There were two distinct periods of play for Arsenal this season: the first 30 League games where Wenger played with a back four and the final 10 game stretch of the season, which includes two FA Cup matches, where Wenger played with a back three (five). Here is a chart with the data:
The first thing that stands out is the remarkable similarities between shots taken (14.9 v. 14.4) goals scored (1.9), Shots in Prime allowed (3.2 v. 3.4) and Big Chances allowed (1.5). For those who don’t know, Shots in Prime are shots taken within 11m of the goal line in a hemi-circle from the penalty spot and Big Chances are shots taken from extreme close range or with very few defenders (usually 1v1) in front of the goal.
A note on methodology: I compile all of my own shots stats. I use the 442 Stats Zone App after every Premier League game to map where the shot was taken from and whether Opta marked it as a “big chance” or not. Using that data, collected now over three years, I find that Big Chances are the most important shot in the game. They are scored at a 50% rate and they account for 50% of a team’s goals. Shots in Prime are not really correlated as well to goalscoring and once I remove the Big Chances I find that they score at about a 10% rate. This is very close to the bulk average for all shots across the pitch. Shots from outside the 18 yard box are the most numerous type of shot and the least effective, scoring at a 3% rate. Now, back to the chart.
Where there are differences are in the attack. Arsenal created 0.8 more Big Chances per game with the back three than they did with the back four. This leads to an increase in Shots in Prime as well, since those two numbers overlap. These numbers lead to an increase in expected goals per game from 1.9 to 2.2. That doesn’t seem like much but over a 38 game season that’s 11 goals and would probably have been enough of a difference to ensure a top four finish. So, the back three, which is typically thought of as a defensive formation, actually led to a renaissance in Arsenal’s attack.
This also shows up in Alexis’ goal-scoring and assists. Alexis scored or assisted 9 total goals in the final 10 matches. In the previous 10 matches that number was down to 6 and included a stretch of 6 matches with just two goals.
Further, when I compare ONLY the matches contested between similar opponents (Chelsea, Man City, Man U, Leicester, Middlesbrough, Stoke, Sunderland, Everton, Tottenham, and Southampton) we see once again almost no change in the defensive stats but a significant improvement in attack. Even more improvement in attack than against all opponents.
Arsenal created 1.3 MORE Big Chances with a back three against the same opponents than they did with a back four. And the resulting better shots selection increased Arsenal’s expected goals for by 0.7 per game. Those are remarkable numbers and I feel comfortable saying that it looks like the back three significantly improved getting the ball forward.
If I’m allowed a bit of pontification, I’d suggest this was down to putting Ramsey in the middle. Ramsey is a much more attack-minded midfielder than Coquelin or Elneny and while his runs into the opposition box make some fans crazy (me) this evidence suggests that they may improve Arsenal’s attack.
But where the back three didn’t make any improvement was in Arsenal’s defense. This is exactly contrary to what most fans have expressed in the forums, on twitter, and to me personally. Now, before I go any further, you have to be consistent here. You have to either accept the shots data both for and against Arsenal or not accept this data at all. It would be inconsistent to say that Arsenal’s attack improved but then discount the same data to say that Arsenal’s defense also improved.
Similarly, it would be inconsistent to cry “small sample size” to argue against my analysis of the back three while also saying you think the back three worked well. The “small sample size” argument cuts both directions. I don’t believe that 25% (10) of a 40 match set is a small sample so I am comfortable with my conclusions.
Back to the defense. There are a number of problematic stats here. This first is just total number of shots. Comparing like for like, Arsenal allowed 2.8 more shots per game with a back three than a back four. The number of Big Chances allowed per game stayed steady but Shots in Prime went up marginally. As a result, I expected to see more goals allowed, about twice as many.
“Ah ha! But we DIDN’T! This means the back three worked!” you might say. Ehhh, no. Instead, what I saw was a massive increase in the shots saved by Cech and Ospina – an increase over both the 30 game larger sample and the like-for-like sample. Those 46% Big Chance saves and 69% SiP saves are astonishingly high. Hugo Lloris was the best keeper among the top six teams, he saved just 39% of the Big Chances he faced. The average across the top six teams was just 31% with David de Gea saving just 20%. And the highest SiP saves was Man City with 56% and Liverpool and Man U saving just 33%.
Using the shots faced data and my own expected goals formula I gave Arsenal’s keeper either a + or – rating on each of the 40 games I sampled this season. +1 means that the keeper saved 1 expected goal, -1 means the keeper failed an expected save.
This chart shows a remarkable increase in Arsenal’s goalkeeping in the final 10 matches of the season. Not quite as good as the first 10 matches, but consistently good at the end of the season with no negatives.
What I would expect to see from an improved defense would be a reduction in the number of shots, the number of big chances, the number of shots allowed in prime, and thus a reduction in the expected goals allowed. What I see instead is simply an improvement, an unsustainable improvement, in Arsenal’s saves.
One final statistical note. Arsene Wenger is a huge proponent of scoring first and scoring early. He has said many times that the team who scores first goes on to win most of the time (and he’s used different percentages, but they are all above 70%). I can’t confirm this stat but what I can confirm is that in the first 30 League games either Arsenal scored or conceded on average in the 29th minute. But when Arsenal switched to the back four, three the first goal scored or conceded drastically changed and moved to the 49th minute on average. This is quite remarkable and there could be a number of explanations that exclude the back four/back three debate – late season fatigue or dropping Mustafi are two examples. But it is true that in the 10 game set directly prior to Wenger changing to a back three, Arsenal conceded the first goal 6 times and did so on average in the 11th minute. And Arsenal lost all 6 of those matches.
It doesn’t look like the back three made Wenger’s Arsenal more cautious, their attack remained steady. Nor does it look like the back three made Arsenal more solid in terms of shot concession. But it does look like there may have been a psychological boost to the team, which when coupled with an increase in Keeper prowess cut out the early mistakes and gave Arsenal some room to breathe life into a dead season.
@7amkickoff
Stats: my personal database
The stats only suggest that we become serious about winning games when top 4 is at stake.
Exactly, it’s all about the effort put in by the players and the fact that the manger decided to step up when top four chances were in danger.
And what Tim has not mentioned is that when we switched to a back five(three?), the defensive stats are eerily similar to the usual back four but this new formation was at its infancy. Getting used to the formation takes a while and these 10 games should be considered more as a developmental phase. Considering that aspect, I think we have done remarkably well in the final 10 games and we should improve in the new season when we have trained all through the summer. At least, we can hope(until Mustafi lunges high up and reverts us to a back… Read more »
I have said many times here that formation doesn’t mean anything until the defensive organisation of the team improves. Teams who defend well will know how to defend against all teams whether it’d be with a back three or back four. Juventus beat Barcelona with the same formation arsenal played for last decade 4231. Even Tottenham with back four comfortably beat arsenal with a back three. Back four formation Was not the reason for our amateurish defending against west from goals. The late resurgence after the formation change has more to do with our oppositions and a bit of luck… Read more »
I completely agree with you, Clins. Just watching the game you could see that we were very sketchy – very lucky to get wins from some of the games. Look at the ‘Borough game, for example, they were a nothing side but we weren’t playing at a level you would want a challenger to play.
I admit to not reading every word of this article. That said, it seems like the big difference is that we won the matches with a back 3 and didn’t with a back 4?
And what Tim gives here is his opinion on ‘Why’ Arsenal won those games.
Going nearly half the season unbeaten with a back 4 shouldn’t be so easily dismissed.
Well said, I find it sad so many supporters forget how well we was doing after the Liverpool game and until Cazorla got injured.
But you have to consider we win games every season when top four chances are in danger. we have made a habit of really turning the screw when we’re in trouble. I’m not sure how much of this is formation or application.
What also needs to be kept in mind is the opposition faced in the last 10 games. We had some of our toughest games towards the end when we faced Spurs, City, Chelsea, Stoke, utd etc.
Stats may not show the whole picture completely. All of us could see the solidity in the team and the way the system suited the playing style of so many players.
This is a good point
Exactly
Yeah, I think he kinda thought of this, which is why his second chart compares the last 10 matches to earlier matches against those same 10 opponents.
I’d like to know how easily dribbled we were through the middle post back-3? Prior to that, we all know our 2-man midfield would regularly get run over. This was made worse with Ramsey’s regular absconsion and Xhaka’s lack of mobility. It seemed the back-3 allowed Ramsey to galavant on his opposition box adventures by having one of the ball-playing CBs step in as a DM next to Xhaka, meaning he wasn’t the one responsible for holding the fort and could focus on dictating play. Basically if the evidence shows a back-3 brings out the best in Xhaka and Ramsey… Read more »
Who on earth votes this up? “I never read it but here’s my view on the article.”
If anyone wants my view on A la recherche du temps perdu, just ask. I haven’t read it, but…..
Leicester proved last season just how psychological football is. Wenger knows this too, perpetually comments more on it than tactics. Im sure he knows these findings all too well. Interesting to see.
Season before last* (time flies)
Quite agree. Plenty of other high profile examples over recent years; pre vs post Mourinho era at Chelski as well as pre vs post Ferguson era at Manure. Players are ‘only human (after all), don’t put the blame on’ them. Its down to the Manager and their ability to inspire/fire-up the players and to a lesser extent the impact on team psychology when things aren’t going well
Its not up to the manager, Leicester clearly lost it last season until Ranieri is fired. But I do agree that Ranieri has a good record in signing some of the best talents, inc Lampard, he was also the reason that Cheslea signed Robben and Drogba. You can see how bad they played without Kante, Vardy, Mahrez has also failed reporduce that season
Hello,
We were playing much better at the start of the season. If that form continued we wouldn’t have gone to the back 3. I would assume the 1st half of the season numbers are probably higher than the full average. If so, it would be interesting to see the last 10 games before the change to a back 3 vs the period with the back 3 as that was the performance period Wenger was trying to improve on.
I remember conceding shots with the new system, but I don’t recall them being very troubling. Not all “shots in the prime area” should be considered equal.
He did include xG against. Seeing that most of your xG comes from shots in prime positions, you could just take that to gauge how good the shot situations were.
The back 3 helped us get the ball up the pitch and generate offense more efficiently. We’re Arsenal, the defending would still be dodgy if we had 6 at the back. The engine had stalled and was in need of a boost.
interesting article and analysis…..never the less i think from my own experience of watching the games arsenal did seem to be more solid defensively. stats are important but don’t always tell the whole story….there was a calamity to our defending before the back 3 was introduced that made chances easier to score….that’s just my observation…..and regardless of where shots where taken from i think the goal keepers improved in part because of better pressure on the ball and less area of the goal to shoot at due to more defenders and better shape overall. the end of the season wasn’t… Read more »
Great analysis. Let me echo some previously mentioned constructive points. 1) The back 3 afforded some defensive security that may not be captured by raw stats. The back 4 effectively devolved to a back 2 at critical parts of games and season. Our wing backs are given such license to ‘bomb’ forward that the 2 center backs would consistently be 1:1 on the break. Therefore, the opportunities created against us were much better, so all shots are not equal. 2) We created more opportunities for multiple reasons. First, the wing backs provided natural attacking width rather than having a left… Read more »
The stats didnt take into account the fact we change a major tactic at the end of the season and how fast we adapted to the back 3. Imagine if we had continue another 10-20 games, or had practice it during the pre season.
Strong stats for a newly adapted tactic for sure. We also missed big players like Cazorla, Koscielny, Mustafi through the end of the season
You “believe” 10 matches is a big enough sample size? Fortunately, when it comes to math, we don’t need to rely on beliefs. Ten matches is too small a sample to draw conclusions with much certainty. Moreover, your conclusions assume all other variables affecting outcomes were equal, which they were not. So, take it all with a pinch of salt. These are interesting numbers to consider is all.
So, it seems the sample size seems to be big enough that YOU can talk with absolute certainty? At least Tim admits the flaws and biases in his analysis.
My analysis only looks at shots and goals and not actually to outcomes. I am highly confident of the shots data correlating to goals. However, if you’d like to do an analysis of the other factors and see if there’s a correlation between say, air temperature and goals, or negative comments on twitter and match outcomes please feel free. As for sample size, I know exactly the mathematical problems with these samples: individual matches are highly problematic (though it’s funny how people will “rate” a team based on 1,2, or even 3 matches), 10 games are problematic, and even the… Read more »
One last idea: 10 games is a misnomer. I’m actually looking at a 181 shots sample of 958 shots set. That gives a confidence interval of +/- 6.5. Which isn’t bad.
Just a suggestion, but isn’t the most relevant comparison that between the late January-April malaise, and the period following the introduction of the back three?
The team was in a mess, and the shift to back three did seem to make at least a qualitative shift. Including the data from August-December seems to obscure that a bit.
Exactly, the whole point is surely that the shift in formation – albeit initially haltingly – helped us to achieve a significant improvement in results and performances relative to our desperate form of the spring. That there isn’t a stark difference compared to our early-season unbeaten form when we were still in the title race is hardly surprising or a searing insight.
Cat: that is an excellent point. Clearly the back 4 of the early season was not the broken mess of the matches leading up to the switch. The season was really a story of thirds: initial excellence, the ides of ineptitude, and the switch of systems. While small sample size is clearly a problem for stats like big chances created/yielded being treated as anything but tentative, the indication is that the switch helped and is worth persisting with unless we can find the cazorla-inspired balance of the first half again. Breaking the season into three sections instead of 2 would… Read more »
I’d love to see a comparison between the last 10 games and the 10 immediately before them. I know they are small sample sizes but what the team was able to achieve playing with a back four in September/October wasn’t the same after January.
I’d also suggest it’d be interesting to see how we performed against big teams (prem top 6, psg, Bayern). Our results against these teams improved markedly after the change of formation. I wonder what the stats would show.
Maybe this is wrong, but I know the numbers are an average over the last ten games. I wonder if we were worse at the start of that run, as everyone got used to the formation?
Did we concede more Big Chances & Shots in Prime during the earlier games than the late ones?
It’s just the Tottenham match early where Arsenal truly struggled, most of the remaining matches are very similar. However, your suggestion we got better is belied by Everton who took 22 shots at Arsenal and I thought should have had more than one “big chance” credited to them. The other oddity is that Arsenal scored early in the final two matches of the season.
How many of Everton’s 22 shots were after we had to play a back four with ten men?
The point of the back 3 is to have an extra player in midfield. If you can have the same defensive performance with one player moved from defence to midfield, the team as a whole will improve and the result will be better offence.
In other words the move to back 3 worked, but not in the way that people might have assumed.
Other teams might move to a back 3 and have the same performance in offence but better defence.
Also, having three central defenders clogs the middle of the box up, so even if you get a prime shot, the shots may be a bit more difficult and thus easier to save
And this
An extra midfielder?
I’m pretty sure that Wenger moved a player from midfield (Coquelin) into defense (Holding). Wenger’s fullbacks have always operated more like wingbacks and his midfield has typically operated with just one deep. He kept all that the same, changed Bellerin for Ox and allowed Ramsey to play as a 4th forward.
This formation was, for me, just dropping Coquelin and adding another center back.
I would argue that the wingbacks are more like wingers than fullbacks, even when considering how Wenger likes attacking fullbacks we can see how the wingbacks have not had the same pressure for def duties, yes required to defend but more like defend like a winger than a fullback. I will point towards Monreal and his form at LCB, being able to go over and perform LB duties while Gibbs got back and helped to defend, moving into the CB position when timing allows. As for dropping Coquelin for Holding (example, I know), I see it more like a drop… Read more »
The back three allowed the wing backs more freedom in attack and allowed the likes of Sanchez to defend less and become more a supporting striker than side midfielder and it improved his out put from the same position. Ozil also improved cos the number of players around him whenever we attacked improved too. We won the chelsea match and out played them cos ozil was a fifth midfielder and bellerin because of his speed and improved match reading covered the right flank well and it freed up ramsay. I think the team played better because they could attack without… Read more »
As the bloke said on Gunnerblog’s “flamingo shirt” video the other day. Agree.
I think maybe the outcome depends on the fullbacks at your disposal as well as potential changes in personel that comes with switch in formation. Our fullbacks are technically gifted and fast (plus we used a winger on the right some of the time). With their pace, especially Hector, they can recover when we lose posession and thus have the luxury of being more offensive minded. We also put Ramsey in the middle instead of Coquelin. The sum of it was better attack in our case. With other options we might have ended up with better defence. Certainly Wenger always… Read more »
I think there’re several reasons why the back three helped our attack.
1)Build up is much better.
2)(As noted in the article) Ramsey has more freedom.
3)The positioning of Özil and Sanchez. I think the position behind the striker suits em both better than playing as a Winger/pure number 10.
Do Xg and Shots allowed take into account the number of players between the shot position and goal? Even though we may have been allowing a similar number of big chances they’re going to be more difficult to score from with more men behind the ball.
Could this be an explanation for the drop in goals against?
I tip my hat to you, blogs sir.
What a well researched, and well written statistical masterpiece.
It wqs @7amkickoff, not me
Surely the comparison should not be with the first 30 games of the season – but from the point when the wheels came off, that is after the Chelsea home game?
Wheels came off at Everton away. But agree it would be good to see that analysis.
For me the use of stats to compare these two periods of the season is the definition of apples and oranges. During the most temultuous period of our recent history any analysis has to include the psycological aspect. If the club was in harmony during this period and all that was changed was the formation, or Ramsey being moved, then I would be far more willing to give weight to the stats. We have heard stories of the players having to concentrate more because everything was different with a back three. Perhaps if the manager had managed to incite a… Read more »
Fascinating numbers. Out of curiosity did shots on target allowed and blocked shots differ from the back 4 to back 3? Wondering if that helped keep down the actual goals conceded.
Also would the rest of the league serve as a baseline? E.g. if the entire league had a trend of higher number of shots in the latter half of the season, that may change the reading of the before-after comparison.
Huh, interesting question.
Yes, Arsenal increased the number of blocked shots from 2.4 per game to 3.5. An increase of 1.1 per game.
0.7 of that increase were from shots outside the box
0.5 of that increase were from shots in the box but not in prime
Arsenal also blocked 0.1 FEWER shots in prime (negligible)
However, when I look at the like for like opponents the numbers are EXACTLY the same between back three and back four.
Thanks 7am for another well researched, thought provoking and informative piece.
Almost everyone focused on the defensive element of the back 3 except for the manager who spoke mostly about its psychology and the attacking benefits.
Despite the acknowledged small sample size with both keepers improving in big chance save percentage so drastically do you give any credence to the popular idea that a qualitative increase in difficulty of those big chances occurred with a back 3?
In those final ten games we had some of our toughest opposition. Chelsea, city, United, sp*rs, Stoke and Southampton away. To compare a varied sample size of 30 to a small sample size of our toughest opponents is not statistically helpful to form any sort of solid conclusion. Anyone with eyes could see we were generally more defensively solid with a back 3.
Re sample size, just because you think it is big enough doe not make it true, imagine telling that to the FDA. My guess, as a scientist, and an actual professor, is that none of those differences, excepting maybe the goalkeeping, are not significantly different. Does not mean the differences are not there, just that your sample size is not big enough to prove it with any sort of statistical certainty.
* uses phrases like “statistical certainty”
* claims to be a scientist
I’ve more faith in 7AM’s analysis than the FDA. Hands down.
I’m least confident of the keeper stats because saves and goals are the least numerous of all my stats.
I also make three samples of the season here: 30 games, 12 game like for like, and 10 game set with the back three. I can’t think of another way to sample Arsenal in 2016/17 where the central question is whether the back three made Arsenal more defensively proficient.
But 10 games is a misnomer. I’m actually looking at a 181 shots sample of 958 shots set. That gives a confidence interval of +/- 6.5. Which isn’t bad.
Thanks.
Use fisher exact test on your data. For example, if you compare 50% saves from 10 shots in prime to 25% from 40 shots your fisher matrix would be
5 5
10. 30
The p value would be .143, which means a 14% chance the difference is random. FDA requires 0.05 for most studies, but many people think that is too high.
You’re getting a little pedantic here. Whether or not 10 games is acceptable as a sample size statistically-speaking, we nevertheless like to make claims about what switching to the back three did for the team. The major narrative so far has been that it made us better defensively. 7amkickoff is just using some numbers to say, hold on, that may not necessarily be the case. The sample size is the same vis a vis our experiment with the back three regardless of what conclusion we make about it. The only other possibility is saying nothing about it at all, and… Read more »
“The p value would be .143, which means a 14% chance the difference is random.”
You don’t understand p values. They are a measure of the probability of the data given that the model (null hypothesis) is true. They are not a measure of the probability that the model (null hypothesis) is true, given the data.
I really like 7am Kick off’s analysis but I think this report also illustrates how stats can also be a ‘myth’. I’m sure the stats would show that both Walcott and Giroud are efficient and effective goal scorers but I’m also sure most of our eyes will tell us different.
I must agree though, I thought Cech was pretty sensational in the final 10 games and in most of those matches I never felt we were going to lose as I usually do. I suppose this highlights that psychology in sport is impossible to measure.
That’s the beauty of perception and the reason statistics exist. We see what we want or expect to see. Still perfect examples you’ve cited though.
Thanks 7am. The 6 defeats thru early goals was an interesting point. Last season, as counterintuitive as it was, every time we let a goal in early I rarely felt we would win. I put it down to the psychology that once breached we continue to feel vulnerable maybe because of our own awareness we are an attack minded team. That psychology (except at Bournemouth) actually inhibits our attacking play reducing clear cut opportunities. How many times did we feel confident about recovery after going behind? Rarely I think. We always need to score first and preferably early and hence… Read more »
So we didn’t come close to perfecting the back 3 but it did help to refocus the team psychologically, we were more patient offensively (goals were scored later in games), we spaced the pitch better (Xhaka running the show) and played more vertically in attack, and Alexis, Cech and Ospina deserve more credit for those critical goals and saves.
I think the team that finished the season would have given Bayern a couple of proper games.
I personally feel that the biggest reason why Arsene changed the system was that Bellerin, Mustafi and Walcott were both out of form at the same time. And in 10 games prior to the switch we got pinned a lot from right side. Arsene could have put Ramsey on the right but he with Cazorla out he, Arsene did not had that luxury. Now you can not remove all these players at once and still put a team that fits in the system because of the lack of cover at RB and Gabriel doesn’t provide you the pace to do… Read more »
Good article.. you were teasing us with the stat for some time but finally a complete article! Those save percentages are clearly unsustainable.. could playing more in this system improve our defensive numbers??
It’s interesting that you say the positive save stats are unsustainable. the stats for the first and last ten games are similar (generally more positive than negative) accounting for 50% of games. Seems to me that confidence runs through a team and lifts individual performance. Had the team played with confidence in the awful mid section of the campaign perhaps (Arsenal) chance conversion and save stats would have been more positive too?
For me, it’s not so much “small sample size” as it is that you’re comparing 10 games in a new system to what’s been done at Arsenal for the bulk of 20 years. Working into a new system takes time, so it’s hard to compare what the boys did to what they’ve been doing for YEARS. I’m not saying that the stats won’t be similar at the end of a full season playing this way, but I do believe it’s a bit early to say 100% either way for now.
One thing about the Shots in Prime and the increased saves ratios that I feel you overlooked is that SiP only looks at where the shot was taken and not the quality of it. With a back three there is a much higher (generally speaking 50%) probability that the attacker is not taking the shot unchallenged and this will particularly affect headed finishes. I do agree that 69% is extremely high but a full analysis would require taking the quality of the finishes into account as well. Also, ten games IS a small sample size. An average of 1.5 big… Read more »
This article though well-made and informative has successfully managed to dampen my upcoming season’s back 3 formation xcitment somewat ?
Nice analysis and much appreciated. Of course it can never take account of intangibles such as perhaps Wenger actually did some defensive drills, told the players to, you know, mark a bit and on the offensive side hit corners in the opposition box a big higher than knee level.
Not a criticism of you , Tim, but of our glorious leader.
Interesting read, tho my feeling is Wenger has no tactics and doesn’t understand defence. He’s never bought a top defender, inherited Adams and co and would never have payed money for Campbell. Arsenal won’t win the big things under Wenger.
I suppose he lucked into Koscielny and, though maybe a touch early, Holding as well? I’m not saying there aren’t any problems in defense, but to say that he’s never bought a top defender is quite reductivist.
The same size problem comes in when you assign all “shots in prime” or “big chances” the same weight. There are shots in prime, then there are shots in prime, if you get what I mean. Some of them you should definitely score, some of them you shouldn’t. You could give away 5 shots in prime in a game but still be better defensively than a game where you give away 1 shot in prime, if the 1 shot in prime is a really big chance and the other five may be headers at tough angles or ridiculous volleys that… Read more »
What i saw was that with a back three we have less last ditch tackles and desperation defending because we were less likely to be beaten over the top on the counter.
It also frees Ozik and Alexis up from tracking back as much so leaves them more freedom to attack
Nice piece, cheers
Very interesting Tim. I suggest actually calculating p values to give a sense of how statistically significant your findings are. I suspect that you will be disappointed.. You should probably calculate the power statistic to get a sense of how big a sample size you need to get a statistically significant difference. This might help with how you frame future arguments. High p doesn’t invalidate your findings, but it pushes them from dispositive to merely interesting or suggestive. As with many pursuits, football requires decision making under conditions of uncertainty. A statistical inference, even a relatively weak one, can reduce… Read more »
Data scientist by profession? 🙂
Another possible reason for the team’s better performance with a back 3, particularly in offense, is the “novelty” of Arsenal playing with a different system. A number of teams seemed to have figured out how to stop an Arsenal offense which had become quite static and formulaic. Perhaps the improved offensive performance isn’t necessarily due to a back 3 being better than a back 4, but due to just a “change”, something new for opposition sides to figure out or deal with.
Best comment of a very good thread.
Missed a key thing, the period with Cazorla and the period without in the back 4 part. Just to point out what I mean, AFC kept what… 50% clean sheet with the back 3? This if I am not mistaken is the same ratio for clean sheets with the back for and Cazorla playing. However we started to let in goals more often and found it harder to keep clean sheets once we lost the little magician who could turn a def movement into an attacking one, instead it became a def movement after a def movement and that pressure… Read more »
The biggest difference moving to 3 at the back was in every game we simply looked much more like we wouldn’t concede. I felt comfortable in every game that we were fine and not going to let one in.
Love your analyses – thanks!
Interestingly, the first reason Wenger gave for changing to a back three was that it gave us more opportunities in attack. So you’ve “proved” he does know.
The only thing I really don’t understand is why, when you put so much weight in creating big chances, you don’t like Ramsey’s runs?
Surely the shape with a back 3/5 leads to less of the goal to aim at regardless of position? I assume this purely on the basis that 3 central defenders narrow covers more across the penalty box compared to two.
The extra man also cover for fullbacks arriving late from transition.
Great analysis, I do have a question though, when you ran your numbers did you happen to split the numbers between top 6 match-ups vs not top 6 match-ups? I’m wondering if splitting out teams where the quality of their play is roughly the same as Arsenal’s would show anything interesting? I think the spuds match will still look bad but Arsenal were not fully used to the system then (imo) but I do wonder if we reduced some chances somewhere along those lines? It may be cutting the sample size up even more but it could show something.
Really interesting analysis. Thanks for expending all the effort 7amkickoff – columns like this are the reason this site is my regular go-to for arsenal related commentary.
I like the back three plan a lot!
I also think the Ox is the dogs bollox in that wingback role.
Interesting stats. What would be great is if there was a way to measure the distance of the nearest defender from the ball before the opposition took a “prime shot”. It seemed to me, at least at a superficial level, that we didn’t leave as much space in the back with 3 CBs. So perhaps the presence of an extra body in defense caused the opposition to either hurry their shot or made the goal-post smaller so to speak? If we were able to reduce the time and space for the opposition to shoot, that certainly would, at least partially,… Read more »
Interesting read but given the 20 game unbeaten run at the start, and really poor form following I think it’s difficult to quantify which formation is best. What it did do, as wenger suggested, is focus the team and the change helped shake the team out of some really poor form. It was good to see and I liked seeing ox, Rambo, holding show some good form. I do think however the big difference was concentration and work rate, and if that same effort was applied in the previous ten games then we would have stuck with 4 at the… Read more »
Pretty sure we hardly got hit on the counter once playing a back 3, which to me was our biggest and most taken advantage of weakness against the bigger teams. I bet a lot of those shot stats whether saved or not came through a crowd of players and out of frustration.
I’m hoping Chambers stays at Arsenal and gets a chance alongside Holding.
Out of all the players we sent on loan he’s the only one to have his stock rise.
I think there are Shots in Prime and Shots in Prime where the defender is in your face putting you off. I think the Back 3 gave us a lot more of the latter. This is my interpretation from watching the games. Back 3 gives us one extra defender to plug gaps at the back and build up play, and also more width. We are less susceptible to being pressed as the opposition need to commit more men forward. Also, With Xhaka pulling the strings and the wide men who are more free – we created more opportunities. Basically, we… Read more »
Some great points well presented. However points are the most important statistics and the facts don’t lie. I believe that with the full backs pressing on we become easy to play against. If we lose possession a quick ball out wide draws out centre backs out of position and leave masses of space in behind. With an extra defender covering that space we feel more secure. That being the case our opponents have to look for different ways to play against us. Hence the strange stats. Again this is only my opinion and I can’t claim to be an expert.
Is it possible that something that isn’t captured in the stats affected the goalies’ ability to stop the shots? Eg, maybe things didn’t develop quite as suddenly, giving more time to get ready for the shot, etc?
I think Middy above made a good point about Santi Carzola. The back 4 worked for us well only when we have Santi and Coq paired in the mid. If it is possible, and I know time is precious, it will be really interesting to see the Carzola and Coquelin axis versus everything else. My guess is that the win rate is a league winning win rate. Given the sparsity of data available as Carzola was injured since September, 7am will have to dig through to the season before. What the data does not tell us is how much easier… Read more »
My memory of when we tactically changed was when we were beaten 3.0 at Palace. We had in excess of 70% possession but they just hoofed it to Benteke and Zaha and Townsend came in from the wings to overload our centre halves. The Fat Walrus gloated and even said on Sky how easy it is to beat Arsenal using that system. Wenger listened for a change and we played 3 centre backs from then on in.
Thanks Sam…..glad you’ve gone though.
“Liverpoop” 😀
Excellent, very interesting article. Very good analysis. Really good work. Thanks!