How I Learned to Stop Worrying and Love Quakes Away Games

Luchi Gonzalez has the Earthquakes in 5th place in the Western Conference just a bit over one-third into the season. Photo credit: Sara Leyva for Sara Nicole Mindful Photography and Quakes Epicenter.

Special acknowledgment to Quakes Epicenter patron Trevor Wojcik, a data scientist in the economics field, for constructing a season simulation technique for me in the preseason. I was able to use Trevor’s insightful work to extend to an in-progress season simulator.

Thanks to a new playoff format from MLS, Luchi Gonzalez's comments from his first press conference that "playoffs are a must" for his first year as coach, which admittedly seemed like a stretch at the time, now feels like, "that's it? just make the playoffs?". In case you've missed it, in the new MLS playoff format, nine teams from each conference will "make the playoffs" with the eight and nine spots in the West vying for the opportunity to play LAF...er...the top seed via a one-game playoff. Just squeaking into the playoffs feels like a really, really low bar.

Within the Quakes Front Office, however, the expectations are higher: we want a home playoff game. Technically, with the new format, that's a home playoff best-of-three series, so at least two playoff games. The top four teams in the Western Conference will get that chance. Fivethirtyeight, which uses last season's results in their predictions (barf), gives the Quakes a 73% chance to get at least the ninth spot, and 64% to get any spot one through eight. I'm not feeling their prediction, largely because of too much dependency on last year's results.

Setting up an in-season simulation

One of the inherent issues with an in-season simulation is the amount of data for each team varies due to the number of games they've played, but also how many games they've played at home and away. As we've discussed ad nauseam before, the biggest factor for success in Major League Soccer is not money, like in the English Premier League, but the built-in home-field advantage and how teams do within that, by either building a strong home fortress or finding ways to win on the road. Points get you in the playoffs, and more points give you home-field advantage for longer in the playoffs so you have a better chance to hoist the MLS Cup. Enough points will even get you a trophy. Not enough points might get you a wooden spoon.

The Philadelphia Union were 2022's only team without a home loss, getting 12 wins and five draws. Only six teams managed a positive goal differential on the road. The Earthquakes and Rapids only managed one road win. Austin, on the other hand, won as many road games as home games. Nashville, who finished fifth due to goal differential with 50 points, only won six home games and seven away games. Do you know how many home wins the Quakes have just over one-third into the season? Five, tied with LAFC.

So there are many paths to a top-four spot; many paths to the likely 50+ points it will require to get it. But how many away wins does the path require? We brought back our season simulator to find out.

In the preseason, we simulated full seasons. However, with 13 games already having been played by the Earthquakes, we needed to start the simulator from this point. San Jose has banked 18 points. They get to keep those in the simulation. All the MLS teams will keep their existing results and points since we are simulating game-by-game.

The other good news is that every team in MLS has played at least 5 home games and 5 away games, giving us a decent idea of what their performance range for both is.

We've established before how our simulator works. It looks at four things:

1. Home/Away Goals For and Goals Against
2. Home/Away xG For and xG Against
3. Home/Away Standard Deviation on Goals For/Against
4. Home/Away Standard Deviation on xG For/Against

Now at this point in the season, xG is a better predictor than goals of future performance. As referenced here, MLS is a dumpster fire of unpredictability, but xG is a better predictor of future performance anyway.

Here is a chart from that ASA article showing that at 13 games into the season, the Expected Goals Ratio is the best predictor of future points per game, regardless of the league.

With that in mind, we are going to tell the simulator to use two-thirds xG results and one-third actual results. In case you are wondering, here's how that breaks down for the Quakes:

In effect the Quakes have overperformed defensively at home, a sign things could change in the future. Holding opponents to a 3.5% conversion rate at home isn't likely to hold up for a full season. At the same time, they are slightly underperforming on the offensive side, so that could get better.

That almost-two-goal overperformance defensively at home is balanced by the fact they have underperformed defensively by two goals on the road, also a sign they could improve there as much as they could regress to the mean (or worse) at home.

At either rate, there is very little in these numbers to suggest things will drastically change for San Jose, and I feel good about using a two-thirds xG/one-third goals mix. The Earthquakes have been decently consistent in their performances, both home and away.

The standard deviation of their past performances and the quality of their opponents at home and on the road will help us determine the most likely outcomes, but I had to double the standard deviation value of the current season across all teams to create the typical level of chaos of a full season in MLS. The initial simulation showed the home teams were unrealistically strong, and by doubling the standard deviation we allow ourselves room for some of the unpredictability that happens in an MLS full season that we'll discuss later in the article.

As usual, we are about to simulate the remainder of the season 10,000 times.

So how many away wins does it take for the Quakes to get to the center of a home playoff spot?

(Cue the owl licking the tootsie pop to find out)

The answer most often across 10,000 simulations was three away wins to get any first-round home playoff spot, one through four. However, to get just fourth place, it was a toss-up between two and three away wins (581 to 600). The incredible thing is it's possible to get no wins on the road and even win the Western Conference. Wild.

The simple reason for that is total wins. To get first in the West, San Jose probably need to win eight or nine of their remaining 11 home games. To get fourth, it's six to eight wins. They don't have to continue their current home-winning pace, but it's likely easier than winning more than two games away from home.

In our simulation, the most likely number of points for the Quakes, given home-and-away 2023 performances so far, was 49, but the most probable range is between 45 and 54.

It would take something akin to the disastrous finish of 2019 to miss the playoffs altogether. To get a home playoff spot, 51 points is a good bet, but it may take more. Last year, it took 50 plus a goal differential tie-breaker, and in 2021 it took at least 50.

It all starts with FC Dallas at home

Looking ahead to Saturday's game against FC Dallas adds much more pressure. The simulator gives the Quakes an almost 55% chance to win based on San Jose's home performances and FC Dallas' 2 wins/2 draws/2 losses performances on the road:

• Win = 54.6%
• Draw = 28%
• Loss = 17.4%

To get a home playoff game, the Quakes almost have to win, for multiple reasons: 1) FC Dallas is in possession of the most realistic home playoff spot, 2) winning over FC Dallas not only gives the Quakes three points, but also takes three from FC Dallas' possible points, and 3) not getting a win at home coupled with losing on the road to FCD in the season's penultimate game could be the biggest factor in preventing them from getting the final home playoff spot if the Quakes maintain their otherwise current pace. Getting a draw at home, so long they win the away game against FC Dallas wouldn't be a disaster but would create more stress on the remaining schedule.

A league unlike any other

As a reminder, MLS is the most unpredictable soccer league in the world's top 20 or so leagues. As I mentioned, I had to double each team's standard deviation to get to an acceptable level of MLS chaos. Where a simulation like this can completely go awry is in not understanding how things can drastically change in an MLS season. Here are a few places where wrenches typically get thrown into the gears:

The summer months

Ah, those midwest and Texas summer months of heat and humidity that cause even the most aggressively-pressing teams to falter and back off. Are you listening, St. Louis City? Or maybe it could be wildfires up and down the West Coast that impact player health and fitness.

Regardless, the summer months often completely change the way teams play games, and, thereby, change the trajectory of some teams every season.

Cards and injuries pile up (also Leagues Cup)

There's been a lot of talk about the depth of the Earthquakes' roster on The Aftershock. We are starting to see that this team does not have reliable backups at a few positions. Yellow card suspensions happen more frequently. A red card could be a two-game double whammy and affect a key defensive player.

At some point, a lengthy injury could hit an attacker that is holding the attack together. The lukewarm news is that it hits almost every team, however, the Quakes may be less able to deal with it than some others at the top of the table.

The Leagues Cup is likely to bring additional pressures, and maybe injuries, but also the likelihood of a couple of weeks off for teams that don't advance. How do teams stay sharp that are eliminated in the early rounds?

The secondary transfer window

Teams like Sporting KC turned their season around with an impactful secondary transfer window this season. LAFC, on the other hand, ruined their chemistry for a bit with questionable moves in the 2022 window.

Coaching changes from poorly performing teams could unlock the potential of underperforming players and change their team's fortunes or have no tangible impact on the table whatsoever.

The Quakes have an opportunity to bolster their position for a playoff run. They could also ruin it with the wrong moves. Is moving Cade (and/or Benji) to Europe in the cards? If so, who can they add to make an impact? Can they get more attacking production in the central midfield with a move without impacting the defensive solidity that is getting them consistent wins at home? These are some big questions.

Conclusion

It's impossible for our simulation to factor in these unknowns. All we can do is say what it is likely to take in wins and points to get that home playoff spot.

From this, I would say nine wins is a good target for the rest of the season. The more the merrier, of course, but nine will put the Quakes right there. Maybe we should track a Magic Number for this on The Aftershock.

After all, the simulation says the Quakes are most likely to finish between third place and sixth place, even based on the current road performances. The most probable finishing table position so far? Fourth, and a first-round set of home playoff games.

I'm switching from away game angst to home game angst.

It starts by beating Luchi Gonzalez's former team, FC Dallas, this Saturday at PayPal Park.

Enjoy the home game.