The world’s best amateur soccer tournament?

The College Cup concludes this weekend. While the tournament includes only US-based college teams, it is a very strong international tournament in terms of the player pool. 27% of the participants self-identified international hometowns from 55 different countries.  Just look at the hometowns of the final 4 teams:


This final 4 map is also interesting for the insights into the recruiting philosophies of the 4 highly regarded coaches, but that is analysis worthy of a different day.

A question I’ve been stewing on lately: is the College Cup the best amateur international soccer tournament in the world?

The NCAA has gone to very great lengths in its attempts to insure that none of these participating players have ever been directly compensated to play soccer. May tournaments are easy to rule out with millionaire players on the field. Just look at the Olympics.

Where amateurism ends and professionalism begins is a fuzzy line, I would imagine. My understanding is that in England FA players must be at least 17 years old before they can sign a professional contract, although most do not sign professionally until they are 19. Thus, do the U17 World Cup or the UEFA U17 Championship or various Development League tournaments fully qualify as amateur tournaments? And if so, how would those teams compare with the best US collegiate teams that are filled with players in their early 20s, many of whom were graduates of elite youth academies from around the world. Plus for the younger ages, player development is the objective, not results, while the single-elimination college cup is pretty much the most cut-throat tournament format in existence.

In the US, the Professional Development League (PDL) makes allowances to accommodate elite college players, and many of the PDL teams are overflowing with talent. However, the PDL season is very short, with players rotating in and out around their other obligations. Do the PDL play-offs that are more akin to AAU basketball all-star teams have a chance against the collegiate elite?

To me, the College Cups is the tip top of the amateur game, and merits more respet than it often gets. I welcome anyone who can set me straight on this argument that the College Cup is the best and most competitive amateur tournament in the world.

The Road to the D1 College Cup

Like the entirety of the college soccer season, the national tournament has flown by in a whirlwind of travel and games stacked upon games. The 1100 players on the 48 teams who qualified for the 2015 D1 Men’s National Tournament traveled over 70,000 miles trying to reach the College Cup Final 4 held this weekend at the beautiful Sporting Park stadium in Kansas City. Stanford University earned the distinction of traveling the farthest, accumulating over 9000 miles in spite of earning a first round bye and playing their next two games at home.

The official tournament site and bracket are here:

Here is how it all played out geographically:











The primary lesson to be learned from all this travel is that home field advantage is HUGE, HUGE, HUGE. Or it could be that the Selection Committee got the match-ups spot on. Or possibly it’s that the seeded teams earning byes were rewarded handsomely. Or maybe it is that college parity isn’t quite as balanced as many claim. Regardless of the reasons, the home teams won 80% of the games, scored 50% more goals than the away teams, and won almost 2/3 of the overtime games.  15_stats_through_4_rounds

Geography of the 2015 College Cup

The preeminent US men’s college soccer tournament kicks off this week with 48 teams and 1100 players vying for the prestigious distinction of being crowned the National Champion. It is easy to argue that this is the most competitive amateur soccer tournament in the world. Although it consists only of US-based collegiate teams, the teams include players from 55 different countries, so the tournament is truly international in play.

Tournament format

The 48 teams all come from the top tier of collegiate soccer, the National College Athletic Association 1st Division, or NCAA D1 for short. Other divisions in the NCAA, NAIA, NCCAA, and NJCAA have separate national championship tournaments. In D1, 205 total teams compete in 24 separate conferences, with most teams playing between 17-20 games in an intense and exhausting 10-11 week season. How the players are also able to complete full collegiate academic schedules is a remarkable feat of persistence. Side note: if any NCAA officials ever read this, please note my vote as a professor to switch to a 10-month season!

Each of the 24 conferences receives 1 automatic bid to the tournament, with 21 of the conferences awarding the bid to the end-of-season conference tournament winners and 3 conferences awarding the bid to the regular season conference champion. A tournament committee chosen by the NCAA selects the remaining 24 teams via “at-large” bids based mostly on the criteria of win-loss-tie records and strength of schedule (as determined by a Rating Percentage Index).

The tournament is single-elimination format. The top-16 seeds in the tournament receive a bye for the first round. The remaining 32 teams compete in geographically-based games, meaning matches are assigned based on proximity of the country whenever possible.

The official site is

Geography of teams

Here is a map of the tournament teams:


States with multiple teams in the tournament. (Note: tables are screenshots. Didn’t have time to format for web…)


Below are the tournament teams ranked by the proportion of players identifying international hometowns in their rosters. Processing note: the rosters used in this analysis consist of those players with record statistics on the NCAA web-site, and not the often-larger rosters listed on school websites. For example, redshirt players are not included in this analysis.


Here is a map of the number of players with international hometowns. In general, the East Coast schools have more international hometowns. (Note: a much-more extensive analysis of these patterns across all college levels is one of the key objectives of one of my manuscripts in progress).


Geography of the format

The tournament selection committee weights by geographic region, with the intent to assign first round matches in relative proximity. Here is a map of the first round match-ups.


Here are road travel distances for the teams. Some of the teams will likely fly – FIU & Utah Valley, looking at you!


Geography of players

Each player self-identifies their hometowns on their team rosters. These hometowns do not necessarily indicate the players’ nationalities, but there is likely a high correlation. Of the 1100 players with recorded statistics on the NCAA website that are listed on the 48 rosters, 806 (73%) identify hometowns in the US and 294 (27%) identify international hometowns. The NCAA reports that 24% of D1 men’s players are international student-athletes, so international players are slightly more represented in the tournament than across D1 soccer as a whole.


Players with US hometowns:


Players in the tournament with domestic hometowns ranked as a proportion of state population:


International map of hometowns:


Top 20 foreign countries providing players for the tournament:


Although international players are 27% of the total player pool, they scored 38% of the goals and recorded 36% of the assists. They also were more efficient on target, with an overall average shot:goal ratio of 7.6 vs 8.4 for domestic players.


Here are general performance statistics for the top 20 foreign countries. The top value for each category is highlighted in yellow.


Same statistics for tourney players by state. The maximum for each category is highlighted in yellow. This first table is for states with at least 10 players:


And for states with less than 10 players:


Best wishes to all the teams and players for an excellent tournament! And as a professor, I also wish the players the best of luck completing your school work during this rush😉


NCAA D1 Men’s Soccer Player Home States, 2015-2016

Context: As part of a larger study on the internationalization of US College men’s soccer I have compiled a rather large database of player rosters and performance statistics. When time is available, I hope to post some tangential analysis from this data set. This first part is a geographic analysis of home states of players in NCAA D1 Men’s Soccer for the 2015-2016 season. Other NCAA divisions, NAIA, and NJCAA are also in the works, including historic data and sub-state geographies (hopefully at least metro areas). I also hope to break down positions and performance statistics (goals, etc) by state and spatially quantify the recruiting markets of teams. International player data will be dumped here after some manuscripts are (hopefully…) published. All in good time.

Methods: Data were assimilated from each team’s website and processed using a combination of Excel and python scripting (thanks in particular to the Beautiful Soup and GeoPy Python libraries). Spatial analysis and mapping use ArcGIS. Statistical analysis and visualization use Excel and Tableau. (Eventually I’ll spend the time to get those beautiful interactive Tableau charts embedded, but I’m not there yet…)

Caveats: There are probably some players who were not on the web-site rosters at the time of web-scraping (approximately Aug 15-Sep 5, 2015), but this is probably a small number. Hometowns are those self-provided by players for their team rosters, so may or may not match where players went to high school, for example. Players who self-report international hometowns are excluded from this data set.

Side note: I’d love to do an analysis by youth clubs, but college teams are inconsistent with reporting this information on the roster pages, so club data are incomplete. Similarly, I very regrettably did not have time to collect college women’s soccer roster data.

Below are maps and charts of the distribution of self-identified home states for 4,435 players in NCAA D1 Men’s soccer. All states supply players except for South Dakota. California supplies the most players (546), but Connecticut has the most number of players per 100,000 population (3.86) . Geographically, the most over-represented states are along the eastern seaboard, and the most under-represented are in the Deep South and Upper Great Plains and Upper Intermontane West regions.







Excel File of Raw data: NCAA_D1_mens_soccer_players_by_state_2015.xlsx

Collegiate players at the World Cups

Top Drawer Soccer analyzed the rosters of the 2015 Women’s World Cup (article is behind paywall) and calculated that 95 women from 15 of the 24 teams played college soccer in the US. I won’t recreate that list here out of respect for their paywall. The AP also reported on some of the stories of these international collegiate players.

In comparison — from the best I can tell on quick search — 12 players at the 2014 Men’s World Cup played college soccer for 2 different teams. A few players transferred schools during their careers, but only the most recent school is listed here:

  • Akron: DeAndre Yedlin (USA)
  • Boston College: Alejandro Bedoya (USA)
  • Cal State Chico: Chris Wondolowski (USA)
  • Furman: Clint Dempsey (USA)
  • Maryland: Omar Gonzalez (USA), Graham Zusi (USA)
  • Notre Dame: Matt Besler (USA)
  • Rhode Island: Geoff Cameron (USA)
  • South Carolina: Brad Guzan (USA)
  • St. Louis: Brad Davis (USA), Vedad Ibišević (Bosnia & Herzegovina)
  • UCLA: Nick Rimando (USA)

Here is the comparison in map form:


Football stadiums around the world

I recently stumbled across this Google KML layer of football stadiums around the world. While Google Maps & Google Earth are unparalleled for browsing, they can be challenging for seeing trends. I was unable to determine who the creators of that data set were (thanks for the remarkable effort to whoever you are!), and I have absolutely no idea on the accuracy or completeness of this data set, but here is a map of 5340 football stadiums: