The analytics movement is working its way into college soccer – this level of the game presents a new and different landscape to investigate. It is a different game in many ways (for better or worse), and much has been made of its place in the American soccer development system. Adding insight from this landscape to the overall analytics discourse is my main goal, and any feedback or further discussion is more than welcome.
I look forward to diving into this. Thanks again for your interest – let’s get to it.
Alex updated his chart with an MLS-only version a few months into the 2015 season. The addition of a “Possessions per Game” column led me to think about what this may look like in college soccer. A popular perception of the college game is that it relies more on speed and athleticism, and less on skill. What would a similar chart from college soccer say about this notion?
Before looking at the college numbers, it may be helpful to represent Alex’s original chart in another type of visualization. Unfortunately, I have not seen an updated version of the data, but this still works for the purposes of this comparison:
Based on the points per game at this stage of the season, it appears that passing style had little influence on a team’s place in the table. Taking a look at the teams who later made the playoffs, the same idea holds. The averages for this group do not differ greatly from the league averages… and the teams are widely scattered:
The biggest comparison that jumps out from these charts – there is no overlap between the college and MLS numbers. The ACC charts are on a completely different part of each axis. This means the following:
No team in the ACC played as many passes per possession as the lowest MLS team. Furthermore, teams had at least 50 more possessions per game in ACC games than in MLS games. This suggests that the college game is a much different style than MLS.
Similarly to MLS though, passing style seems to have no major impact on success in the college game – for both conference points per game and final results for the season. In 2015, for example, Wake Forest passed the ball a lot, and spent a large portion of the season as the #1 team in the country. Clemson were a bit closer to the conference average, and rode that balance to the College Cup. Syracuse were at the complete opposite end of the spectrum from Wake Forest, playing a much more hectic style – and were right there in the College Cup with Clemson.
All of this suggests that at a team level, these passing style numbers indicate just that – style, more than substance. Perhaps you can’t predict how well a team may perform based on their passes per possession, but you can learn a great deal about how they will play. For those of us whose job entails a great deal of opposition scouting, this metric can help us determine whether to prepare for tiki-taka, counterattacking football, or mortar shells being launched down the field.
But why is there such a difference in the passing numbers between college and MLS?
A college soccer game differs from an MLS game in a few easy to pick out ways – the clock counts down from 45:00 and can be stopped by the referee, players who leave the match in either half can return in the second half, and the season runs only from August to December, at its very longest. Finally, logic tells us that the players at the professional level likely exhibit better quality than their college counterparts.
To further investigate how these factors play into the drop off from MLS to the ACC, I plotted the average number of passes attempted by a team in each 15 minute period of a match. Data used for this chart came from the same ACC dataset, plus WhoScored.com’s detailed MLS stats. The results look like this:
The college line largely mirrors that of MLS, particularly in the first half, with the ACC teams keeping within 4-7 passes of the MLS teams. This can presumably be explained by the increase in quality of player from college to professional. But what happens to the college players at halftime?
In the first 15 minutes of the second half, the gap between MLS teams and ACC teams widens to 16 passes, and never really recovers. Something in the game has clearly changed at this point. Looking at the box score or play-by-play of a college soccer game, things often get a bit chaotic in the second half in terms of personnel. NCAA substitution rules allow players who leave in the first half to return after halftime, and players who leave in the second half to re-enter once more. Perhaps the introduction of several substitutes, and subsequent re-introduction of starters, actually serves to hurt the speed and quality of play?
Research by Conte et al. found that men’s Division I soccer games featured substitutions at a rate 375 percent greater than that of professional leagues – the standard three permitted substitutions per game. Furthermore, 60 percent of these substitutions were re-entries - an action not permitted in professional soccer – and 90 percent of substitutions occurred after the 30th minute.
A frequently seen substitution pattern in college games involves midfielders and forwards coming out late in the first half, then returning either at halftime or 10-15 minutes into the second half. This is the same time period in which the drastic drop off in passing numbers occurs. Are the substitutes not quite as adept at passing as the starters? This is a possibility – after all, the starting eleven are starters for a reason. Are players re-entering the game actually a bit removed from the flow of the game once they return? Whereas the majority of players in a professional game remain on the field for the full 90, thus remaining entrenched in the match, perhaps college players lose a bit of their grip on the game when they leave and come back.
Conte et al. present another potential influence of substitutions. The huge increase in subs is to be expected not only due to the allowance of such moves by the rules, but based on the packed schedule of NCAA soccer. Teams play upwards of 20 to 25 games in a short time span, with most weeks featuring two games. For a player to play 90 minutes midweek, plus 90 more minutes on the weekend – sometimes for several weeks in a row – is a very grueling task. It is likely that coaches turn to substitutes more frequently than they may prefer simply to keep their top players available for the next match. Additionally, teams often spend the days between these games recovering and regenerating, rather than working on the passing that we see them struggle with in the second half.
All of this raises a potential new argument for one of the hot topics in college soccer at the moment – the extension of the NCAA season to last the full academic year. Many supporters of the change tout the reduced time demands on athletes on a weekly basis, as well as a reduced risk of injury with less frequent games. Could another result of spacing out the season actually be higher quality soccer?
Players able to focus all of their energy on one game each week, rather than two, would likely perform at higher levels. Plus, if teams become less reliant on substitutions – whether the rule changes to eliminate re-entry, or coaches no longer feel the need to rest their top players – then presumably the passing numbers we discussed may improve. Though they may never reach the same level as MLS, we could feasibly reduce or eliminate the big drop off seen after halftime of college matches.
College soccer’s position in the American development system has come under scrutiny for a perceived lack of quality – many feel that by the age of 18-22 players should either be in a professional environment already, or they are not going to be good enough. But data such as these passing style statistics suggest that perhaps we are doing the college players a bit of an injustice. Over the first half of games, they DO have a level of quality approaching that of their MLS counterparts. Are we actually derailing their progress somewhere around halftime through substitutions and re-entries? Could we reduce this interference with their play (and potentially their development) by changing the NCAA schedule? Further investigation into this possibility may help college soccer improve its level of play, the time demands on its athletes, and its place in the development of American players.
Click here to view the Tableau Public workbook, which will allow you to play around with the visuals shown above. If you're unfamiliar with vizzes in Tableau, try hovering your mouse over graphics to see additional information.