Who’s the Most Dangerous Passer in the NWSL?

No, it’s not who you think it is.

Last year, I took David Sumpter’s Pro Soccermatics course and learnt about expected threat (xT) models. He wrote a great article on Medium explaining xT and action-based xT models, which I recommend reading.

I wanted to know which players have the highest xT in the NWSL so I built an NWSL xT model. This model estimates the value of a pass based on how likely it is to lead to a shot, and how dangerous that shot is. It identifies the most dangerous passers - those creating the highest xG opportunities through their passing. I was somewhat surprised by the results - although once I looked at other metrics, my findings made a lot of sense.

Here are the top xT players for each NWSL club, based on the raw sum of xT created. This is intentional, as it highlights players who not only know how to create danger but do so consistently and at a high volume.

North Carolina Courage

North Carolina Courage currently have the league’s top xT producer: Ryan Williams. I know what you’re thinking: what is a defender doing top of a threat model? I wondered the same, but a closer look at the data won me over quickly. Let’s start with the graphic below, its difficult to argue that those ten passes (Williams’ top xT-creating passes) aren’t dangerous, even with a basic understanding of football.

We also know that Courage manager Sean Nahas is all about possession. So perhaps it’s not surprising to see a Courage player topping an action-based Expected Threat metric.

But why a defender? Well, Williams has the third highest number of completed passes this season which provides more opportunities to rack up xT. Defenders tend to pass the ball the most, and if goals are produced through build-up play that starts with defenders passing among themselves, those defenders still get credit in a model like this. So, get used to seeing defenders perform well in this xT model, because if they’re part of a passing team then the passes have to start somewhere.

Williams ranks third in progressive passes this season, and fifth for crosses into penalty area. But the other reason I think Williams ranks so highly in this xT model is Ashley Sanchez. Williams and Sanchez combine exceptionally well down the right side of the Courage’s attack. They have combined for three goals and one assist in NC’s last two games. Sanchez is the top receiver for progressive passes for the Courage.

San Diego Wave

The Wave’s two central midfielders, Savannah McCaskill and Kenza Dali, rank first and second for progressive passing in the league. Add in a top-three ranking for passes into the penalty area and crosses into the box for Dali, and it’s not surprising she has the second-highest xT in 2025 so far.

Looking at the Kenza Dali xT heatmap above you can tell she mostly operates down the left side of the pitch. However, her highest xT is accumulated from the top-right area, due to her in-swinging corners coming from that side and any second balls that follow.

Dali’s highest xT in a single game (1.66) came against Bay FC, where she also assisted the winning goal from a corner. Four of her top ten xT passes in that match came from corners.

Bay FC

Taylor Huff is Bay FC’s top xT performer so far in 2025. One thing that stands out from her xT heatmap is that her threat isn’t limited to any particular area of the pitch.

Huff is the third-highest cross contributor in the league and has delivered the most corners. The heatmap above reflects this well, with two of her highest-xT passes being corners, while many others are longer passes from the flanks into the box. She’s also in the top ten league-wide for passes into the box. Huff is Bay’s joint-top progressive passer (alongside Malonson).

Washington Spirit

Considering the attacking talent Spirit boast, you might struggle to believe that a centre-back comfortably tops their xT leaderboard. Not only that, but the top three are all defenders:

1. Rebeca Bernal

2. Tara McKeown

3. Esme Morgan

The reason is these three most often play the ball into the attacking third.

Bernal’s heatmap looks distinctly different from the more attacking players we’ve seen so far, but aligns with expectations for a centre-back. She appears to accumulate xT by playing numerous passes from just outside her own penalty area, and whilst none are individually high in xT they add up over time. Her most dangerous passes occur when Spirit push up and she steps just inside the opposition half.

Interestingly, one of her highest-xT passes was a backpass to the keeper inside her own six-yard box which is as a result of the model crediting recycled possession within build-up play.

Gotham FC

Gotham saw Spirit with three centre-backs as their top xT merchants and decided to go one further: they’ve got a goalkeeper in their top three. Absurd, I know.

Ann-Katrin Berger is not just a shot stopper though, she was brought into the club for precisely this reason - to act as an additional outfield player, helping Gotham create overloads. Berger’s xT for 2025 is 5.36. That’s higher than any player from Chicago Stars, Angel City, Bay FC, Kansas City Current, Portland Thorns, Orlando Pride, Racing Louisville, Seattle Reign and Washington Spirit. Yes, I just listed 9 out of 14 of the NWSL teams to make my point. Sorry.

Enough about AKB, because Emily Sonnet is Gotham’s xT leader. Sonnet has the highest number of completed passes in the league, which significantly benefits her in this pass-based xT model. Gotham have had 50% or more possession in 7 of 8 matches this season, and Sonnett is often the one recycling possession.

Sonnet’s highest-xT match this season was a 4–1 loss to Portland Thorns, where Gotham had 61% possession, their highest this season.

Sarah Schupansky created the most xT in a single game for Gotham this season, against Seattle Reign. She leads the league in number of crosses and is second for number of corners taken.

Houston Dash

Paige Neilsen leads the Dash in xT this season, dominating the right-hand channel of their back three. Her best xT performance (1.51) came against Angel City, where Houston enjoyed 68% possession.

Neilsen leads the Dash in number of progressive passes, and is second in number of completed passes.

Kansas City Current

Finally, a name you were probably expecting: Debinha. Her xT heatmap shows that it doesn’t really matter where she picks up the ball, she can do some damage.

Her top ten dangerous passes all land in the same square on the pitch (bar one), showing her consistency. Debinha leads the league in completed live-ball passes that led to a shot attempt and ranks fourth for passes into the penalty area.

That’s seven of the fourteen NWSL teams covered. I’ll follow up with the remaining seven in my next article.

Behind the xT model

Position-based xT assigns a value to every point on the pitch based on the probability that having the ball there will lead to a goal. If a player moves the ball from a less dangerous to a more dangerous area, they increase their team’s xT. This change in scoring probability is the expected threat (xT).

Action-based xT works slightly differently. Rather than focusing purely on location, it takes the nature of the action into account. The threat lies in how the ball is moved, not just where. Cross-field balls receive higher value, while backpasses aren’t penalised, since they can help retain or advance possession later on, unlike in a purely location-based model.

I used the latter in this analysis.

I first created possession chains within matches. Then, I trained an XGBoost classifier to predict the probability of a chain ending in a shot. For chains that did end in a shot, I used a linear regression model to calculate the probability that the shot would become a goal. The product of these two values gives the action-based Expected Threat.

This model estimates the value of a pass based on how likely it is to lead to a shot, and how dangerous that shot is. It allowed me to identify the most threatening passers - those consistently creating the highest xG through their actions.

Data source: xT model: Opta. All other data: FBREF.