I was doing some digging around with some statistics to figure out who were the two best wide midfielders in MLS this past season, and one statistic jumped out at me - "Fouls Suffered." It turns out that the two wide midfielders who I would consider the most dangerous and probably have the most potential on the US Men's team also were fouled a LOT more than their nearest competitors. Before you read on, any guesses as to who those two players might be? The "fouls suffered" (FS) statistic might help to understand how dangerous a player is, and how hard it is for his peers to keep him in check. A player with a high total of FS is probably more difficult to defend than the average player, probably is a bigger part of his respective team's attack than the average player, and is more of a "gamebreaker" than the average player. By gamebreaker, I mean a player who by himself at any point of the game, can change the course of that game with a burst of individual skill. I don't think this statistic is the end-all, be-all, litmus test of whether or not a player is ready for the Men's National Team. There are very dangerous players who do not make the top of the list in FS. But when we discuss whether or not "PlayerX" deserves a shot on the MNT, it may be worthwhile to take a look at how many fouls "PlayerX" suffered in MLS that year, to help determine how much of a real threat "PlayerX" might be. Like I said, not the litmus test, but I think it can be helpful. Perhaps some follow-up posters will give some logic behind why Landon Donovan and Edson Buddle, for example, have a relatively low total of FS. But in doing the digging, I think these guys are exceptions to the rule, and that the general theory is this: If a player has a high total of "fouls suffered" in league play, that player is probably on the MNT radar. Here is a breakdown of every player who was fouled more than 40 times in MLS in 2003. Look and see how many of these players are either US or foreign internationals. I also included some players who didn't have 40 FS, but who I thought if given more minutes, probably would have. I also included two columns to show: 1. The average number of points each player scored over a 90 minute period (AP/90) 2. The average number of fouls each player suffered over a 90 minute period (AFS/90) Take a look: Code: [size=2] POS PLAYER MIN PTS FS AP/90 AFS/90 MID Jose Cancela 1142 9 64 0.71 5.04 MID D. Beasley 1969 19 96 0.87 4.39 MID Kyle Martino 1765 8 79 0.41 4.03 FWD Carlos Ruiz 2331 35 87 1.35 3.36 FWD A. Moreno 1117 14 40 1.13 3.22 FWD Brian McBride 2183 27 78 1.11 3.22 MID Ben Olsen 2239 15 80 0.60 3.22 FWD Brian Ching 1235 14 44 1.02 3.21 MID Andy Williams 1360 11 45 0.73 2.98 FWD Taylor Twellman 1893 34 57 1.62 2.71 MID Dema Kovalenko 2401 12 68 0.45 2.55 MID Amado Guevara 2264 16 64 0.64 2.54 FWD Damani Ralph 1985 28 56 1.27 2.54 MID Eddie Gaven 691 3 19 0.39 2.47 MID M. Etcheverry 2006 19 52 0.85 2.33 MID Cobi Jones 2574 12 65 0.42 2.27 FWD Mike Magee 1709 16 43 0.84 2.26 FWD Edward Johnson 1265 6 31 0.43 2.21 MID Oscar Pareja 1982 6 48 0.27 2.18 FWD Joe-Max Moore 1187 15 28 1.14 2.12 MID R. Mulrooney 2330 7 54 0.27 2.09 FWD Pat Noonan 1646 27 37 1.48 2.02 MID Preki 2678 41 60 1.38 2.02 MID Ricardo Clark 2590 7 58 0.24 2.02 MID Brad Davis 1910 17 42 0.80 1.98 M/F Jamil Walker 365 8 8 1.97 1.97 MID Shalrie Joseph 2466 6 54 0.22 1.97 FWD Landon Donovan 1882 30 38 1.43 1.82 M/F Brian Mullan 2723 21 54 0.69 1.78 F/M Clint Mathis 2019 19 40 0.85 1.78 FWD Ante Razov 2270 34 44 1.35 1.74 M/D Evan Whitfield 2329 3 44 0.12 1.70 FWD Chris Carrieri 2458 14 45 0.51 1.65 FWD John Spencer 2265 33 40 1.31 1.59 MID Kerry Zavagnin 2487 4 41 0.14 1.48 D/M Wes Hart 2623 2 40 0.07 1.37 FWD Edson Buddle 1509 24 20 1.43 1.19 [/size]
That's an interesting statistic, but I think it is a bit misleading. Not totally, but a bit. It seems to work best for attackers and attacking midfielders, but then you have a guy like Razov who trails only Preki and Ruiz on points (tied with twellman), yet gets fouled less than Ricardo Clark, a D-mid with offensive tendencies. So, it is a statistic of how "good" the player is, or how un-elusive he is? Getting fouled also, at times, means that he wasn't able to avoid a foul with better ball skills or speed. Granted, some guys come in studs up on you and you have to go down, but other times people look to get fouled. So, you have variables like how often and how convincingly the player dives, and how often they face teams that foul more than others. I think position variability like "Do center attacking mids get fouled more than wingers?" can account for a bunch of motion up and down this list. Also, you have size to consider, so a guy like McHead "suffers" a lot of fouls and keeps going without a call (or gets an advantage call) whereas a guy like Magee with a slighter frame falls down because he can't out-stregth a Jim Curtain. I'm not disagreeing with you, I'm just wondering aloud. It does clarify a general "on radar/off radar" status, but position top to bottom can't be taken too literally as Magee, much as I like him, is not as dangerous as Razov, Donovan, Preki, Noonan etc. Good work though. Maybe you could do a stat that was a combination of the two. There were 4838 fouls committed and 1284 points scored, making fouls ~ 3.75 times more common than goals. So if you take your AP/90*3.75+AFS/90 you get a hybrid stat that weights scoring and gets a bit closer to how dangerous they are, though I am by no means a statistician here. Also, your stat accomplished the "on the radar" hurdle, so figuring out how dangerous they are may not really be important as points per game does that in an established way. Code: PLAYER MIN PTS FS AP/90 AFS/90 Threat Level Taylor Twellman 1893 34 57 1.62 2.71 8.785 Carlos Ruiz 2331 35 87 1.35 3.36 8.4225 Jose Cancela 1142 9 64 0.71 5.04 7.7025 D. Beasley 1969 19 96 0.87 4.39 7.6525 Pat Noonan 1646 27 37 1.48 2.02 7.57 A. Moreno 1117 14 40 1.13 3.22 7.4575 Brian McBride 2183 27 78 1.11 3.22 7.3825 Damani Ralph 1985 28 56 1.27 2.54 7.3025 Preki 2678 41 60 1.38 2.02 7.195 Landon Donovan 1882 30 38 1.43 1.82 7.1825 Brian Ching 1235 14 44 1.02 3.21 7.035 Ante Razov 2270 34 44 1.35 1.74 6.8025 Edson Buddle 1509 24 20 1.43 1.19 6.5525 John Spencer 2265 33 40 1.31 1.59 6.5025 Just a thought. You did nice work (don't mind me, I'm just playing with your numbers).
I've been meaning to do a little work looking at MLS' stats and see if there are any stats that have large correlations with goals scored (and goals allowed) on a team by team basis (like a multiple regression thing). My guess is that fouls suffered as mostly a function of time on the ball, and closeness to the opposition goal. Donovan, for some reason, tends to not spend a lot of time on the ball, whereas Beasley (and Cancela) spent quite a good deal. For example, if you look at the Revs goal yesterday, you'll see Cancela spend a considerable amount of time doing ball tricks trying to find the right pass to make, before slotting it to Fabbro, who dummied, and Noonan ran onto it and finished. So number 10s and dribblers, and possibly guys that get lots of high balls played to them probably have the highest rates. I'm thinking that for most teams, these are usually among their best offensive players.
Dives Nice work, Nutmeg. If you backed out all of the dives from guys like Ruiz and Guevara, I think you would find them much lower on the list. Of course, many of Ruiz's dives turn into PK goals for him.
Re: Dives I think there is less here than meets the eye. Players who hold the ball get fouled more, those who pass quickly don't. Landon plays very fast 1-2s, so you can't get at him with the ball. Your sterotypic latin midfielder who dwells on the balls gets hit alot more. I would not call that a plus in and of itself, nor is a sign of quality.
Jamison "That's an interesting statistic, but I think it is a bit misleading. Not totally, but a bit." I actually think you are right. I think this statistic gives you a really good starting point, and then you have to dig a little deeper with each individual player to read the real story. For example, there has been a long thread lately on right midfielders, and Ben Olsen's name really hasn't been a prominent one. It should be. Why? Because whenever we speculate whether or not a player will be successful at the international level, one of the things that is consistently brought up is whether or not that player can create matchup problems and be a catalyst for the US attack. Ben Olsen has shown in league play that he definitely can be that type of player. A guy who is tough for his peers to defend, a guy who over the course of a season (a shortened one for Benny at that) has forced defenders to resort to hacking is a very good fit for the right midfield spot on the National Team. We could really use a guy who can be a gamebreaker on the right. Contrast Ben's numbers to those of his two closest competitors, and the statistics portray a story of one guy who can really be an impact player on the international stage - Olsen, and two other guys who will at most be serviceable role players - Klein and Ralston. The more gamebreakers, the more guys you have on a team who can individually cause problems for the opponent, the ones who can break their mark down 1v1, the more dangerous and potent your team will be as a whole. On the flip side, I think that defenders like Whitfield and Hart should definitely not be on this list. When you contrast their fouls per 90 with their points per 90, it would to me indicate a couple of guys who are too slow to react on the ball, who hold on to it too long and put their teams in bad spots, and who have no business even thinking about the international game. BTW, I really like your adjusted "threat level." It is interesting how accurate those rankings really are.
Re: Re: Dives Bruce S I think there is less here than meets the eye. Players who hold the ball get fouled more, those who pass quickly don't. Landon plays very fast 1-2s, so you can't get at him with the ball. Your sterotypic latin midfielder who dwells on the balls gets hit alot more. I would not call that a plus in and of itself, nor is a sign of quality. I think these numbers indicate a starting point. From there you can determine whether or not a guy is just sitting on the ball too long (like perhaps Kyle Martino) or whether or not this is a player who is really taking defenders on and forcing them to foul (like perhaps Beasley, Cancela, and yes, even Ruiz). If you're not ready to follow up with some extra analysis, this theory is incomplete.
Great stuff. One thing to keep in mind is that stats like this are NOT supposed to end the discussion. They are supposed to contribute to the discussion.
You probably need to throw a lot of Kovalenko's out because they are just retaliation fouls. I can't believe I just responded to a thread about stats and soccer!
The post above sort of beat me to it (sort of), but I'd like to see the same fouls stat recreated as a differential score, i.e. fouls suffered less fouls committed. I do think there's a tendency for guys like Kovalenko and Ruiz to draw fouls just because they whack the other guy a lot and are in the middle of things all the time.
... and according to MLSnet.com, only 32 goals were scored by the top eight teams from (non-penalty) free kicks. Only 10 of these came directly from free kicks. By and large, free kicks are not good scoring opportunities. So what did Kyle Martino accomplish by getting repeatedly hacked to the ground? Very little, especially since (IMO) MLS refs are criminally reluctant to card players for professional fouls. As you suggest, a lot of the most effective players draw fewer fouls than they otherwise might --they help their team by keeping the attack moving. Along with Donovan, Buddle, and Razov, these guys include Chung, Spencer, and Klein. Some of them do it through quickness, others through their strength.
I think you are making an unwarranted assumption. Getting fouled is not a cause, it's a symptom. I am sure that a Kyle Martino or Jose Cancela doesn't go out onto the field intending to get fouled. Except for Klein, all the guys you mentioned are strikers who I would expect to be in or near the 18 yard box, where fouls have notably graver consequences. Midfielders, on the other hand, are more likely to be far enough away from the danger zone that defenders will think the risk of fouling is worth taking. I like this metric, but another factor that distorts it is the advantage rule. Good passers are likely to get fouled while getting the ball to a team mate. The ref correctly waives play on, and the victim gets up (if he can) and goes on with the game. As a Revs fan, it seems I see this happen to Cancela 2 or 3 times a game, and I'm sure it's the same with other players of that type. On the other hand, I think this might compensate for the diving that milkduds was complaining about, since it's often the same players. For defenders, is there benefit in looking at the ratio of fouls committed to fouls suffered? I've long thought that Alexi Lalas is better than he should be because of his seeming nack from drawing fouls while defending. Should this thread be moved to the Stats forum?
You can look at it as a symptom, but in Martino's case, it's a symptom of failure. He had a total of two goals and four assists this year. He's a talented player, but opponents were able to neutralize him by fouling. That's why he ended up seeing some bench time. I agree that Cancela is a different story; players try to foul him, but he is still able to produce. Incidentally, Klein is not the only midfielder I listed ... I also included Chung and could have added Ralston, who had a successful national team run this summer. EDITED to note that I agree that this kind of thread should probably be in the Stats forum ... one problem with the Stats forum is that posts there don't appear on the first page of BigSoccer -- perhaps Nutmeg was concerned about trying to get more visibility?
I didn't see much of Martino this year, so I'll assume your description is accurate, but I still wonder about the causality. I would normally figure that a player gets fouled a lot if he has a lot of possession and has shown that he can be dangerous with time and space. It seems to me that defenders fouled Martino because they figured they had to, and then he turned out not to be able to adjust. I would be surprised if they fouled him more frequently because he couldn't adjust. After all if you concede enough free kicks, one of them is likely to lead to a goal. And there is persistent infringement to worry about. So I would think that it wasn't the fouling that was symptomatic of Martino's failure, it was his inadequate response to it. Or maybe it was really that awful haircut.
Careful ... this gets back to treating "fouls suffered" as a cause, not a symptom. As I mentioned before, free kicks very rarely lead to goals, less than 1% of the time. Martino was fouled 79 times and didn't draw a single PK. A reasonable guess is that all those fouls only set up one goal all season. But other than this one quibble, I agree with your reasoning to a large extent. Martino's exceptional ball skills (and ridiculous hair ) do make him a target.
Re: Re: One way to tell if a player is ready for the MNT I assume the stat you quoted earlier refers only to cases where a DFK goes in the net for a goal. If that is the case, does anyone have any data on on free kicks that lead to a goal more indirectly? I would define these as cases where the ball goes in the net without having left that end of the field.
Re: Re: Re: One way to tell if a player is ready for the MNT No, only 10 of the 32 goals came from a DFK that went straight into the net. These are MLSnet numbers, so I don't know precisely what they included.
I posted it here, because as you can see by the title of the thread, I was relating this particular statistic to the potential a player might have on the US Team. I saw it as more of an analysis of the US playing pool, which is in large part what this form is about, than I did a purely statistical question. Anyway, back to topic. I mentioned Martino in my earlier response to Bruce S. I think I sufficiently prefaced this metric at the beginning of the thread by saying this is not the end-all, be-all statistic that will tell you whether or not a player might be ready for international soccer. It may be a tool for analysis. One of many. I happen to think it is a good starting point. I think when you compare the number of fouls Martino suffered to his production, you start to see a picture of a player who is holding on to the ball too long and is getting beat to hell because of it. Comparing Cancela to Martino is, I think, a very useful comparison. Both play the same position. Both get fouled a lot. But Martino hasn't produced in the form of setting up his teammates, or creating opportunities for himself. Then the discussion focuses on, "Why?" Like Superdave said, I think statistics serve as a useful tool in starting and/or contributing to discussions and comparisons. They are not meant to be the end of the story.
"Producing" points and goals is an individual statistic that is heavily reliant on how the TEAM is doing. Which is one reason why a player like Brian McBride, who has consistently been at least one of the top USMNT forwards, has consistently been outscored by the likes of Jason Kreis at the MLS level (well that and injuries). Looking at the fouls suffered statistic just is another tool to try and identify who the top players are, regardless of whether they are on a good MLS team. It's value lies in that it does provide an indication why Bruce Arena chooses Kyle Martino despite his lack of MLS production. Of course, watching the game could tell you the same thing (and I think those that argue that Martino is not a good MLS player are wrong). Of course, those of us without time or access to the 90% of MLS games that BA watches, might want some shortcuts. Thanks Nutmeg.
Many free kicks don't occur in dangerous areas of the field. A foul at midfield or on your own defensive end is unlikely to result in a goal, hence reducing the % of goals scored. A foul with 15 yards of the penalty area is far more likely to result in a goal. Of course no statistic is kept for where fouls occur, so its a bit misleading to just go off total fouls. Beasly has a tendency to draw many fouls, but the ones on either side of midfield are generally speaking worthless as an offensive measure. What I do think fouls suffered is a good measure of is overall aggressiveness on the offensive end. Players who push and force the action are far more likely to be fouled than one who does not directly challenge a defender. Conversely, I think fouls committed is probably a pretty good measure of defensive aggresiveness, the likelihood of trying to pressure and win balls. Probably not a good thing for a defender, who you dont want commiting fouls and giving dangerous free kicks, but for a wing midfielder or forward a high fouls commited number would probably be a good thing, as it indicates that they are actively pressuring the other team, and as such are likely to force more bad passes or win more 50/50 balls that turn into dangerous opportunities. Similairly, point production is a usefull stat but also misleading if taken in a vacuum. Twellman for example I think is partly as productive as he is because he has been the focal point of the attack for New England when healthy. On the otherhand, a player with little service is much less likely to put up large totals. Which ties into a third consideration. Touches on the ball. How many times a player touches the ball in a game is of some signifigance in this discussion. It is hard to be fouled or gain a point if your own team limits your touches via style of play or just poor passing. A third possibility is that a player is simply not able to get open. All of which gets to the heart of the statistical problem in soccer, not that statistics are meaningless for the game, just that which statistics are signifigant, and which are insignifigant is as yet undetermined, largely due to the neglect other parts of the world have for them. The basic idea, of counting and measuring events and using them to identify broad patterns in the game is as sound as it is in any sport, its just that the metrics are poorly developed. Thats also partly due to the nature of game not being stat friendly. Baseball on the other hand is very stat friendly in many respects, and in recent years there has been an excellent effort to understand and gain more meaning from them. Unfortunately, I think that these discussions are interesting, but maybe ultimately doomed as we are, metaphoricly speaking, given a set of dinnerware for the task of digging the panama canal.
I think it's worth repeating that the average playoff team scored a grand total of four goals from free kicks all season, and only 1.25 of those were direct shots. While I'm sure there were more goals from 30 yards out than 60 yards out (excepting Guevara-to-Woly), no free kick outside the area has much chance of producing a goal. (Kenn T. recently pointed out that goals from corner kicks are also very rare.)
The average playoff team scored around 44 goals, so fouls resulted in roughly 9% of a teams total goals, so its not an insignifigant number. Admitedly the chance of conversion even on a free kick is small even if you toss out the ones where the likelihood is near zero. However, to toss out some numbers not grounded in fact at all, if an attacking player on average draws 1 foul a game within 15 yards of the box, and on average 4% of free kicks within 15 yards are converted, that player being fouled is adding over 1 goal a year. Its not earthshattering, but it certainly should be taken into consideration.
Getting fouled is better than nothing ... but if an attacking player sets up just one ordinary run-of-play shot per game (~10% of converting), he adds 3 goals per year. Our best indication is that most goals are scored when defenses are caught out of shape. As a consequence, it's important to pounce when the opportunity arises, rather than allowing the defense to commit a tactical foul and regroup. An unrelated point: IMO, it's interesting to see that MLS's most durable attackers (Preki, Chung, and Ralston) have all been relatively good at avoiding fouls. By contrast, a guy like Tab Ramos suffered a lot of fouls and had his career cut short. This bodes well for Landon, but not for Damarcus.