Looking For Progressive Passers In The FAWSL Using StatsBomb Data

Two things I’ve been putting off for a while are taking a look at StatsBomb data and getting into women’s football. This is nothing new for me, I’m often hesitant to take up new things, but I’ve decided that this summer I’m going to use the World Cup as a gateway into women’s football, which also offers a perfect chance to get two things done at once and take a look into StatsBomb’s free data – especially for the women’s game.

For my first look into StatsBomb data I’ve decided to dive into the passing data and try to create a progressive passing metric and look for progressive passers in the FA Women’s Super League.

The caveat is that I’m not familar with the women’s teams and players, so I’ll also be using the 2018 men’s World Cup to help guide the method, before applying it to the FAWSL.

Needless to say, the data is all from StatsBomb and I followed this piece by @FutballAnalysR to get the pass and shot data from R into a csv. I also did most the heavy lifting in Excel, as I’m not great with R and my Python is rusty after not programming anything for about a year. It’s also worth mentioning I excluded goalkeepers, throw-ins and free kicks too and only used completed passes for player totals.

Method

Method 1

What I’m trying to do is inspired by (read: shamelessly stolen from) @MC_of_A, who often posts numbers for progressive passes and runs. The method I’ve seen for this was in an SBNation piece about RB Leipzig. To quote the article, where they were talking about Naby Keita, it says this:

To quantify Keita’s production in midfield, I developed a statistic for ball progression in midfield. This counts passes that take the attacking move at least 10-15 yards toward goal compared to where the ball had been over the last three actions (less progression required nearer to goal), and it counts runs on the ball that beat an opponent and progress the ball a similar distance.

@MC_of_A

To try and replicate something similar my idea was to take the maximum point of a possession (as it goes) and then look how each pass progresses the ball in comparison to that. I’m not very good at explaining it, but each pass has an attribute for the furthest forward the possession has reached up until that point, then if a pass moves the ball forward beyond that point, by a certain amount, it’s deemed as progressive.

To start I only looked at the x-coordinate, with the objective just to move the ball up the pitch, and decided a pass would be deemed progressive if it goes 20 units (pitch length is 120 units) further than the furthest point in the possession.

To test it I jumped into the 2018 men’s World Cup and took to the Germany vs Mexico game to check how these progressive passes looked (Germany players put up good numbers for progressive passes).

I felt a bit disappointed by the results. While the names seemed okay with Luka Modric, Toni Kroos, Toby Alderweireld, Jerome Boateng and Kevin de Bruyne having the top five total progressive passes (which is pretty crazy considering Kroos only played three games and Boateng only played two), watching videos of progressive passes felt disappointing.

For an example, this pass by Mesut Ozil is deemed progressive when it only looks as though he pushes the ball to the side.

The reason it was flagged up as progressive was because the furthest point the possession reached before the pass was where Ozil picked the ball up, the end location of Boateng’s pass, so the end location of Ozil’s pass is deemed as progressive compared to where he picked up the ball – although doesn’t look progressive from where he passed the ball from.

A more extreme version of this can be seen below from Arsenal’s Danielle van de Donk.

I was in two minds about this situation. On the one hand, they do progress the ball, but it’s not really a progressive pass and if we want to find progressive passers, rather than just ball progressors in general, they should probably be cut out.

Method 1.1

To try and remedy the above situation I set a new rule that the player had to progress the ball 20 units further than the possession has reached, but also the pass itself has to progress the ball at least 20 units.

The amount of passes being flagged as progressive dropped, but the top five progressors for the 2018 men’s World Cup was similar. Toni Kroos was now first then Alderweireld, Modric, Boateng and Paul Pogba.

Watching a few clips of passes that made the most progression from the FAWSL showed that this was better, but another issue was raised: The goal had no effect on progression. The two most progressive passes in the FAWSL were players hitting long balls towards the corner flag and a team-mate getting it and being penned in the corner.

With these being the most extreme cases of progression, it isn’t a huge issue, it’s not as though every pass being deemed progressive is a long ball into the corner, but it raised a question about how the distance from goal should feature. If a pass is progressive you’d assume it not only moves the ball up the pitch but also closer to goal.

Method 2

To look at players who both move the ball forward and closer to goal I altered the method by looking at the distance from goal of the ball during the possession, rather than just the x-coordinate.

I took the distance from goal at each pass and should a pass move the ball 20 squared units closer to goal than it had previously reached in the possession, it was deemed progressive.

This sees a slightly bigger change to the top five at the men’s World Cup from last year. Modric is first again, then Kroos, Neymar, De Bruyne and Raphael Varane.

Method 2.1

Making the adjustment to rule that passes must move the ball 20 square units closer to goal from the furthest point in the possession and the pass itself has to move the ball 20 square units closer to goal shakes up the top five slightly, with three new names being in there this time.

Modric is still first, Kieran Tripper is second, Simon Kjaer is third, Alderweireld fourth and Ivan Rakitic fifth.

Trippier and Kjaer being put in is interesting, @SaturdayOnCouch often mentions Trippier as having good progression numbers so that doesn’t seem too alarming, but Kjaer is definitely a name I didn’t expect to see in the top five, especially with just four games played.

Using this method, the most progressive pass of the men’s 2018 World Cup goes to Se-Jong Ju of South Korea against Germany, which should come as no surprise given the position of Manuel Neuer at the time.

Summary

Overall, each method has its different pros and cons and I’ll use them all to try and find some progressive passers in the FAWSL.

There’ll always be things to nitpick, with passes being treated in isolation rather than in the context of the style of the team, the move the pass is in, the options available for the passer and the intention of the passer, but I feel okay about the names that came up when testing with the men’s World Cup data.

Progressive Passers In The FAWSL

After boring you with the method, it’s finally time to try and find some progressive players in the FAWSL. I’ve decided to highlight two young players who have strong numbers for progressive passing, before posting the rankings for the league.

It’s also worth noting I took the position of the players by taking the most popular value they were labelled with in the data.

Leah Williamson – 22- Arsenal

Looking at the four methods of progressive passing from above, the battle for the most progressive passes is between Arsenal’s Leah Williamson and Manchester City’s Steph Houghton. Williamson is 1st with Houghton 2nd for the first and third method, while Houghton’s 1st and Williamson 2nd for the second and fourth method.

With Houghton being England captain and Williamson only being 22-years-old, it’s Williamson who I thought I would mention here.

Playing for a dominant Arsenal side, both on xG and in the league table, Williamson put forward some strong passing numbers. She drops slightly in the method that rules the pass must progress the ball 20 units, which hardly seems surprising when looking at some clips of her. She seems more than comfortable carrying the ball out of defence and moving it on, as can be seen below.

With these clips showing her operating as the right sided centre-back and making a pass for a wide player, it seems as though she should make a good fit for the national side, bringing the ball out of defence and finding an overlapping Lucy Bronze out on the right.

A lot of the time when finding players with good numbers for either progressive passes or passes into the final third, I’m left disappointed as it’s usually because they punt low probability long-balls forward, which makes it nice to see someone like Williamson who looks comfortable on the ball and looks to pick a pass to progress the play.

I’m not good at judging defenders even if I watch lots of them, so I can’t comment much on her defending ability, but Arsenal did have the best xG against last season. With how good they were last season it shouldn’t come as a surprise, but if at 22-years-old Williamson started 18 out of their 20 games, I get the impression she must also be very good.

My reaction to seeing Williamson is similar to when I saw Alexis Saelemaekers in my Wolves piece – I very quickly went from knowing nothing about them to being all in on them within a few clips. Her numbers are strong and the clips more than back them up.

With Steph Houghton being 31-years-old, Williamson could be the long-term replacement for her in the national team and it’d be nice to see Williamson get some minutes during this summer’s World Cup.

Keira Walsh – 22 – Manchester City

Moving into midfield another player that shines – and another member of the Lionesses World Cup squad – is 22-year-old Keira Walsh for Manchester City.

With it being Arsenal vs Man City in defence with Williamson and Houghton, it’s the same in midfield with Walsh and Dominique Bloodworth of Arsenal.

Looking only at midfielders, Bloodworth is 1st and Walsh 2nd for all four methods I did above, but having already mentioned an Arsenal player, it’s Walsh who I thought I’d mention – plus Bloodworth spent time in both defence and midfield last season.

Watching clips of Walsh it doesn’t take long to realise the impressive passing range she has, she seems to pull off just about every type of pass. The below video shows some clips of her hitting long diagonals to switch the play.

The video below shows some other long passes (some could be classed as switches too, it’s just how I labelled them when I saved them).

Finally, the below video shows some other passes of hers, with three nice vertical looking passes coming in the same game against Brighton.

A couple of clips also make it seem as though Walsh is good when it comes to resisting pressure, which is supported by her dribbling numbers on WyScout, attempting ~2 dribbles p90 with a ~70% success rate in the FAWSL last season. Once I get over my programming rustiness it’ll be interesting to see how she fares when using StatsBomb’s data for dribbling and pressure.

I only looked at her passing clips, so there’s likely more/better examples of similar things out there, but the below turn against Birmingham was my favourite clip showing her resisting pressure, although the pass at the end of it was disappointing.

Walsh didn’t put forward strong numbers for progressive passes in England’s opening World Cup game against Scotland, with much of the progression being done by the defenders – and full-backs especially when looking at the third and fourth methods – but there’s still time for her to utilize her passing range in the rest of the tournament.

Rankings

To put some of the names above into context a bit more I thought I’d post the top names for each method, both passes and passes received. I’ve been lazy and used total numbers, rather than getting the minutes played, but I’ll get around to sorting the data out soon.

Method 1

Progressive Passes

General

NameTeamProgressive Passes
Leah WilliamsonArsenal156
Steph HoughtonMan City153
Magdalena EricssonChelsea142
Josanne PotterReading141
Kirstyn PearceReading108

Midfielders

NameTeamProgressive Passes
Dominique BloodworthArsenal96
Keira WalshMan City90
Danielle van de DonkArsenal70
Sophie IngleChelsea65
So-yun JiChelsea55

Passes Received

NameTeamProgressive Passes Received
Beth MeadArsenal154
Nikita ParrisMan City143
Vivianne MiedemaArsenal137
Charlie WellingsBirmingham City119
Brooke ChaplenReading109

Method 1.1

Progressive Passes

General

NameTeamProgressive Passes
Steph HoughtonMan City134
Leah WilliamsonArsenal111
Magdalena EricssonChelsea98
Josanne PotterReading96
Dominique BloodworthArsenal75

Midfielders

NameTeamProgressive Passes
Dominique BloodworthArsenal75
Keira WalshMan City67
Sophie Ingle Chelsea47
Laura CoombsLiverpool43
Danielle van de Donk Arsenal41

Passes Received

NameTeamProgressive Passes Received
Nikita ParrisMan City119
Beth MeadArsenal114
Vivianne MiedemaArsenal92
Brooke ChaplenReading78
Clarke WellingsBirmingham City75

Method 2

Progressive Passes

General

NameTeamProgressive Passes
Leah WilliamsonArsenal180
Steph HoughtonMan City157
Josanne PotterReading121
Millie BrightChelsea119
Kirstyn PearceReading116

Midfielders

NameTeamProgressive Passes
Dominique BloodworthArsenal103
Keira WalshMan City78
Jill ScottArsenal63
Danielle van de DonkArsenal56
Sophie IngleChelsea55

Passes Received

NameTeamProgressive Passes Received
Lisa EvansArsenal151
Nikita ParrisMan City143
Beth MeadArsenal135
Hannah BlundellChelsea99
Alisha LehmannWest Ham96

Method 2.1

Progressive Passes

General

NameTeamProgressive Passes
Steph HoughtonMan City116
Leah WilliamsonArsenal110
Millie BrightChelsea70
Kirstyn PearceReading63
Sophie Bradley-AucklandLiverpool63

Midfielders

NameTeamProgressive Passes
Dominique BloodworthArsenal49
Keira WalshMan City34
Laura CoombsLiverpool34
Sopie IngleChelsea31
Kim LittleArsenal31

Passes Received

NameTeamProgressive Passes Received
Nikita ParrisMan City94
Beth MeadArsenal84
Lisa EvansArsenal80
Alisha LehmannWest Ham62
Kayleigh GreenBrighton61

Conclusion

There’s a lot of fine tuning needed, mostly with the distances of what’s deemed as progressive and making an adjustment depending on how close to goal the pass is made, but I’m quite happy with this method at the minute.

From here I have a few things on my to-do list to make it better. At the minute it’s the fine tuning mentioned above, getting the minutes played of players rather than just looking at totals, looking at attempted progressive passes rather than just completed to see how efficient players are at progressing it, before diving into other aspects of StatsBomb’s data like pressure and how that effects these numbers and looking at other events other than just passing.

I should probably whittle the methods down rather than having all four, but I like all four of them and each of them feels as though they do a slightly different job which can be useful in finding different types of ball progressors.

It’d also be worthwhile to test whether any of this actually matters, or whether looking at things like long balls and passes into the final third tells you the same as this progressive passing method.

Also, given this piece and the World Cup is my first real venture into women’s football, please feel free to tweet/DM me (@_POTP) with recommendations of people to follow or pieces to read regarding women’s football and the World Cup in particular.

Finally, thanks again to StatsBomb for giving out this data for free, hopefully I can stop being so lazy and get some more content out using it.

Leave a Reply

Your email address will not be published. Required fields are marked *