Home > Confederations Cup 2009, Tournament Statistics > Is the U.S. Really Close or Really Lucky?

Is the U.S. Really Close or Really Lucky?

Greg Lalas, editor of goal.com, wrote recently that “if the U.S. showed anything against Spain in the semifinals, it’s that the time is coming very soon when they no longer merely endure against the superpowers, but actually prevail.”  And Jen Chang recently wrote in his fantastic blog that the U.S. proved in its match against Brazil that the victory over Spain was no fluke.  And while I would like to accept that the U.S. is now a substantially better team than it was even a couple weeks ago, the numbers made me wonder how big a role luck played in our success.

For example, the U.S. shot incredibly well in the last couple matches.  Ridiculously well.  In Stage 2 (the semi-finals and finals), the U.S. scored goals on 67% of its shots on goal and 24% of its total shots.  Those numbers are out of this world.  Spain, no offensive slouch, scored on only 19% of its shots [on goal] and 9% of its shots on goal.  And Brazil?  21% of its shots on goal and 9% of its shots.

In fact, if we look at the top five non-U.S. teams in the Confederations Cup, their statistics are sobering for U.S. fans:  Brazil, Spain, South Africa, Italy, and Egypt combined for 36 goals on 371 shots (150 on goal).  That is a rate of 24% of shots on goal and 9.7% of their total shots.  This is consistent with academic studies that find, on average, 10 shots for each goal scored.  (Pollard, Ensum, and Taylor 2004)

The U.S.’s scoring rate in Stage 2 was 2.4x the historical average for soccer.  As good as our team is, it is hard to believe that our attacking players are so substantially better than their counterparts on Spain, Brazil, South Africa, Egypt, and Italy.  So, I wondered what if the U.S.’s scoring rate had been consistent with the historical and tournament average?  How likely is it that we would have beaten Spain?


Here’s one way to try it:

I simulated the results of 100 matches using randomly generated numbers (ranging from 0-99) to represent shots taken.  For each shot the U.S. took against Spain (8), it received one randomly generated number.  If that number was less than or equal to 23, the U.S. scored a goal on that ‘shot’.  (Consistent with its 24% scoring rate on shots taken).  Spain was given 19 numbers for each match (19 shots taken against the U.S.) and we used its 9% scoring rate for Stage 2, so any time its number was <= 8, Spain scored a goal.

In 100 matches, even with the U.S. shooting 2.4x the historic average for soccer, the U.S. beat Spain only 40 times.  There were 26 draws (which would have been settled by stoppage time or a penalty kick shootout).  And Spain beat the U.S. 34 times.

We also simulated 100 matches, but this time we reduced the US.’s shooting to the historic average (10%) and kept Spain’s the same (9%).  In those 100 matches, the U.S. won 17 times, there were 18 draws, and Spain won 65 times.


Ultimately, even if the U.S. scores at the exceptional rate of almost 1 goal for every 4 shots taken, it still only beats Spain in 54% of matches decided by the end of regulation.  And if the U.S. returns to the traditional rate of 1 goal for every 10 shots taken, it beats Spain in only 21% of matches decided by the end of regulation.

On the plus side, the U.S. appears to have increased its chances of getting points from its matches with quality teams.  Were we paired with Spain in the Group Stage of the World Cup, we’d have a 35% chance of getting at least one point from the match.  And if we rose to the event and performed exceptionally well, our chances of a point climb to 66%.

Ultimately, as much progress as the U.S. has made, it still cannot afford to take half as many shots on goal as a high-quality opponent.  Even discounting our sample, research has concluded that Shots on Goal Differential “is the most important variable to decide the outcome of a match in a tournament.”  (Papahristodoulou 2007)

  1. Craig
    July 1, 2009 at 12:44 pm

    I would be very interested in seeing finer-grained statistics on the success of shots on goal… things like angle, distance, defenders involved. I doubt it significantly diminishes the degree of luck involved in the U.S. victory over Spain or performance vs. Brazil, but my feeling is that the first chance against Spain and the second against Brazil were both relatively high-percentage chances. (Then again, the second against Spain and the first against Brazil did not appear to be high percentage chances, and we still made goals out of them.)

    • Michael
      July 1, 2009 at 2:12 pm

      Check out the link in the post to the Pollard et al. article. It analyzes many of those things (i.e., angle, distance from the goal — e.g., concluding that each yard further away from the goal, decreases the odds of scoring by 15%).

      Prior to posting I had thought about how good a couple U.S. chances were — you point to two virtually sure things. But with so many statistics all coming out about the same (1 goal per 10 shots historically and 1 goal per 9.7 shots in the Confederations Cup), I went ahead with it.

      Because shots on goal are higher-percentage shots, I considered using them rather than pure shots in the simulation. I opted not to because the U.S. had so few (2 versus Spain).

      USSoccer.com does provide a ProZone tracking system for MNT matches that identifies shots with a great deal of specificity — where it was taken, whether it went high, etc. When I have a chance, I’ll work with those and post what I find.

  2. wjmooner
    July 2, 2009 at 10:22 am

    Have you seen any breakdown of shots by half from the Confed Cup, or perhaps even shots before and after the USA’s goals? If I recall they took 9 shots in both games, but it sure felt like 6 or 7 of those came in the first half of each game. If we hadn’t gotten the lead both games, I’m sure we attack more and end up with more shots. We definitely went into a shell against Spain and Brazil. Not so much against Egypt since we knew we still needed a third goal for most of the second half.

    • Michael
      July 2, 2009 at 4:01 pm

      I’ll see if I can find that. I’m not sure we would have taken that many more shots regardless: only 10 (5 on goal) against Italy and 9 (2) on goal in the group-stage Brazil match.

  1. June 26, 2010 at 6:47 pm
  2. July 15, 2010 at 1:04 am

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: