What it takes to win the Champions League

By Jacob Beckett (@jacobbeckett22)

There's something great about knockout tournaments, especially involving teams that are completely unfamiliar with each other. NCAA March Madness and the World Cup are perfect examples; seeing your favorite team play against a relative unknown like Murray St. or Ghana carries a little extra intrigue than your average game against a conference opponent. For MLS fans, CONCACAF Champions League embodies this opportunity.

Increasingly, CCL has been painted as an MLS vs. Liga MX referendum, one in which MLS teams steadily gain but never overtake Mexico's dominant position in the region. But lost in that narrative is that CCL includes teams from a handful of other countries, too. These MLS-Liga MX matchups will get the majority of publicity (starting with Sporting KC-Toluca in the first knockout round), but seven other MLS/Liga MX teams have to knock off Central American or Caribbean opponents before those glamour matchups are set in stone. If you think these first round matchups are just a formality, just ask an FC Dallas fan how their campaign went last year (spoiler alert: they lost to Panamanian side Tauro before they even got to face a Mexican team).

So what kind of chance is there that we'll get four premier MLS vs. Liga MX matchups in the CCL quarterfinals? What's Toronto's Panamanian opponent going to be like in their first ever CCL campaign? How should we expect Houston to perform against a random Guatemalan team that doesn't appear to have existed prior to 2010? These are the questions I'm interested by, but have very little information to answer. I set out to remedy that situation.

First, let's take a look at how history sees the first round matchups we have on deck for 2019 CCL. This chart shows how each league's teams have done in CCL dating back to 2004. For example, the line for Monterrey shows the record for all Mexican teams against Salvadoran opposition in CCL play.

CCL Matchup Teams Games Between Feds. Overall PPG Home GP Home PPG Away GP Away PPG
1 Alianza (SAL) 7 0.29 3 0.67 4 0
Monterrey (MEX) 2.43 4 3 3 1.67
2 Herediano (CRC) 29 1.62 16 2.19 13 0.92
Atlanta United (MLS) 1.21 13 1.85 16 0.69
3 Independiente (PAN) 25 1.04 14 1.71 11 0.18
Toronto (MLS) 1.76 11 2.64 14 1.07
4 Sporting KC (MLS) 62 0.87 27 1.37 35 0.49
Toluca (MEX) 1.89 35 2.29 27 1.37
5 Guastatoya (GUA) 13 0.54 6 1 7 0.14
Houston (MLS) 2.15 7 2.71 6 1.5
6 Saprissa (CRC) 38 0.95 20 1.45 18 0.39
Tigres UANL (MEX) 1.82 18 2.39 20 1.3
7 Atletico Pantoja (CRB) 36 0.75 17 0.82 19 0.68
New York Red Bulls (MLS) 2 19 2.26 17 1.71
8 Marathon (HON) 23 0.83 12 1.5 11 0.09
Santos Laguna (MEX) 2 11 2.82 12 1.25

Unsurprisingly, MLS and Liga MX teams have pretty good historical records against the other CONCACAF leagues. However, Costa Rican teams have been a historically tough out for MLS, indicating that Atlanta might have a rather tough draw against Herediano. 

Historical results (especially those dating back to MLS Era 1.0 and maybe even 0.5) are probably not the best predictor of future results, so I sought out an alternative. Believe it or not, there are not a ton of advanced statistics out there for the domestic leagues of Central America (at least not that I could find). Luckily, there's still a relatively simple way to determine team ratings and even stack up leagues from different countries against one another: Elo.

Like any good writer tasked with explaining some sort of math formula, first I'll link the wikipedia page for anyone looking for details on what Elo ratings are. For those that need a shorter recap, Elo is a simple way of rating teams in which the only input is the historical result of games. The ratings are essentially self-correcting - a team is awarded rating points for winning games (adding more points for beating better opposition), while teams lose rating points for losing games (removing more points for losing to worse opposition). Therefore, teams that win a bunch of games in a row end up with a higher rating and teams who lose a bunch end up with a lower rating.

For my CONCACAF Elo ratings, I began by looking at those last 15 years of CCL games for a baseline of how each league has historically performed against one another. Perhaps as you'd expect, Liga MX and MLS are the top two leagues, particularly in recent years, followed by teams from Panama, Costa Rica & Honduras. Teams from El Salvador and Guatemala are a ways behind, followed by everyone else, who I bucketed together for simplicity (these include teams from Nicaragua, Belize & several Caribbean countries).

The chart below shows the percentage chance a generic CCL team from each league would have against one another, based on the Elo ratings I calculated. Odds vary quite a bit for home vs. away; the way to read this chart is that a generic MLS team has a 58% chance at home against a Liga MX team, but when the Liga MX team is at home they have a 69% chance of winning. One quick note: Elo doesn't exactly predict ties; if a draw occurs, the lower rated team by Elo steals a small number of points from the better team. So the percentages below are actually the chance that a team wins Elo rating points, not the match itself. 

Away Team
Home Team Chance MEX MLS PAN CRC HON SAL GUA OTH
MEX - 69% 83% 84% 87% 92% 93% 98%
MLS 58% - 80% 81% 84% 90% 91% 97%
PAN 39% 44% - 65% 70% 80% 82% 93%
CRC 38% 43% 63% - 70% 79% 82% 93%
HON 32% 37% 57% 58% - 75% 78% 91%
SAL 22% 26% 44% 45% 52% - 68% 86%
GUA 19% 23% 40% 41% 47% 60% - 84%
OTH 7% 9% 18% 19% 23% 33% 37% -
This table shows the percentage chance to gain Elo points, from the perspective of the home team.

Now that we have baseline ratings for every league involved in CCL, the next step is to zoom into each league. Maybe this year's Guastatoya squad is far better than any past Guatemalan qualifier for CCL (they aren't), so looking only at historical results would understate their chances. By tracking the last few seasons' results within each CONCACAF league, we can again develop Elo ratings, this time at the individual team level. I used each league's rating from CCL competition as the starting point for individual team ratings to ensure that the league's relative strength remained a consideration.

The individual team ratings yield the following chances for each team to progress, and eventually, win the 2019 CCL:

Team Quarterfinal Odds Semifinal Odds Final Odds Champion Odds
Alianza 22% 6% 3% 1%
Monterrey 78% 47% 33% 19%
Herediano 48% 22% 13% 6%
Atlanta United 52% 25% 15% 8%
Independiente 34% 7% 1% 0%
Toronto FC 66% 22% 6% 2%
Sporting KC 46% 32% 13% 5%
Toluca 54% 39% 17% 8%
Guastatoya 48% 9% 2% 0%
Houston Dynamo 52% 10% 2% 1%
Saprissa 38% 29% 14% 7%
Tigres 62% 52% 31% 18%
Atletico Pantoja 2% 0% 0% 0%
New York Red Bulls 98% 54% 28% 14%
Marathon 24% 6% 2% 0%
Santos Laguna 76% 40% 21% 11%

There are plenty of caveats that go along with these percentages and they should probably be taken as rough estimates. My ratings didn't account for any domestic cup competitions because that was just too much work to include. My data source was also a touch spotty in some of the Central American leagues, so a game here or there might've been skipped or logged incorrectly. Finally, the ratings are based on games completed through the last full season; teams have changed a lot since then (new jerseys, new coaches, new players). Most of the leagues aside from MLS have also already started play, and none of 2019's results have been considered yet. So if you're looking at Monterrey's chances, it's likely that their rating has only risen given their early hot start in Liga MX.

Now, with all that said: a few potential bullets on each MLS team's potential path to the Concacrown:

  • I'll start with the Red Bulls, whose first round chances are quite possibly inflated thanks to their opposition being pretty unknown. Atletico Pantoja is from the Dominican Republic and qualified by winning the 2018 Caribbean Club Championship. The only other team from the DR to play in CCL was Cibao last year, and they got wiped out 7-0 by Chivas in the round of 16. No Caribbean team has won a CCL game of any kind since 2015, and as far as I can tell, no MLS team has fallen to Caribbean opposition since 2010 when an LA Galaxy lineup featuring Gregg Berhalter and Eddie Lewis lost (at home!) to the now-defunct Puerto Rico Islanders.

  • The flipside of that is Houston, whose chances might be understated since my ratings don't include Open Cup games. Instead, they see the Dynamo as only a mediocre MLS team who failed to make the playoffs. Plus, it appears Guastatoya has two points from their first five games in the Guatemalan Clausura, so Houston should probably feel decent about their chances.

  • Sporting KC against Toluca is the only MLS-Liga MX matchup in the first round, and it looks like one that could potentially go either way. KC is a well-regarded MLS team by ELO (3rd highest in MLS after NY and ATL), while Toluca's results have been pretty meh since topping the 2018 Clausura table a season and a half ago.

  • Atlanta is the highest rated MLS team according to Elo but may still have a fight on their hands with Herediano (see the numbers above regarding MLS vs. Costa Rican opposition). Then again, Herediano has started this season pretty poorly (four points from their first five games), so I'd probably expect Atlanta to advance.

  • Finally, we come to Toronto. TFC actually had the highest Elo rating for any MLS team back in late 2017, but their team has undergone a massive change of fortunes since then. Maybe the best thing going in their favor is that their first opponents, Independiente, are first time CCL participants and don't seem to have any players even on the fringes of the Panamanian national team.

Here's to hoping this article adds the teensiest bit of context to the most mysterious games on your team's schedule this season and may this be the year that MLS wins! 

...or at least racks up enough moral victories that we can keep the 'constant league improvement' narrative going.