I have an idea for a new feature of my column. It’s a running comparison of shot quality (what people call “expected goals”) over the course of a match. I divide the game into five minute increments and for each shot taken I weight the shots by location. For shots outside the 18 I give it a 0.03, inside the 18 gets 0.11, inside the 6 yard box gets a 0.3, and penalties get 0.7. This creates a chart that looks like this:
What you see here is that Arsenal started slightly better, Stoke got a big boost from the penalty, then Arsenal surpassed Stoke after halftime and never looked back. Stoke matched Arsenal around the 80th minute, but then leveled off and Arsenal finished the match strong.
I could do the same thing for pass quality. Here I weighted all passes (0.01), final third passes (0.1), key passes (0.3), and through balls (0.7).
Arsenal’s pass quality (attackingness?) took a huge bump in the same pattern as the shots quality. Stoke matched Arsenal pretty close for the first 30 minutes of this game, but then Arsenal put on the afterburners and got what I consider a well deserved win.
Anyway, let me know what you think of these new charts in the comments below.
Match Stats
6 – Shots by Theo Walcott (led all players)
2 – Shots by Cheerful Charlie Adam (led Stoke)
1 – Shot by Charlie Adam from the penalty spot
1 – Shot by Charlie Adam from kickoff
1 – Blatant stamp by Charlie Adam on Alexis, which allowed Iwobi to get the ball and score
1 – Shot by Mesut Özil, which was Arsenal’s only big chance, which was a header, which he headed over the keeper
77 – Passes by Xhaka (led all players)
30 – Passes by Imbula (led Stoke)
5 – Key passes by Alexis (led all players)
2.7 – Key passes per game Alexis is averaging this season (leads Arsenal this season)
2.6 – Key passes per game by Ozil this season
4.2 – Key passes per game by Ozil last season (led the League)
2.1 – Key passes per game by Alexis last season
4 – Successful dribbles by Walcott (of 7, led all players)
3 – Successful dribbles by Ox (of 3)
8 – Ball recoveries by Xhaka (led all players)
7 – Ball recoveries by Imbula or Adam (led Stoke)
3 – Tackles by Xhaka (of 3 attempted, led Arsenal, Alexis also had 3 but was 3/4)
17 – Passes from Coquelin to Xhaka (led all players)
13 – Passes from Xhaka to Ozil (the second most common pass combination)
1 – Great save by Cech to deny Penis Crouch a goal
1 – Place where Arsenal sit in the League table
1 – “Achievement” which Stoke have accumulated over the last 45 years and which they sing a song about whenever they play against Arsenal. That moment of glory? Breaking Aaron Ramsey’s leg.
@7amkickoff
Sources: 442 Stats Zone App, WhoScored.com, that basket of deplorables known as Stoke City supporters
Bonus Stat
20 – Minutes it took for Man City to shit the bed
Who honestly thought that we’d lead Pep’s Citeh by mid-December by four fucking points?
To be on the safe side now, we must sack this Venga guy, else we risk winning the league.
Me. Pep is shit, I could win trophies with those Barca and Bayern squads. I mean the Man City players dont have a clue what they are supposed to be doing.
I am loving that me prediction that he would be crap in the premier league is coming true. Also delighted that he didn’t end up at Arsenal like most Arsenal fans wanted.
Funny how Bayern are struggling this season when supposedly anyone can win trophies with that squad. Dump on Pep when he loses, but never giving him credit when he wins. I don’t understand this anti-intellectual culture that the British media have created in this country. Why do people want to see Pep fail so badly? Just so they can say that his style doesn’t work against teams like West Brom on a wet and cold Tuesday night. People want Guardiola to fail so so so much in this country when they shouldn’t really have anything against him. He’s hardly a… Read more »
“Dump on Pep when he loses, but never giving him credit when he wins.”
Ummm what?! Guardiola is one of the most lauded managers in world football. Whenever his sides have won something it has always been because of Guardiola and not in spite of him. Same with Klopp. Every win is because of Klopp. Every loss is because his players let him down. It’s just pathetic.
Yes to the graphs would love to see some comparative stats after a time too
Awesome graphs. Interesting to see pass quality per player over time (don’t ask don’t get…) If ever you have the time.
I like the idea of the graphs, especially the shooting one (you could argue that the passing one would look similar most games). One refinement would be to add on a dot/ball when each goal is scored. That would allow us to see the bigger picture of how the game has changed eg a goal has been scored and the other team spends a spell on attack. Or that we scored and still dominated.
It would also show efficiency of goalscoring v shots.
Nice. I can’t decide if the incremental points approach (line) of the shot quality is useful. Versus a potency bar for each block. Bar would show zero, that we did nothing for 10-15 mins in the first half, what the flat line shows, although the line shows we’re on 0.3, but so what? Though it shows that at 80 mins both teams were level. So I can’t decide. HOWEVER a bar for the passing IS useful, because tracking a consistency figure for passing is very insightful regarding game control ebb and flow. Love your work. What I do know is… Read more »
Loving Xhaka at the moment, he is fulfiling all my CM/DM requirements at the moment. His game has an element of Xabi Alonso to it and I’m Mm Mm loving it!!
Also that was never a penalty
very apt comparison to alonso.
I think the passing and shot breakdown is great, and it means I need to read less. That said, through balls seem a little disproportionately high to key passes. They (through balls) may be rarer, but key passes still are potential assists. Maybe 0.4 and 0.6 would be fair?
Through balls are statistically the most dangerous pass in the game.
As far as the charts go, I think a running average/density would make much more sense. Right now it’s a bit hard to read what’s going in separate parts of the match. Plotting possesion together with shootyness would also be an interesting one.
Anyways, where are you getting your stats from? I should be able to whip up some interesting plots as examples
@Reidar do you mean average/density every 5-minute interval or average/density over the entire game? I honestly love the chart as it shows whether the team has dropped, maintained, or improved quality over the course of the match versus the opponent. It would be interesting to see the chart when the team plays against big teams e.g. Barcelona or Bayern. I guess my only comment on the chart is the points against the quality of passes, which will need tweaking over time as the system is used more frequently, as currently there is not enough to justify the difference in the… Read more »
That data is available on whoscored.com
https://www.whoscored.com/Teams/13/Fixtures/England-Arsenal
Right now, I am using the slider and hand coding the data.
The new graphs are interesting . It shows passing dominance and goal threat in similar graphs . The Leicester ManC match would be interesting on the graphs I imagine . Sterile dominance vs clinical counter attacking ?
One change you could make would be to add the major events (goals , red cards , half time ) . Thus we can try to find a narrative – stats are just used to spot a pattern and make a narrative after all ?
Another thing would be to see if certain injury/substitution changes the slope of either graph substantially …
COYG
Really dug the graphs – especially if I didn’t see a game, like, it was 2-1 for a long time, were we patiently stroking the ball around or hanging on for dear life? The graphs help answer that.
The graphs are really cool, good idea. I wonder if you could do a version of both that’s weighted for possession or maybe even territory? Often it looks like the other team has the best chances while we suffer from “sterile domination”, to quote Wenger.
Thanks as always! These new graphs are pretty amazing. I was a little surprised to see how close both teams were on the shots chart, until Arsenal really pulled away.
How did you devise the value system for the passes exactly? And did it turn out as you planned?
Any interest in retroactively applying them to a few previous games even? Thanks again! Go Arsenal!!
Penis Crouch is the main highlight of your summary mate. Nicely done. ???
LOVE THE GRAPHS!!! Data visualisation FTW!
1 – Goals Charlie Adam allowed because he was too busy stamping on Alexis rather than stopping Iwobi. What an idiot.
@7am
Love the new chart. I say keep it going for several games and see what kind of representation this chart will make.
BTW, are those shot on goal and completed pass or not?
Totally unrelated but Wolfsburg just announced that Julian Draxler won’t play for them again and will be sold next month. I’ve been hoping he’d join us for the past 3 years, but looking at his unprofessional behavior I don’t think the Boss would be interested. Thoughts?
I think Wenger’s plan for Draxler was to convert him into a center forward. Now that we have Alexis tearing up trees in that position, Draxler is at best a bargaining chip to get the Chilean demon bunny and his faithful canine sidekicks to stay in London for the foreseeable future.
He’d be a great player for us but we’re already very strong in those positions. You could argue that it would “kill” someone’s Arsenal career but who? Ox and Iwobi would look the most vulnerable, even Welbeck who might find the team has “left him behind”. Tricky one for the manager, but my instinct would be to bring him in and let them all fight it out!
Am I the only one who noticed ‘Penis’ Crouch? Or did I miss a reference?
Pretty sure stats just doesn’t like Crouch and who could blame him?
Crouching Tiger, Hidden Penis
Lanky Crouchy, Total penis.
Bit harsh on penises really.
Why was Iwobi’s goal not a big chance, wasn’t he 1v1 with the keeper?
Wtf is penis crouch hahaha
That’s a really good chart. Gives a very good overview of the ebb and flow of a game.
Fan of the new charts Tim. Always appreciate your work and so glad it’s done for the red and white. They remind me of the scoring charts ESPN does for the post game basketball reports. Very helpful visual, and I’m curious to see what your numbers show as you get more data.
Cheers from an 8amkickoffer
Ozil’s face always seen so black and blue, seems like he had a boxing fight.
Ozil is assisting by scoring goals.
Love the new stats and graphs. It gives a different view of the game than just shots, shots on target, possession and territory.
why not just plot it as given value instead of keep adding it all up? then we can see the frequency of better shots/passes over game time.
then how about plotting the shots over passes so we can see how effective the team’s possession?
Hi – Yuur charts. LOve it. It would be good to have some idea – or even a very good idea of volume so that when we/they are making better shots/passes more of the time we can see that dominance.
You should add number of defenders between the player that is taking a shot and the goal. On this, it seems Crouch’s chance was a guilt edge, but it really wasn’t as he had to climb over two players to get to the ball.
Graphs are really nice. First one is more shot distance than quality but I guess distance isn’t a bad proxy for quality. Would be really cool to so this over a long run of games and then calculate the average pass quality/shot quality for each game and make a graph for the season on that! Definitely though put any sort of statistics into graphs and I’m there. Really like the idea.
Basically what you are trying to show here is momentum and for that may be merging both is a better idea or atleast multiply shots with team’s shot accuracy of the season. And i think for attackingness you should subtract back passes or defensive passes too cuz otherwise a team will always have a better graph and ‘their manager’ will claim that league table/their position is unjustified.
Just a quik one in the Ramsey chant.. Is that a real thing??..
Graphs are a great idea. Maybe extras if you have the time. Thanks
Graphs look good but it would be interesting to see the same data NOT accumulated. Over time you would see if Arsenal’s shot quality peaks in any areas, like the final 20 minutes for example
Would rather have more players with high assist stats than one with record breaking numbers.
As for the graphs, well, can’t have enough, but only supported by analyzes. Why were we equal in exp.goals at 80 minutes? Didn’t seem such a close match?!
It was Even Stevens for a short while around the 75-80 minute mark, and for a few minutes there was as much chance of Stoke scoring as there was of us scoring, but then Right Charlie Adams happened, and Stoke got depressed and gave up trying.
The penalty gave them a huge boost. Unlike some xG people, I think pens should be included in the analysis.
For Christ sake do they find pleasure in breaking Aaron’s leg? What a group of twats
How do you pull such detailed raw data? I don’t see an option for time-series info anywhere on fourfourtwo, which I imagine is the site compiling the information
17 – Passes from Coquelin to Xhaka (led all players)
I should bloody well hope Coquelin did lead all players in the Coquelin to Xhaka passing stat
He had a very up and down game though… he was rubbish in the passes from Xhaka to Ozil stat. 🙂