Chivas USA: Taking Inventory Of 2010′s Formations

Posted In Features - By Richard Farley On Wednesday, March 2nd, 2011 With 2 Comments

In mid-December, Kevin McCauley and I started a site. Then, I got busy. Again, though I feel OK about it. I posted twice. Kevin never posted at all. Thanks, buddy.

But before I aborted that project, I was able to start (what was supposed to be) a series of posts which examined how teams set-up in 2010, attempting to bring something to Major League Soccer blogging that’s enjoyed by almost every major league around the world. Rather than basking in the (perversely proud) ignorance of blind tribalism, we wanted to give our domestic league the same respect we see writers across the world give their favorite competitions.

With such lofty ambitions, I was ripe to be shown a hypocrite, though nobody ever called me on it. That’s because nobody ever read the site. I only posted there twice before completely disrespecting my idea, finding excuses (“but this job paaaaaaaaaaaaaays (almost nothing)”) to walk away.

This is my confessional, my apology to myself, my pittance for being so arrogant. More than that, I’m still sorry I’m not finding time to write more pieces like this. I feel like I’m not treating my domestically league fairly (not that it cares).

This was the second in that series. We’ll move the first over some time soon.

Yesterday I looked at the different formations Carlos del los Cobos played with Chicago during Major League Soccer’s 2010 season. Today, it’s Chivas USA’s turn.

The reference points are largely the same. I want to see what Vazquez’s preferred set-ups were, and to do this I’m relying heavily on Football Lineups. It’s a great site with a ton of free information. I can only imagine how much dedication its volunteers must have, though today I’m a little more skeptical of some of the information they’ve cataloged. I’ll get into that in a minute, as Football Lineups still gives us more than enough data to build on.

Before I break out the tables, here are the abstract-esque observations I can make:

  1. Chivas USA was a much better team playing 4-4-2 than the more often used 4-2-3-1.
  2. The 4-2-3-1 probably should have been abandoned once Sacha Kljestan was sold.
  3. The team performed better with Justin Braun as part of a forward partnership (as opposed to playing as a lone striker).
  4. Chivas was better with Jonathan Bornstein in the team and on the left side, though not necessarily at left back.
  5. Martín Vazquez never found a good deep midfield alignment.

We’re going to look at general (three band) and more specific (four band) formations, break down striking tandems as well as choices at the creative and defensive midfield positions. However, adding a twist to yesterday’s reference points, I’m going to pay special attention to how Martín Vazquez used Jonathan Bornstein and (while he was there) Sacha Kljestan. Those were Chivas’s two most recognizable players coming into the 2010 season. Some would argue they were Vazquez’s best, but with each playing in multiple spots (as well as playing very little), I wanted to see if we can see any relationship between usage and results.

Again, let’s start general. In this first table, I’ve crammed each of Martín Vazquez’s formations into either 4-4-2 or 4-5-1. Like Chicago, Chivas played a mix of each, with one formation being clearly more successful. For reference, Chivas USA averaged 0.93 points per game in 2010.

Formation Times Started Ratio Record (W-L-D) Points Per Game Difference
4-5-1 17 56.7% 3-13-1 0.59 -37%
4-4-2 13 43.3% 5-5-3 1.38 +48%
Table 1 - Chivas USA formation frequency, results

Those numbers lean heavily toward the 4-4-2′s relative success, though this is where a caveat about Football Lineup’s recording needs to be mentioned. In four matches near the season’s end, Football Lineups has either Alan Gordon or Giancarlo Maldonado playing an central, attacking midfielder in a 4-2-3-1. Those matches are:

  1. September 10, versus New England Revolution (Gordon),
  2. September 19, versus Kansas City Wizards (Gordon),
  3. September 25, at Phildelphia Union (Gordon), and
  4. October 3, versus Los Angeles Galaxy (Maldonado).

Alan Gordon is not going to draw many comparisons to Wesley Sneijder or Mesüt Özil. That’s the main reason for this aside. While I’ve kept those four matches in the 4-2-3-1 set, the smell test hints those may be instances of a 4-4-1-1, a formation that I would clump in with the 4-4-2 results. It would be interesting to review those matches and see if Gordon’s (and Maldonado’s) positioning and defensive responsibilities make Chivas USA’s formation more 4-2-3-1 or 4-4-1-1. For now, I will trust the person who’s taken the time to update Football Lineups.

The distinction between a 4-4-1-1 and a 4-2-3-1 is one example of why not all 4-5-1s and 4-4-2s are made the same. Here is a more detailed breakdown of Chivas USA’s formations, extending them out to four bands where necessary. As you can see, 4-2-3-1 was Vazquez’s most used approach:

Formation Times Started Ratio Record (W-L-D) Points Per Game Difference
4-2-3-1 14 47.7% 3-10-1 0.71 -23%
4-4-2 10 33.3% 4-4-2 1.40 +51%
4-4-1-1 2 6.7% 1-0-1 1.33 +43%
4-5-1 2 6.7% 0-2-0 0.00 -100%
4-1-4-1 1 3.3% 0-1-0 0.00 -100%
4-1-3-2 1 3.3% 0-1-0 0.00 -100%
Table 2 - Chivas USA detailed formation frequency, results

Overall, there was a huge difference in Chivas USA’s performance when playing with two forwards. Referring to Table 1, they saw a 85 percent improvement in points per game when using more than one striker. The results with a lone striker, pro-rated to a season’s worth of games, would have had Chivas USA finish behind DC United.

The other thing to notice is how much better Chivas was with a a “flat” midfield. I put flat in quotes because we’re talking about relative flatness. Some players are always more defensive and deployed deeper than others. The same could be said for systems, though, and when Martín Vazquez used a approaches like 4-4-2 or 4-4-1-1 - systems that do not inherently have two bands in midfield - Chivas was more successful.

That dichotomy’s highlighted in the next chart. In Table 3, we look at all the combinations Vazquez used in deep sitting midfield positions. We also note the club’s performance when Vazquez deployed formations that didn’t specifically call for that role.

Deep Midfield Times Started Ratio Record (W-L-D) Points Per Game Difference
None 12 40.0% 5-4-3 1.50 +61%
Blair Gavin, Ben Zemanski 6 20.0% 2-3-1 1.17 +25%
Ben Zemanski, Paulo Nagamura 3 10.0% 1-2-0 1.00 +8%
Other combinations used: 2 times - Rodolfo Saragosa/Zemanski, 1 time - Marcelo Saragosa, Gavin/Saragosa, Gavin, Michael Lahoud/Gavin, Nagamura/Lahoud, Rodolfo Espinoza/Nagamura, Saragosa/Nagamura
Table 3 - Chivas USA deep midfield frequency, results

While the Blair Gavin, Ben Zemanski partnership was not bad, the team was still better off when they didn’t break their midfielders up into two bands, though the causation is less clear. It’s possible the team doesn’t have players who are adept at playing this role, making the team weaker when those players are called on. It’s also possible that the team didn’t execute the formations well, and given using multiple deep lying midfielders often means sacrificing a forward, perhaps this was a trade-off Chivas USA should not have been making.

One piece of evidence describing that trade-off is the breakdown and success of Vazquez’s forward deployment. Starting Justin Braun alone was Chivas USA’s preferred set-up, but looking at Chivas’s record when Braun played along side Maldonado or Jesus Padilla, sacrificing a striker looks less viable.

Forwards Times Started Ratio Record (W-L-D) Points Per Game Difference
Justin Braun 12 40.0% 3-8-1 0.83 -10%
Justin Braun, Gianluca Maldonado 4 13.3% 2-1-1 1.75 +88%
Jesus Padilla, Justin Braun 3 10.0% 1-1-1 1.33 +43%
Other deployments: 2 starts - Braun/Gordon, Maicon Santos, Chukwudi Chijindu, Maykel Galindo, Gordon, 1 start - Santos, Maldonado/Gordon, Padilla/Maldonado
Table 4 - Chivas USA forwards frequency, results

To illustrate the point about Braun, here are the matches where Braun started alone juxtaposed against ones where he was part of a tandem. The ratio and difference columns have been adjusted to only take into account games Braun started at forward.

Justin Braun, Forward Times Started Ratio Record (W-L-D) Points Per Game Difference
Starting Alone 12 57.1% 3-8-1 0.83 -30%
Part Of A Tandem 9 42.8% 4-2-3 1.67 +40%
Table 5 - Justin Braun at forward frequency, results

Chivas USA lost 18 matches in 2010 but only twice when Justin Braun started as part of a striking tandem. Had Braun only been part of a two forward set a handful of times, that statement might be empty; however, Braun was half of a forward tandem nine times.

One reason the two striker set may have been more successful is the loss of Sacha Kljestan. Chivas USA’s best player coming into the season, the attacking midfielder was sold to Belgium’s Anderlecht shortly after the start of the season. Perhaps Chivas was using tactics driven today’s Kljestan’s presence, failing to adjust quick enough to life without him.

Kljestan would start only ten matches for Chivas, with the club performing slightly, insignificantly better with him in the team:

Sacha Kljestan Starts Times Started Ratio Record (W-L-D) Points Per Game Difference
Did Not Start 20 66.7% 5-12-3 0.90 -3%
Started 10 33.3% 3-6-1 1.00 +8%
Table 6 - Sacha Kljestan starts frequency, results

Kljestan’s absence is more evident when looking at Chivas USA’s performance when starting 4-2-3-1. Despite only playing 10 times for Chivas in 2010, he was still Vazquez’s choice central, attacking midfielder, starting twice as many times at that position than any other player.

Central, Attacking Midfielder Times Started Ratio Record (W-L-D) Points Per Game Difference
Sacha Kljestan 6 20.0% 2-3-1 1.17 +25%
Others 10 10.0% 1-9-0 0.30 -68%
Other deployments: 3 starts - Alan Gordan, 2 starts - Blair Gavin, Osael Romero, 1 start - Ben Zemanski, Giancarlo Maldonado, Michael Lahoud.
Table 7 - Chivas USA central, attacking midfield (4-2-3-1) frequency, results

The Football Lineups caveat applies here. Four starts amongst the “Others” deserve questions. Without those four included in the sample, Chivas USA went 0-6-0 playing 4-2-3-1 without Sacha Kljestan.

One final, interesting story line from Chivas USA’s 2010 season was the use of Jonathan Bornstein. Bornstein came into the season best known as a left back, but Martín Vazquez ended up starting him in four different positions.

Jonathan Bornstein Starts, by Position Times Started Ratio Record (W-L-D) Points Per Game Difference
None 13 43.3% 3-8-2 0.62 -34%
Left back 10 30.0% 3-6-1 1.00 +8%
Left midfield 4 13.3% 2-1-1 1.75 +88%
Center half 2 6.7% 0-2-0 0.00 -100%
Right back 1 3.3% 0-1-0 0.00 -100%
Table 8 - Jonathan Bornstein starts by position frequency, results

Chivas was a point per game team with Bornstein starting. If you take has three starts in the middle and on the right of defense out of the equation, Chivas USA’s point ratio with Bornstein starting goes up to 1.21, or a 31 percent improvement on their normal performance.

Overall, Chivas was better when they used a simpler set-up in midfield and employed two strikers. Much of that may be owed to losing Kljestan, with the team performing very poorly in the 4-2-3-1 without him. Chivas USA may have had a better season had they gone exclusively 4-4-2 after Kljestan’s departure.

About Richard Farley -

Displaying 2 Comments
Have Your Say

  1. [...] Richard Farley: Chivas USA: Taking Inventory Of 2010′s Formations [...]

  2. baumer says:

    Really enjoyed this analysis. Hope they keep coming. One thought is that the formations themselves may have been dictated by the opposition. Meaning that Chivas may have been dropping an extra midfielder deep against stronger opponents and simply the superiority of the opposition may explain or at least influence the success of that setup. Have you considered away to normalize the results based on the strength of the opposition? Perhaps this is going too far in a league that has so many mechanisms to induce parity.

Leave a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>