks 10th, while Albright comes in at 33rd, lagging behind Taylor in both the shooting and foul-related components.
Oh, and another interesting tidbit, as I showed in the original table, Frankie Hejduk (!) is rated as the "most offensively dangerous" defender according to this criteria, and by a fairly wide margin over the next defender, Wes Hart who is ranked 41st. I rarely get to see the Crew play, but I'm left scratching my head trying to figure out why anyone would find it necessary to foul Frankie. ;-)
Which, as I was pointing out, completely ignores goals scored. This kind of nonsensical rating system is what gives sports stats a bad name.
I'm not really sure what you're saying here, but 2003 numbers for PK's drawn were tabulated by ChrisE over on the stats forum. It's also possible to tabulate some info about a lot of the free kicks that led to goals. Some (esp. direct free kicks) are likely to be mentioned in match reports, such as this one, where a goal followed a Rusty Pierce foul on Landon Donovan. www.cnnsi.com/soccer/us/news/2003/08/16/revs_quakes/ This one was: Pierce on Donovan, 19 yds out, left of center, scored directly
The fouls near the box that result in goals are easy to find because goals are pretty well documented. So are PKs, which is why you could account for the number of PKs drawn. The problem I see are counting the fouls around the box that do not result in goals, which are not well documented. I think the number of those fouls may be a better predictor of the future # of PKs a given player will draw on average than the number of PKs drawn in the past, since PKs are handed out infrequently and are hence subject to a high degree of random variation.
Your last point (relating total fouls and PK's to fouls-near-the-box) is an excellent one ... even without data for non-goal-causing fouls, we could try and model this relationship. Wish I had more time to spend on it right now.
Let me remind you that the subject of this thread is whether a measure that is NOT goals is useful as an indicator of "dangerousness" in the sense of offensive prowess. There could be many explanations for why Twellman's ratio of goals to shots on goal is higher than Albright's - maybe on average Albright shoots from much further away from goal than Twellman does, maybe his shots on goal are higher and therefore easier to save, or maybe he is more prone to shoot directly at the keeper, or maybe you're right and your statistic proves beyond the shadow of a doubt that Twellman is "God" and Albright "sucks". I analyze soccer statistics because I enjoy doing so, not to please you, so why post if you have nothing constructive to add?
I offered constructive criticism by pointing out that you can't sensibly pretend that shots on goal are a useful substitute for goals scored. Now, here are two questions for you: 1) Why post to a thread that you don't seem to have read? 2) Why post to a thread when you aren't interested in thinking about other people's response to your ideas?
So just for fun, I decided to re-run these statistics for the teams that made it to the conference finals in MLS playoffs, and see how they compared to the regular season. I also thought Jamison's "threat level" (Named JTL) statistic was interesting, so I included it in this chart. The following is a table of every forward and midfielder who played at least 45 minuts on a team that made it to the conference championship: Code: [size=2] PLAYER MIN PTS FS AP/90 AFS/90 JTL Justin Mapp 45 1 1 2.00 2.00 9.50 Ante Razov 251 5 4 1.79 1.43 8.16 Damani Ralph 273 3 13 0.99 4.29 7.99 D Beasley 252 1 15 0.36 5.36 6.70 Andy Williams 236 2 6 0.76 2.29 5.15 Chris Armas 281 3 3 0.96 0.96 4.56 Jesse Marsch 270 0 7 0.00 2.33 2.33 Igor Simutenkov 142 5 2 3.17 1.27 13.15 Chris Klein 297 5 0 1.52 0.00 5.68 Preki 297 1 9 0.30 2.73 3.86 Wolde Harris 261 2 3 0.69 1.03 3.62 Kerry Zavagnin 297 1 6 0.30 1.82 2.95 Alex Zotinca 90 0 1 0.00 1.00 1.00 Eric Quill 186 0 2 0.00 0.97 0.97 S Armstrong 130 0 1 0.00 0.69 0.69 Josh Wolff 46 0 0 0.00 0.00 0.00 Jose Cancela 281 3 14 0.96 4.48 8.09 Dario Fabbro 242 3 3 1.12 1.12 5.30 Pat Noonan 281 4 1 1.28 0.32 5.12 Jason Moore 140 0 6 0.00 3.86 3.86 Shalrie Joseph 281 0 10 0.00 3.20 3.20 Brian Kamler 138 0 2 0.00 1.30 1.30 Steve Ralston 281 0 3 0.00 0.96 0.96 Landon Donovan 303 6 8 1.78 2.38 9.06 Jamil Walker 167 3 4 1.62 2.16 8.22 Rodrigo Faria 115 2 2 1.57 1.57 7.43 Manny Lagos 145 2 2 1.24 1.24 5.90 R Mulrooney 303 3 7 0.89 2.08 5.42 Ian Russell 158 2 2 1.14 1.14 5.41 Ramiro Corrales 72 0 3 0.00 3.75 3.75 Brian Mullan 303 2 5 0.59 1.49 3.71 Ronnie Ekelund 213 2 0 0.85 0.00 3.17 D De Rosario 57 0 2 0.00 3.16 3.16 [/size]
Next comes a positional comparison. First up: forwards: Code: [size=2] PLAYER MIN PTS FS AP/90 AFS/90 JTL Igor Simutenkov 142 5 2 3.17 1.27 13.15 Landon Donovan 303 6 8 1.78 2.38 9.06 Jamil Walker 167 3 4 1.62 2.16 8.22 Ante Razov 251 5 4 1.79 1.43 8.16 Damani Ralph 273 3 13 0.99 4.29 7.99 Rodrigo Faria 115 2 2 1.57 1.57 7.43 Dario Fabbro 242 3 3 1.12 1.12 5.30 Pat Noonan 281 4 1 1.28 0.32 5.12 Wolde Harris 261 2 3 0.69 1.03 3.62 D De Rosario 57 0 2 0.00 3.16 3.16 Josh Wolff 46 0 0 0.00 0.00 0.00 [/size]
wide midfielders: Code: [size=2] PLAYER MIN PTS FS AP/90 AFS/90 JTL Justin Mapp 45 1 1 2.00 2.00 9.50 D Beasley 252 1 15 0.36 5.36 6.70 Manny Lagos 145 2 2 1.24 1.24 5.90 Chris Klein 297 5 0 1.52 0.00 5.68 Ian Russell 158 2 2 1.14 1.14 5.41 Andy Williams 236 2 6 0.76 2.29 5.15 Jason Moore 140 0 6 0.00 3.86 3.86 Ramiro Corrales 72 0 3 0.00 3.75 3.75 Brian Mullan 303 2 5 0.59 1.49 3.71 Brian Kamler 138 0 2 0.00 1.30 1.30 Alex Zotinca 90 0 1 0.00 1.00 1.00 Eric Quill 186 0 2 0.00 0.97 0.97 Steve Ralston 281 0 3 0.00 0.96 0.96 S Armstrong 130 0 1 0.00 0.69 0.69 [/size]
defensive midfielders: Code: [size=2] PLAYER MIN PTS FS AP/90 AFS/90 JTL R Mulrooney 303 3 7 0.89 2.08 5.42 Chris Armas 281 3 3 0.96 0.96 4.56 Shalrie Joseph 281 0 10 0.00 3.20 3.20 Kerry Zavagnin 297 1 6 0.30 1.82 2.95 Jesse Marsch 270 0 7 0.00 2.33 2.33 [/size]
attacking central midfielders: Code: [size=2] PLAYER MIN PTS FS AP/90 AFS/90 JTL Jose Cancela 281 3 14 0.96 4.48 8.09 Preki 297 1 9 0.30 2.73 3.86 Ronnie Ekelund 213 2 0 0.85 0.00 3.17 [/size]