How Long Does It Take a Team to Mesh? by Kevin Minkus

By Kevin Minkus (@kevinminkus)

While beginning a season 0-3-0 does not a happy fan base make, Sunday's win over Philadelphia has some Chicago Fire fans feeling at least a little better about the team's rebuilding process. Throughout the beginning of the season, coach Frank Yallop has frequently stressed that the team needs time to adjust to each other. After all, they brought in three new designated players during the off-season, and are returning players who accounted for only 63% of last year's minutes (the league average over the last four seasons is around 71%). It should take a while for all of those new pieces to mesh from the somewhat disjointed side we've seen into a coherent whole. But, given the Fire's level of roster turnover, how long should we expect the meshing process to take?

The term “meshing” is a slippery one, and can be defined in any number of ways.  Is it when a team's roster turnover no longer informs its results? Is it when a team's results sufficiently indicate its performance for the rest of the season? Is it when a team reaches the level of performance it will remain at throughout the rest of the season (if, in fact, a team can ever be expected to do so)?

Each of these definitions could be argued as valid, and I'm sure there are many other possible definitions not considered here. As it stands, though, these are the three I will analyze, using MLS data since 2011, in hopes of arriving at an answer to the question of how long it takes a team to mesh.

Let's start with the first definition- meshing defined as the number of games in which roster turnover still directly informs a team's results. 

This graph shows the correlation between points after x number of games and the percentage of a team's field minutes returned from the previous season. 

A positive correlation suggests that as roster stability increases, so does points earned. Numbers below the red line are not considered statistically different from zero (at 90% confidence). Note that the correlations in general aren't huge, but they do exist. As you can see, the correlation between roster stability and points peaks at game three, and remains statistically significant until game five (after which it remains insignificant until close to the end of the season).

A similar pattern exists if we look at defensive stability, though the correlation becomes doesn't become insignificant until after 8 games:

These two graphs, then, suggest (though perhaps not convincingly), that it may take as few as three or four games for a team in general to mesh, while it may take as many as eight for a defensive unit to come together.

Now let's take a look at the second definition- meshing defined as the point at which a team's results through some number of games “sufficiently” indicate what its results will look like for the rest of the season.

To do this, I've split teams into two groups- those with “high” roster turnover (in the top 50%), and those with “low” roster turnover (in the bottom 50%). I then regressed the team's final points total on the team's points total after x games, for each of the two groups. The Rsquared values for each of these regressions are graphed below, with the linear models from the set of all teams included as well. So essentially what we are looking at it is how well we can predict how a team will finish the season, based on what they've done after a given number of games.

Through six games, each game is about as predictive for each group, meaning that how well a team with high roster turnover does through six games is just as indicative of how that team will finish as how well a team with low roster turnover does through six games. That is to say, we don't gain any extra predictive power by knowing a team's level of roster turnover.

By game seven, though, high turnover teams begin to out-pace low turnover teams- by game seven we have a better idea of how high turnover teams will finish the season than low turnover teams. 

By game nine, the R2  value for high turnover teams is at .546, which is pretty high. We would expect predictions made using this nine game point total to be on average only about seven points off the final season total. That gets us pretty close for being barely a quarter of the way into the season.

 Though it's a normative statement not a positive one, and you could really draw the line anywhere, I would probably suggest that nine games is as good a place as any to set the limit on meshing based on our second definition. At the very least, we can say that after nine games we should have a decent idea of whether the rebuilding process will be successful in year one.

Finally, let's turn our attention to the third definition- meshing as the point at which a team reaches its consistent level of performance.

Let's investigate this phenomenon a little bit. 

Here's a graph of the three game rolling expected goal difference (at x = 4, the value on the y axis is the xGD from games two, three, and four, for example) for Sporting Kansas City last season- a decently representative mid-table team.  Expected goal differences provide a pretty reasonable statistic for gauging how good a team is.

It's pretty much all over the place. 

A three game rolling points per game graph of another mid-table team from last year, the Vancouver Whitecaps, tells a similar story:

These graphs point to something which I think is an important (though perhaps obvious) point to make; it's mostly unreasonable to expect game by game measures of a team's strength to converge over the course of a season. (Metrics like xGR (expected goal ratio), TSR (total shot ratio), and points per game will converge, but usually only when they're being calculated on aggregate.) There are a lot of reasons for this. Injuries, international call-ups, strength of schedule, and mid-season transfers are all factors which affect a team's consistency of performance. Teams, save maybe the very dominant and the very bad ones, just go through peaks and valleys throughout the year. They have good games and bad games. 

What does this mean for meshing, then?

Well, we've already seen that how a team performs at the start of the year can be predictive of where it finishes, particularly for teams with high turnover. The point above, though, suggests that how a team starts the year isn't necessarily indicative of how it will perform throughout the year. 

For teams who haven't quite come together yet, then, there is certainly still hope of righting the ship. Given the above analysis, I would expect the effects of having new players brought in to the system to begin to wear off by game four or five (though this may take a bit longer this season because of international call-ups). By game nine or ten, a team should have a decent idea of how well it has done in rebuilding its roster. If things remain bleak at that point, there is still the possibility of finding some success, but it may come only in limited doses.

Location Adjusted Total Shots Ratio by Drew Olsen

Millionaire Malcolm Forbes was famous for his quote, "He who dies with the most toys wins." And while that might not be the most moral mantra for life, sports fans have a hard time arguing with the logic. After all, a game is about runs, points or goals, and after enough of those it's about shiny trophy cases. But in the world of sports analysis there is no such victory in the absolute. Analysts need to explain how those runs, points or goals came about. In the world of soccer especially, there is never a complete answer. Goals are exceedingly rare, so explaining how they grace us with their presence mathematically is difficult, to say the least. We're happy with higher R-squareds and other such geeky descriptive metrics. Have you ever seen a trophy case filled with strong correlations? Nope, all we get is a little blog post, and if we're lucky, some twitter praise. Still, we search.... One of the more popular explanations for winning in soccer is Total Shots Ratio, which calculates the percentage of shots taken by a team in games played by that team. A 60% TSR means that a given team took 60% of the total shots fired in the games they played. The logic isn't all that difficult to wrap your head around. If you can take more shots than your opponent you are likely to score more goals. For the English Premier League, TSR explains 68% of the variance in the point table, which is impressive for one statistic. TSR happens to be less important in MLS.

data sources: AmericanSoccerAnalysis, mlssoccer.com

The variance prediction is just 37% and this is likely due to the lower finishing rates in MLS compared to the EPL, rendering shots less effective. But there are probably a number of other reasons why TSR is less predictive of points in MLS. There are a larger percentage of teams employing counterattack strategies which have significant impacts on finishing rates, which would in turn alter the effectiveness of TSR. But what if the shots were weighted to account for the location of the shots? It would be logical to assume that better teams take better shots and make it more difficult on the opposing shooters. But does that logic actually manifest itself when predicting points? ASA's Expected Goals 1.0 worked pretty well, so a TSR adjusted for shot locations ought to work better than the original TSR.

The first thing required would be a fair weighting of shots by location. To do that I took the ratio of the finishing rate by location and divided by the average finishing rate. Here is the resulting table for adjusting the value of shots.

Location Weighting
1 3.14
2 1.79
3 0.72
4 0.54
5 0.24

For the sake of simplicity I have collapsed zones 5 & 6 into a fifth zone. This table illustrates that a shot from zone 1--inside the 6-yard box--is actually worth 3.14 average shots. And a shot from zone 5 is worth just .24 average shots. Adjusting all of the shots in MLS in 2013 yields the following result when attempting to predict table points.

data sources: AmericanSoccerAnalysis, mlssoccer.com

You can tell from just eyeballing the dispersion of the data points that the location adjusted TSR better aligns with points and the Rsquared agrees. There is a 17-percent increase in R-squared. Not just pure volume of shots, but the location of those shots is vital to predicting points in MLS. It would be interesting to see if location is equally important in the EPL, where TSR is already such a strong predictor.

For the curious, the New York Red Bulls were the team that was best at getting better shots than their opponent. Their TSR improved from 47% to 52% when adjusting for shot location. Real Salt Lake actually took the biggest hit. Their TSR was 53% and their location-adjusted TSR dropped to 48%.

It's only one season's worth of data, but with such an impressive increase in the ability to explain the variance in point totals, it confirms that location does matter, and that teams are rewarded by taking better shots themselves while pushing their opponents -out farther from goal. And perhaps soccer analysts have another statistical toy to add to the toy box---Location-Adjusted Total Shot Ratio.