Why didn't xG reflect Austin FC's dominance against LAFC?

AUSTIN, Texas — All year long, Austin FC has been saying thank you to its haters. The team that many pundits (including this one) picked to miss the playoffs in year two has taken supreme joy in proving its doubters wrong.

At this point, after its 4-1 victory over Los Angeles FC in front of a national TV audience, the hard part is going to be finding anyone left who's doubting.

That is unless you know where to look. American Soccer Analysis has the dopamine hit that Felipe Martins is looking for. Because even after its best performance of the season, you can twist the computers' analysis to say that Austin FC should have lost. Considering expected goals (xG), ASA's model had LAFC with a 1.90 to 1.03 advantage. MLSSoccer.com, which uses a slightly different algorithm, had it at 1.5-0.6 in favor of LAFC.

How is that possible? It has to be a mistake, right?

Felipe should stop reading now. For the rest of you, I'm going to explain why you shouldn't (necessarily) be sent into a fit of rage by this obvious example of disrespect.

Game state matters

If you've been listening to our friends over at Moontower Soccer, you'll know that Austin FC has been involved in a lot of wonky xG matches this season. One thing that almost all of them have in common is an early goal.

There have only been a handful of Verde matches this season that reached halftime scoreless. The truth of the old adage, "goals change games," has been abundantly clear to see.

I asked ASA's lead club consultant, Sean Steffen, to take a look at the xG from Friday, and here's the first thing that stood out to him.

"Austin once again jumped to an early lead on relatively low xG," he said. "This time, to the extreme."

At the same time, Verde did something it has struggled to do throughout the season. For more than an hour, it held LAFC without a shot on goal. José Cifuentes's shot in the 61st minute, that led to the rebound for Cristian Arango's goal, was the first that was on frame.

At that point, Austin FC led 4-0 and had accumulated all of its shots except for one low-probability header by Sebastián Driussi in the 78th minute. LAFC was in attack mode the rest of the way.

"There is a huge urgency deficit," Steffen said. "LAFC is clawing for one, and Austin is, frankly, happy to let them build up as much xG as they want until it actually converts to goals and the scoreline is threatened, which didn't happen."

xG doesn't factor in finishing

One of the deceiving things about the term expected goals is that the actual algorithm used to calculate xG is fairly one-dimensional. According to the ASA website, "expected goals (xG) are the number of goals that can be expected to be scored based on where and how a shot was taken."

That's basically it. Location and the type of shot (foot vs. header) are the main variables. The skill of the individual who took the shot is not part of the equation.

And so, while the majority of 20,738 fans inside Q2 Stadium had the knowledge that Diego Fagúndez was lining up a free kick similar to the one he had just scored on July 12 against Houston, the computers didn't have that knowledge. They saw a free kick 25 yards from goal, and calculated the xG as 0.08. 

But there's another number that was much kinder to Austin, and a better reflection of where it excelled Friday night. That number is xPlace, which calculates the difference between post- and pre-shot xG models. So for that Fagúndez striker, the xPlace figure was 0.52.

Maxi Urruti, who finished the match with two goals and 0.62 xG, had the highest xPlace of any player on either team at 1.24. That's more like it.

Which brings us to....

Driussi is the G-xG King

The larger xG conversation we should be having about Austin FC involves Sebastián Driussi, who continues to defy the models in his pursuit of the MLS Golden Boot. Driussi's goal in the 59th minute against LAFC, his league-leading 19th of the season, brought his G-xG total to 8.20.

That means, based on the factors we discussed above, the xG algorithm at ASA would expect Driussi's goal total to be closer to 11. In fact, Driussi is on pace to break the single-season record for G-xG, shattering the record of 8.56 set by Lee Nguyen in 2014 with the New England Revolution.

Driussi's goal on Friday was a perfect example of what's made his season so special, as he turned a 0.13 xG opportunity into a dagger that finished off the best team in the league.

And yes, it was a tight-angled shot. But if there was anybody inside Q2 who didn't expect him to bury it then they simply haven't been paying attention. Driussi ended the match with an xPlace total of 0.65, a testament to his finishing ability.

You might expect his season-long xPlace total to be higher than 3.06, which ranks third in MLS behind Gabriel Pereira of NYCFC and Taxi Fountas of D.C. United. And that begs the question analysts are asking about Driussi's MVP-level season: is it sustainable?

If you're an Austin FC fan, you desperately want to say yes. Driussi has remained consistent all season, producing moments of magic that are thrilling but no longer surprising. To borrow a phrase from Austin Chronicle reporter Eric Goodman, Driussi leads the league in un-expected goals.

It doesn't really matter where he's shooting from. Driussi's finishing has just been that good.

While it's fun to express faux outrage about the xG numbers constantly failing to give Austin FC its due, it doesn't really mean much.

Unless you're Felipe Martins. In that case, you've found a new doubter to prove wrong.