Thanks to Arielle Dror, John Muller, and Tiotal Football for their advice, Jamon Moore for cooking up any data I requested, and Eliot McKinley for the g+ wheel chart and enduring hundreds of DM’s.

_____________________________________________________________________________________

Ask any data analyst about the utility of statistics and you’ll get any number of answers, from the ability to measure shot quality to the potential to build all-in-one impact measurements, such as American Soccer Analysis’ goals added (g+) metric.

However, the most fundamental utility of data is not found in the answers it provides, but in how it empowers us to ask questions.

For example, while pouring through ASA’s numbers on Sophia Smith in anticipation of a piece on her and the Portland Thorns, I noticed something curious. Despite dominating the NWSL in xG, she came off looking far more mediocre in Receiving g+, ranking only 26th among her peers (xG and g+ data accurate as of October 30th).

Going back to ASA’s explainer on goals added, we can see why this is weird on face value:

“Forwards are getting a lot of credit for being available and receiving [emphasis added] the ball (or winning headers) in dangerous places in front of goal. To some extent, we are imposing our belief on the methodology that the hard work is done in getting the ball in dangerous locations, and that the shot is a formality.”

In other words, Receiving g+ (not Shooting g+) tends to capture a lot of the value associated with xG, as it apportions credit to the attacker for taking their first touch in high-value territory.

Why, then, is Sophia Smith not getting any love as a receiver, despite obviously getting into great goal-scoring locations, as evidenced by her 0.59 xG p90 figure?

Remember, Receiving g+ assigns a number to the reception - not the shot. Thus, while it can be considered a good proxy for xG and vice-versa, they are not, in fact, the same thing, which can produce discrepancies like the one we see with Smith.

It’s for this reason ASA analyst Jamon Moore has been developing something called “pre-carry xG,” which reflects the scoring probability of a shot if it was taken immediately upon the reception. By comparing this xG value to the one that materializes when the shot actually occurs in real life, we can start to see what types of actions produce the biggest differences and how this might explain Smith’s situation.

Headed passes, long balls, normal passes, progressive passes, and through balls have all engendered the biggest “xG-pre” differences over the last few seasons in the NWSL. This makes sense, as the end locations of these sorts of actions tend to occur farther from the penalty area, requiring attackers to do a lot of their own work to produce high-quality shots after getting the ball.

So, although progressive passes and through balls have been proven to be very important to goal scoring, they primarily benefit players by putting them in a strong position to create a good look, rather than directly fashioning the attempt for them.

This stands in stark contrast to crosses*, cutbacks, and penalties, which all occur closer to goal and primarily set up first-time finishes; hence, the “xG-pre” difference being around zero.

When applying these same filters to Sophia Smith’s shot generation, everything becomes a lot clearer.

She’s rarely been the beneficiary of assists that put her immediately in front of goal, getting only a handful of shots off of crosses and none off of cutbacks. Instead, she’s produced most of her opportunities from “individual” actions (shots that aren’t attached to a key pass) and progressive passes.

In total, she has added 2.48 xG through her own initiation.

A prototypical striker like Rachel Daly is the exact opposite. She sits in 2nd place below Smith for total xG but blows the latter and the league away in Receiving g+, almost doubling the numbers of next best Jessica McDonald.

Daly barely adds any extra xG with her touches (0.28 in the aggregate, to be exact) because so many of them are first-time. She’s the ultimate example of a penalty-box predator that excels at getting on the end of passes that require little more than a finishing touch. Even the progressive passes that Daly receives place her closer to goal compared to the historical norm; her pre-carry xG per (theoretical) shot is 0.16 off of this kind of service - that’s 2x Smith’s 0.08.

Clearly, Smith thrives differently.

The Thorns attacker is an aggressive driver that explodes off the mark like few in the world, flying into touches that either push the ball into space or cheekily nutmeg a defender.

The freeze frames tracing her path from touch one to the shot serve to illustrate just how proficient she is at bringing possession into more dangerous zones from her start location.

Her added value from these solo efforts are probably captured well by Dribbling g+, which ranks her 8th among the league and 3rd among those who are classified as strikers.

We can also see this statistically in another way if we break her shot generation down further:

As is evident above, Smith adds xG in most of the action-type categories via dribbling and carrying.

Nevertheless, there are still some big “xG-pre difference” figures on actions that are classified as successful passes (notably through balls and headed passes), indicating a sort of grey area where value added after the first touch gets lost. As Jamon told me: “the dirty secret is that there is no carry event for most shots received from a key pass, even if there probably should be.”

That applies to the moment below, where Smith takes a touch that undeniably adds value by bringing the ball into the box, but the distance is too small for it to be recorded as a carry.

Indeed, while the numbers might be missing a trick, it feels weird to say that Smith’s carrying or dribbling created that chance. Instead, what this really speaks to is the advantage provided by her off-ball movement.

I’m not sure whether carries are attached to these sequences (the distances between the start and end locations are certainly greater than in the prior example), but the larger point is that her runs are another factor driving some of this xG-Receiving g+ discrepancy.

What’s even more interesting is the somewhat unorthodox nature of the 21-year-old’s movement. While Smith is fully capable of making traditional runs between the center-backs, she loves to hover out on the flanks, attacking the blindside of the fullback from the outside, as is visible in the third clip.

*Sophia Smith’s heatmap, illustrating her volume of touches in wide areas (source: Wyscout).*

This positional tendency dovetails well with her carrying and dribbling, where she can find her space to isolate and cut inside, and is something coach Mark Parsons looks to exploit by deploying her in a flexible 4-4-2 diamond.

Smith’s desire to push wide quite often - combined with her lack of first-time penalty box touches - makes her a rather atypical striker. She feels like some kind of mix between an off-the-shoulder #9 and a winger, which is probably why USWNT manager Vlatko Andonovski has played Smith on the right in a 4-3-3 next to false nine Carli Lloyd (and hopefully false nine Catarina Macario in the future), where the former can more accurately mimic her Thorns’ role as opposed to if she operated as a lone forward.

Smith is also rather unusual in her defensive work-rate. Whereas others stationed higher up prefer to skimp on such unglamorous duties, Smith takes pride in them, displaying an intensity and mastery of angles that make her genuinely elite as a presser.

Interrupting g+ adores her, ranking her 1st among strikers and 2nd among all classified in attacking positions at time of writing. Even when opening the sample to everyone, Smith sits inside the top 20.

Based on film analysis, we can infer that this is because her activity against the ball frequently leads to shots, similar to how her Dribbling g+ is likely so high because of how much Smith adds to pre-carry xG via solo brilliance. Taken together (along with stuff she gains from fouls, shooting, and receiving), Smith sits fifth in total Goals Added in the NWSL.

The best part is that we have to consider that her final figure might be depreciated due to the “dirty secret” discussed before, not to mention that event data-based models still cannot properly assess the value of off-ball movement.

However, regardless of what her final g+ stat should look like, what is apparent is that Smith is unlike many strikers you might come across.

While most of them, like Rachel Daly, focus their game around finishing off deliveries in the most valuable of locations, Sophia Smith brings the ball into those zones herself, leveraging her spectacular dribbling, unorthodox movement, and industrious defense to turn good (or non-existent) passes into great ones.

______________________________________________________________________________________

*Why corners produce a negative xG-pre difference in the aggregate is interesting. Maybe players are controlling the delivery and moving backwards or into a worse angle on certain occasions? But that’s for another article.