I'm going to use this thread as an opportunity to break down our goals by the numbers in these past couple cycles against quality opposition. I'd like to think of this as a companion thread to Susaeta's thread, since I didn't feel like clogging up his thread with a ton of stats. I've tracked all of the goals we've scored since the beginning of the last cycle against top 50 opposition (as judged by Elo rankings at the time) and recorded goalscorers and those officially credited with assists and (not officially) credited with secondary assists. I've also divided them into objective and somewhat subjective categories to help give a better sense of how our goals are created. Later in the thread, I'll also break these down by distinct periods over this time frame to show how it's changed (and maybe I'll also break this down further into tiers of quality among opposition even within the top 50). For now here's the general information. (One note however; I've excluded January camp friendlies and the 2009 Gold Cup, since the quality of the opposition in those games generally didn't reflect their rankings in any way.) Since the beginning of 2007, we've played 51 teams ranked 50th or better by Elo at the time. In those matches, we've scored a total of 70 goals (a return of 1.37 goals per game). Penalty Kicks 8 of those 70 goals (or 11%) have come through penalty kicks. Landon Donovan has scored 6 of those goals. Eddie Johnson and Jozy Altidore have each scored 1. The following is a list of players who drew those penalties (each of whom drew one): Jozy Altidore, DaMarcus Beasley, Brian Ching, Clint Dempsey, Eddie Johnson, Fabian Johnson, Oguchi Onyewu. Another came through an opponent's handball. Feilhaber had the final pass before the foul on two of those plays (against Argentina and Italy). Chandler, Dempsey, Donovan, and Spector each had a final pass on those plays as well. Dempsey (twice) and Feilhaber (once) have also had secondary assists on those penalties. I'll note here that I'm also going to include these passes in the total assists and secondary assists figures later, so if you disagree with that methodology, just subtract the above numbers from the ones later on. Set Pieces (Corner Kicks and Free Kicks) 17 of those 70 goals (or 24%) have come off of set-pieces. Two of those were direct free-kicks (scored by Landon Donovan against Honduras and Eddie Lewis against Poland) and one was the result of a long free kick taken by Oguchi Onyewu in our own half, flicked on by Charlie Davies and scored by Connor Casey (against Honduras). Here are the numbers for the remaining 14 goals. Goalscorers: 1. Bocanegra - 4 2. Bradley, Onyewu - 2 4. Agudelo, Bornstein, Conrad, Dempsey, Edu, Gomez - 1 Set-piece Delivery: 1. Donovan - 7 (4 assists) 2. Beasley, Holden - 2 (1 assist each) 4. Adu, Bradley, Rogers - 1 (1 assist each) Crucial flick-on headers: 1. Goodson - 2 2. Altidore, Bocanegra, Dempsey, Donovan Deep throw-in's: 2 additional goals were scored off deep throw-ins. Onyewu scored off a header against Mexico in 2008 after Donovan recovered Onyewu's own long throw-in and delivered an assist. And Charlie Davies scored against Egypt after Spector sent in a short-throw in two Jozy Altidore who sent a low cross to Davies. These two goals will also be included in the "Run of Play" figures below. Run of Play Goals: The remaining 45 out of 70 goals (or 64%) have come from the run of play. Additionally, assist figures will include the final two passes preceding a foul on a converted penalty kick. Goalscorers: 1. Donovan - 10 2. Dempsey - 9 3. Bradley - 6 4. Altidore - 4 5. Buddle, Gomez - 3 7. Davies, Jones - 2 9. Casey, Clark, Cunningham, Feilhaber, Hejduk, Rogers - 1 Assists: 1. Donovan - 12 2. Dempsey - 5 3. Altidore, Cherundolo, Jones, Spector - 3 7. Feilhaber - 2 8. Agudelo, Beasley, Bradley, Chandler, Ching, Clark, Davies, F. Johnson, Moor, Orozco, Shea, Twellman - 1 Secondary Assists (note: I allowed for players to be credited for secondary assists on their own goals -- e.g. Bradley against Egypt): 1. Bradley, Donovan - 5 3. Feilhaber, Gomez - 3 5. Adu, Dempsey - 2 7. Agudelo, Altidore, Bocanegra, Davies, Demerit, Findley, Holden, F. Johnson, Mastroeni, Moor, Shea, Szetela - 1 Assists and secondary assists combined: 1. Donovan - 17 2. Dempsey - 7 3. Bradley - 6 4. Feilhaber - 5 5. Altidore - 4 (Edit: Oops, Donovan was not credited with an assist in the win over Spain, and thus Feilhaber cannot be credited with a secondary assist.) Types of Goals from the Run of Play: Long Distance Goals: 11 of our 70 goals (16%) have come from outside of the box. Four of those have been following a defensive clearance, and four of them have been the result of large defensive errors. Here are our long distance goalscorers. 1. Donovan, Bradley, Buddle - 2 3. Cunningham, Dempsey, Feilhaber, Hejduk, Jones - 1 (Note: If it was too close to call, I gave the benefit of the doubt to it being inside the box.) Long Balls: 3 of our 70 goals (4%) have been the result of long-balls. (Not counting Casey's goal off Onyewu's long free-kick from earlier.) - Bradley's goal against Slovenia in 2010 came off a Donovan long ball and an Altidore flicked on header. - Clark's goal against Paraguay in 2007 came off a Drew Moor long ball and a flicked on header from Twellman. -The penalty that Brian Ching drew (and Landon Donovan converted) against Mexico in the 2007 Gold Cup final came off a Spector long-ball. Flank play: 10 of our 70 goals (14%) have been scored inside the box largely as a result of play down the flank (this category is somewhat subjective though - it includes crosses from wide areas and players dribbling in from wide areas). Goalscorers: 1. Dempsey - 5 2. Bradley - 2 3. Altidore, Buddle, Donovan, Rogers - 1 Main contributor to the goal on the flanks: 1. Donovan, Spector - 2 2. Cherundolo, F. Johnson, Moor, Rogers, Shea, Szetela - 1 *Note: Donovan is credited as both the goalscorer and the main contributor on the flank to his goal against Slovenia in 2010. Full Counter-Attack: Another very subjective category, but by this I mean goals created from a fast-break in or around our own half where the opposition is struggling to get numbers back. (Note: I didn't count goals where we had a fast-break, slowed the play down considerably, and then scored.) 7 of our 70 goals (10%) were scored this way. Goalscorers: 1. Donovan - 3 2. Bradley, Casey, Dempsey, Jones - 1 Assists: 1. Donovan - 3 2. Agudelo, Ching, Clark, Davies - 1 Secondary Assists: 1. Donovan - 2 2. Bradley, Davies, Gomez - 1 Through-Balls: Goals that were largely the result of defense-splitting through balls (although they were not-necessarily the final balls). Again this one is pretty subjective, since I was trying not to include give and go's. 12 of our 70 goals (17%) were the result of such passes. 5 were long range passes; 7 were short range. Play-makers: 1. Bradley (both short range), Dempsey (short and long range), Donovan (both short range), Feilhaber (both long range) - 2 5. Adu (long range), Chandler (short range), Findley (long range), Jones (long range) - 1 Players at the end of the pass (not necessarily goal-scorers): 1. Donovan - 4 2. Altidore, F. Johnson - 2 3. Beasley, Casey, Davies, E. Johnson - 1 High-Pressure: This one's hard to tell because too many of the highlight clips I'm getting these from start too late into the play. But, as far as I can tell 7 of our 70 goals (10%) directly followed a turnover in or around the opponent's third of the field. Players whose initial pressure led to the turnover: 1. Dempsey - 2 2. Cunningham, Findley, Torres - 1 3. ?? (a couple where it appears a turnover was forced but it's too hard to tell by whom) 1 v 1 Goals: These are run of play (or PK) goals where beating a player 1 v 1 while closely marked was key. 5 out of the 70 goals we scored (7%) involved these. (I was pretty generous with these, but it's still possible I missed some). - Altidore vs Spain (2009) - outmuscled defender - Dempsey's drawn PK vs Ghana (2010) - dribbled past defender before being pulled down - Dempsey vs Turkey (2010) - dribbled past defender to score goal - Feilhaber vs Spain (2009) - dribbled past defender in buildup to Dempsey's goal - Shea vs Mexico (2011) - dribbled past defender to assist Rogers' goal With one or two exceptions, I believe that encompasses every goal against top 50 opposition that we've scored in these past two cycles. As I said, when I have the time I'll break it down further by era and quality of opposition. To sum up, here's how we score most of our goals against decent quality opposition (the percentages won't add up perfectly because there's overlap between categories): 1. Set pieces - 24% (with Donovan delivering and most often Bocanegra scoring) 2. Through balls - 17% (7% long range and 10% short range) 3. Long-distance strikes (i.e. outside of the box) - 16% 4. Flank play - 14% (with Dempsey at the end of it) 5. PK's - 11% (with Donovan scoring and any player you can think of drawing it) 6. Full counter-attack - 10% (with Donovan involved in everything) 6. High pressure - 10% 7. 1 v 1 skill - 7% 8. Long balls - 4%
Would be interested in seeing the difference in the breakdowns between Bob and Klinsy. I feel like we've already scored loads more from open play than we did the last year or so for Bob. While the set-pieces have been less effective.
I can tell you that right now (at least against top 50 opposition): Bradley's teams scored 17 goals off non-PK set-pieces and possibly as high as 19 if you count deep throw-in's. That came out to between 0.4 and 0.44 set-piece goals per match for Bradley. Factor in the 7 PK's, and you're as high as 0.6 set-piece goals per game. Well over a goal every other game. That left 34 goals from open play, or rather 0.79 goals per game from open play. By contrast, Klinsmann's USMNT have only managed two set piece goals against top 50 opposition (counting PKs), both of them against Slovenia. That comes out to 0.25 set piece goals per game. From open play, Klinsmann's sides have scored 10 goals in 8 games, or 1.25 such goals per game. The sample size for Klinsmann is pretty small though, and the blowout against Scotland is having a pretty big influence on these numbers. Take that game out, and you suddenly have 5 goals in 7 games or 0.71 goals from open play per game, lower than Bob Bradley's rate. We probably have to wait a bit longer for a larger sample size before we can say much meaningful there.
Kalm, your hard work is much appreciated. Donovan really shines throughout despite the recent perception that he is holding-back somehow. After seeing him miss scoring after an exciting breakaway last night, I'm unconvinced that he has much more to offer the USMNT in qualifiers and WC '12. At or around his peak, he's got to score on those chances.
You want write off LD because he misses a 1v1 in MLS? If anything that rush showed how fast he still is.
Nice work. Could you add the stats on 'chances created'? Goals scored is only a small part of the story.
Plus his assist was a moment of the old Donovan and the new one combined: Speed and attacking mentality along with the experience to wait for the perfect time to make the pass.
He is responsible for 50% of LA's goals since he has come back from the National team break (see my post in the final third thread) - and if Keane hadn't missed several sitters the stat would be even higher. 5 goals and 4 assists in 7 games and you're focusing on 1 breakaway? Basically his play since he's come back from the national team shows how lethargic he was in the beginning of the season. Scotland was the turning point and he hasn't looked back. But good lord, looking at those stats it is impressive to see how much he drives our attack, and Clint Dempsey too. WOW. D&D indeed.
KALM, kudos. What a spectacular post. Great effort, great research, and we should keep this thread around for comparison as this cycle continues - can it be stickied? One player I'm very interested to watch is Michael Bradley. How does his game change? And how do his stats change???? Can't wait to find out!
I just wanted to add a couple things to the original post. These are categories that I initially left out because there was too much overlap with the ones I already included, but then I noticed that there was some interesting information regarding the individuals involved. Short Crosses/Cutbacks These are goals that I did not include in the "Flank Play" category, because I did not think they fit the traditional mold of what you'd expect from a winger or wingback. They are goals that result from centering balls or crosses in or around the box. (I also included a couple cutbacks just outside the box that resulted in goals.) 13 of our 70 goals were scored this way. That's 19% of our total goals. Goalscorers: 1. Donovan, Dempsey - 3 3. Bradley, Gomez - 2 5. Altidore, Davies, Hejduk, Jones - 1 Passers/Crossers: 1. Donovan - 7 2. Agudelo, Altidore, Beasley, Cherundolo, F. Johnson, Jones - 1 Donovan racks up a large number of his assists this way, and it fits with the narrow winger role he plays for the US. Target/Hold-Up Play Two things I want to look at here. First: goals that were scored where a striker playing with his back to goal got a crucial touch. Second: goals that were scored where a striker playing with his back to goal held off a defender/battled in the air to get a crucial touch. In general, 15 of our 70 goals (21%) involved a crucial touch from a striker playing with his back to goal. Those strikers and the number of such goals they were involved in as follows: 1. Altidore - 6 2. Gomez - 3 3. Ching - 2 4. Adu, Davies, Findley, Twellman - 1 (Note: One of those includes a penalty that Ching drew against Mexico in the 2007 Gold Cup final) Even more interestingly, 9 of our 70 goals (13%) involved a striker with his back to goal holding off a defender or battling in the air. Those strikers and the number of such goals they were involved in as follows: 1. Altidore - 6 2. Adu, Ching, Davies - 1 (If Adu's inclusion surprises you, I'm crediting him for his touch on Donovan's goal in last year's Gold Cup final against Mexico.) In short, Altidore's strength/target play is a pretty good asset for us and has probably been slightly underrated over the years.
I might give it a shot at some point, but I'm working off highlight videos here and I can't be sure those include all chances created. If there's a site that records this stuff though, I could compile the data for a much more basic look though. This I can do (but again probably not right away). It would probably be useful for putting some of these goalscoring and assist numbers in context.
To me this is one important quality missing from Klinsmann's current team makeup, and where Feilhaber in the past has been a good option. He is willing to make that killer last pass. Obviously a big part of how the US scores, and maybe why Bocanegra is still so valuable to the team. Amazing how much of the US's production comes from those two. How does the US score goals? 1. Donovan 2. Dempsey 3. Other ways. I love this statistic. Being from the belief that the most valuable asset in soccer is a player who can create opportunities, the Feilhaber exclusion bothers me more and more. My guess is he has a fraction of the minutes played than everyone else on this list, yet he remains productive. Yikes. Not a big number or percentage for the number of long balls played by the US, but it echoes what I have felt for a long time watching the US. When they play quickly through the midfield, their attack is far more dangerous than when they play over the top. To me one of the more concerning elements of the US game. You see an increased emphasis on flank play given the lack of space in the middle of the field in today's game. Seems like there should be more emphasis here. Maybe Johnson out wide will help in this area in the future. If Lichaj gets time on the right, I think he shows signs of being a good wide attacking option as well. Also makes me hope Gatt or Gyau force their way into contention. This shocked me. I would have guessed this number a lot higher. Not a surprise that Donovan was a key factor in these plays, though. Another interesting stat I am trying to wrap my head around. It bothers me that maybe these numbers should be a bit higher. Clearly Donovan is the US's best off-the-ball runner, though. Does the US play with the right guys behind him to find him on those runs? Not sure. Again Feilhaber stands out with shortened minutes. One of the more underrated aspects of his game. If he smells blood in the water, Dempsey can be a very aggressive defender. Not a high number. I think Johnson may add to these totals in this cycle, but 1v1 dribbling skills does not seem like it will be a primary producer for the US in the near future. Thanks again. Really makes me challenge some assumptions. Overall, here are my takeaway thoughts (including what you posted later on forwards with their backs to the goal): 1. The US must focus on putting Donovan and Dempsey in their best spots to create and finish goals. 2. Feilhaber was productive contributor, even against top teams, being ignored now. Tough to understand. 3. Set pieces should be a continued emphasis for the US. At ~1/4 of their goals, it is a big part of how they score.
Thanks for the response. I want to put some of these numbers in context, which may or may not change some of your analysis. He might actually have six total, tied with Bradley for 3rd on the list of assists and secondary assists combined, but I couldn't make out for certain whether he was the player who had the secondary assist (long ball to Beasley on Donovan's third goal) 3:33 into this video: Just to clarify, by long balls I meant long balls into the final third. And I didn't include goals such as the one above that involved picking out players running into space with long lobbed passes. There were also at least one or two goals that resulted from long balls into the middle third of the field, but all of those required something extra from one of the other categories to turn them into actual goals. For example, 1:49 into this video: I included that goal in the "Through Balls" and "Cut Backs" categories, but not the "Long Ball" one. And to be fair to US coaching staffs, I have a feeling that when our defenders go long, they're hoping for something more along the lines of that play than the type I listed in the "Long Ball" category (which are more like Hail Mary efforts). Again, just to clarify I had a certain type of flank play in mind with this category, which was crosses from well outside the box, or if they were not from outside then at least they were the result of dribbling down the flank the way you would expect of a winger. (The one exception was Shea's assist to Rogers against Mexico, which I debated putting in the short crosses category, but decided against it since that is the sort of dribbling I would expect from a winger.) As the "Short Crosses" category later shows, we do use width, just a bit narrowly. And we don't tend to do it through dribbling -- we tend to find runners in space near the edge of the box who then cut the ball back or send driven crosses into the six yard box. Surprised me too. But it also might be the way I counted it. I had a certain type of counter-attacking play in mind with this category -- like I said, where we recovered the ball at least in the middle third of the field or further back, and where we used pace and precision to exploit the opposition struggling to get numbers back in position. The paradigmatic example being Donovan's goal against Algeria, which I could not count because Algeria were ranked too low. (Which might be another problem -- that teams of a certain quality don't often find themselves in that position, or at least not until very late in the game when they're pushing for a goal.) That said, there are goals that we've scored by pouncing on turnovers in the opposition's third of the field and catching the defense on their heels that I didn't include in this category but that still speak to our counter-attacking ability -- think Dempsey's goal against Spain in 2009. There are also goals that weren't the result of turnovers and where the opposition had numbers back, but that still involved the type of skill and precision at speed that make us a quality counter-attacking team: think Davies' goal against Mexico in our 2009 away qualifier. Despite the low numbers here, I'd still argue this is a strength of ours.
Actually, never mind. Looking back through my chart, it seems I did end up counting that pass under Feilhaber's secondary assists. So change that post to say he might only have four if that wasn't Feilhaber in that video. (On the other hand, he was probably deserving of another secondary assist against Spain, but once again since Donovan wasn't recorded with an official assist on Dempsey's goal, Feilhaber cannot be credited with a secondary assist there.)
Great thread, Kalm. From your analysis, other than working harder on set pieces, do you have any other observations as to where future goalscoring opportunities might come from?
I think Susaeta more or less came to the same conclusions I did. But looking through this, I can only tell you what's worked so far, not what might work with a different set-up.Which means most of these points are going to be fairly obvious to anyone who's watched the US in the past 6 years. 1. Long ball is generally not an effective form of attack against quality opposition. Even including long balls we send to the middle third of the field, there aren't very many goals created this way. If our coaches are seeing the same thing we are, I suspect we're only using it as a form of defense rather than attack -- that is to avoid costly turnovers in our third of the field. 2. We're not going to score goals through individual efforts, players taking on opponents 1 v 1 and scoring or assisting. Maybe there's someone out there in our player pool who can do it, but the crop of players we have right now don't do it on a consistent basis. Perhaps our players can't do this because our coaches aren't letting them, but even with Klinsmann supposedly encouraging our players to take on opponents 1 v 1 more, there hasn't been an increase in the rate of goals scored this way. 3. Find Donovan in space however you can. He's had a direct involvement in well over half of our run-of-play goals against quality opposition during the past 6 years, and looking through my list quickly it appears that about 4/5ths of those goals involved darting runs to receive the ball in space or darting runs with the ball into space. I'd suspect those numbers would be even higher if he had been healthy and available for most of the past year. Before I compiled this information, I suspected that Donovan's numbers might not look as impressive when you remove all of the CONCACAF minnows or when you remove his PK goals or set-piece assists. I was wrong. He still carries our team from the run of play against quality opposition. The worry of course is that we're extremely over-reliant on Donovan, and none of us can be certain when he's going to slow down. At some point we're going to have to start preparing for life without him. 4. Have Dempsey crash the box from a slightly deep position (either in the midfield or a bit behind a lead striker) and whip in crosses and good things will happen. Almost all of Dempsey's goals have come this way (either through short or long crosses). Bradley is another good candidate, but this aspect of his game appears to be getting rarer for club and country. 5. Exploit Jozy's strength. A respectable number of our goals against top 50 opposition have come from Jozy holding off defenders. Hopefully these numbers will get even better as he matures as a player. Good target play doesn't have to be ugly, and doesn't have to be incompatible with flowing, attacking soccer. 6. Continue taking shots from outside of the 18. We do it more than I expected, and it tends to work. On the other hand, we don't tend to score this way against top 10 quality teams (Robert Green's flub aside). I don't know if this is because we're not as confident against those teams, or we're not given as much space and time outside of the 18, or if goalkeepers at that level are more consistent. But outside of that range, long distance shooting is a pretty effective tactic for us. 7. Field players who can play defense splitting through balls. It works against all levels of opposition, including the elite (see Feilhaber vs Italy in 2009 or Argentina in 2007, or Bradley vs Brazil in 2012). I don't know if I think Feilhaber or Adu are 90 minute players at this point, but I think both of them are very good with that skill-set. Bradley seems to be developing that skill-set too. Both of the through balls leading to goals that I credited him with came during this recent 5 game tournament. We'll have to see if that's a blip or a more consistent, long term thing going forward. 8. And on that note, our attack benefits from having Bradley on the field. He's our most consistent goalscorer against quality opposition outside of Donovan and Dempsey, he seems to have a habit of making the pass before the pass that leads to a goal, and as I said he seems to now be developing the skill to deliver that killer ball as well. The defensive side might be a different story, but I actually don't think it's a coincidence that our attack has picked up a bit since Bradley's re-introduction into the lineup.
The op and discussion center around the attacking part of the game. A more complete look at value of some of these guys has to take into account the defensive side of the game. Great thread btw. Stuart Holden on corners and Adu on the shorter free kicks are would be a boon on the deal ball situations. I'm not suggesting either needs to in the team anytime soon. It's too bad Brad Davis never quite made it.
Apart from Donovan and Dempsey, the numbers from everyone else are so small that I don't think a coach can make reasonable decisions based on them. For me, this goes back to proving what you can do consistently with your club team. Feilhaber has one goal, one assist in MLS this year. Adu has two goals, one assist. If they're not finding ways to create at that level, should we really expect that they'll produce internationally?
Excellent piece of work however I have 1 minor complaint. Your exclusions seem very arbitrary and somewhat skews the picture. One player that is negatively impacted by the way you chose to shape the data is Kljestan. If you are going to include secondary assists you should also include "chances created", the pass that leads to a shot. Shea, Chandler, and Dolo are negatively impacted by that missing category
I didn't think it was that arbitrary. My goal with this project was to see what works against the type of opposition we're likely to see in the World Cup and the Hex. That tends to be somewhere around the top 50 range. For example, all but 4 of the 32 teams in the past World Cup were within the top 50, and of the 4 that weren't - 1 did not have to qualify and another qualified from a region where they did not have to play a single top 50 team. And the four teams that were competitive in the Hex last cycle were consistently in the top 50 (the other two were not). And I generally don't include January camps when I do these things because the teams we play against don't tend to be representative of their rankings. The same goes with the '09 Gold Cup. Mexico's team was semi-representative but we didn't score any goals against them anyway. As far as secondary assists go, you can ignore those numbers if you like. I've isolated them in that way. Some people tend to find them useful. Some don't. It's your call. And as I said earlier, I haven't included "chances created" because those stats are a bit harder to find across the entirety of the past six years. But if you know a site that's compiled something like that over that entire period, I would be happy to include it.
Except by doing so, you excluded perhaps the most famous goal in US history. Excepting the Barbados game and the one vs. Cuba, it isn't like we light up the score sheet against the "CONCACAF minnows", anyway. Great work nonetheless. It's a lot to digest, but I think you pretty much nail it with the post a few above this one. Obviously, the most striking thing is just how much we rely on Donovan. Impressive from him...a bit scary for us going forward.
It was not clear from your initial post that you had that constraint as the basis for your analysis. I take back my statement that it was arbitrary. So this thread should be "Since the last cycle, how the USMNT scores goals against the ELO top 50, excluding January camps and excluding the 2009 Gold cup. This means from GC2011 you did not count Canads (60), Guadelope (83), or Panama 2 (51). You did count Jamaica (42) and Panama 1 (47). In the recent 5 games you counted Scotland (40), Brazil (2) but not A&B (139) or Guatemala (83). Canada (66) we did not score but would not have been included.