Category Archives: Arsenal

Tony Pulis pontificates on the meaning of referees in the Premier League

Instant Replay Officiating: Be Careful What You Wish For

By Tim Todd: Senior Diaby Researcher

With a twinkle in their eyes, as if they have been struck with genius, nearly every single television commentator, writer, blogger, and fan of the game has decided two things: referees are suddenly the worst they have ever been and that the only way to solve the problem with referees getting calls wrong is… instant replay refereeing. Well, first, I haven’t seen a significant change in refereeing since I’ve been following the game. They have always been awful and random. And second, be careful what you wish for.

I am not a football traditionalist. I don’t think instant replay will “slow the game down”, I don’t worry about how they will stop the game to review the play, and I’m not one of these boring people who worries about how they will implement the system. I’m worried about how instant replay will change the game.

Instant replay will not slow the game down. A typical football match already has a variety of events which result in stoppages: fouls, head injuries, fake injuries, offsides, throw-ins, Chelsea players surrounding the referee every five minutes, and playing a match against Stoke City or West Ham. Instant replay will, at worst, result in a few seconds of extra stoppage time being added on to the game.

I also don’t worry about how they will stop and start the game in order to review a play. The game already has precedent for arbitrary stoppages whenever there is an injury. The referee sees a player down, he stops play (sometimes), the physios jog out, the player is sprayed with Magic Spray, the player is revived, and another sports miracle is witnessed. Play is restarted when sometimes a drop kick is awarded and other times the ball is simply kicked to the other team. This happens in every match, multiple times. And you can’t see how instant replay will be administered?

Or how about the king of arbitrary stoppages, intentional fouling. Any time the opposition is on a fast break, the opponent will look to foul to stop the play. Sometimes the official will call the play dead and force a restart with a free kick and other times, he will award an “advantage” and let play go on. This advantage rule is one of the worst rules in the game. I’ve seen Santi Cazorla get kicked in the Arsenal defensive third, the official play an “advantage”, and then Cazorla gets kicked again and the official decides it’s not a foul, so Cazorla turns the ball over in a dangerous area and the opposition scores. It’s completely arbitrary when the referees stop play.

And I don’t care how they implement the system. Some have suggested managers get challenges, I’m ok with that. So what if Sam Allardyce throws a challenge late in a game in order to waste time or stop a counter attack? Is that any different than his players intentionally fouling in order to stop a play? And before you say “yeah, but they get a yellow card for that” I can show you dozens of fouls every weekend which are done to kill a fast break and aren’t given a yellow card.

The facts are clear that the referees are well out of their depth and have been for some time. This weekend alone there was another case of mistaken identity and the wrong player was assigned a red card. That seems like a mistake that a referee should never make, yet that was the second time in two years.

For several months now I have championed having two referees on the pitch: one in front of the play, one behind the play. I think this will work wonders because right now, the referee only sees the game from the one angle and many times instant replay’s main power is that it sees the incident from the front. Had there been two referees on the pitch this weekend in the Man City v. West Brom match, I can guarantee you that the referees wouldn’t have given the wrong player a red card.

But I don’t think the two official system is going to happen because the game is primed for instant replay. The fans want it and the managers want it. Instead of a second referee on the pitch, they want a second referee in a room somewhere to watch the match and review the plays. Hell, that video replay official doesn’t even have to be at the stadium. He could be in a sound proof room in Guam. With all the people banging the drum instant replay officiating is inevitable.

I have watched the evolution of instant replay refereeing from the inside and I have to tell you, it’s weird. It’s so weird over here in America that when you watch an NFL game and there is an instant replay, the broadcasters call in their “rules analyst” to comment on the instant replay. The reason they do this is because the rules have become so byzantine and bogged down in minutiae that we need to have them explained to us by a lawyer. What was it that Shakespeare said about lawyers?

This is what happens when you introduce instant replay. By slowing the action down you get to see finer details and once you see those finer details, you have to adjust the rules to cover those finer details. For example, one of the main rules people would like to see reviewed is the offside call. And if you do that, you will need to add to the rules an interpretation of when the ball is played forward. Is it a forward pass when the ball is first touched or when the ball clears the player’s toes? Or at the middle-point of the kick? This matters because there are fractional offsides calls already happening and once we slow down the action we now get to see finer details about where the attacker was relative to the exact moment when the ball was kicked.

Notice I’m not saying this is a bad thing, I’m just saying that as a sport moves toward technology assisted refereeing, the technology itself starts to drive the officiating.

The other problem with video referees is that I don’t think two people can watch the same play and make the same call. Take the gif below:

That's a stonewall dive and the minimal contact happens outside the box anyway

Is that a foul or a dive? Koscielny certainly touches N’Zogbia but N’Zogbia certainly dives. Is “any contact” now a foul? I don’t think so, because if that was true then there would be a penalty awarded on every single corner kick. But even beyond my example of corners (since we clearly have two sets of rules: one for the penalty area and one for the field) there is a push among the fans that any contact absolves the player of guilt for a dive. And I’ve had this argument over that very gif above.

There are people who say that is not a dive, that it’s a penalty. For those folks, it doesn’t matter that N’Zogbia goes to plant his right foot and then picks it up and does a perfectly executed judo roll. Koscielny touched him, they say, it’s a foul. And for me? That is a stonewall dive.

What video replay will have to do is clarify this call above. Is any contact now a foul? How much force is required to be a foul? Can we have both foul and simulation? Should N’Zogbia get a card for diving, Koscielny get called for the foul, and the free kick be awarded outside the penalty box? Did anyone else notice the mythical Diaby in the gif? He’s there, folks. He’s there.

Maybe that’s the one question instant replay officiating can answer definitively: does Diaby exist? Until then, the mysteries of the universe, the offside rule, and whether Koscielny fouled N’Zogbia, will remain unanswered. Instant replay isn’t the panacea that many think it will be. It will simply add another layer of interpretation, it’s own layer which is bogged down in tiny details, to an already jumbled mess of rules that govern football.

If you want to fix refereeing the first thing you need to do is fix the laws of the game. Because it seems to me like the real problem isn’t that the refs aren’t seeing things properly it’s that too many of the Laws of the Game are intentionally vague and require judgement of “intent”. It’s also a problem that too many people don’t understand the laws and how officials make decisions because much of what officials decide is clouded in mystery. As if some secret cabal of black cloaked wizards are sitting in their ivory towers sending out proclamations onto the people. Until you fix those problems of interpretation and secrecy  video replay officiating won’t solve anything. In fact, it could make things worse. Much worse.


Arsene knows

Arsenal win sixth straight: people complain about Chambo and Rambo

There was much discussion in my By the Numbers column about Chambers’ stats for the match yesterday. The fact is that Chambers was targeted by Newcastle. They dribbled down his side of the pitch 14 times (12 times down the other side) and Chambers ended the game making 9/13 tackles. Chambers was also 2nd on the Arsenal team in final third passes, 3rd on Arsenal overall in passes completed, and was even 3/4 in dribbles. Writing this brought out the usual “stats don’t tell the whole story”, “lies damn lies and stats”, and my new favorite “I trust my own eyes over stats”¹.

All these stats prove is that Chambers was highly active in this match both on offense and defense. They also prove that he won a high percentage of his duels. Winning a high percentage of your duels, especially when you attempt a large number of duels, shows a player (in my opinion) who is talented and unafraid to take on his opponent.

What has everyone so upset is that he failed 4 tackles. I read the comment “how many times did the opposition blow by him??” or a variant more times than I care to count. The truth is that he was “blown by” 4 times and 3 of those were either in or around the 18 yard box. This is what people mean when they say “I trust my own eyes over stats” because where the missed tackles occur counts as much as how many. People watching the game panicked when he missed that tackle in the box, but thankfully, Newcastle didn’t score off that play. That panic stuck with them more than the 9 times that Chambers tackled well.

In fact, Newcastle actually scored off a missed tackle down the left, the side that Nacho Monreal was on. But it was actually a missed tackle by Aaron Ramsey which set up the chance for their goal. A goal that they scored on the other side of the pitch, mind you.

Chambers could miss one tackle of 100 and if his man went on to score a goal, people would say that the rest of the stats don’t matter. That is just another way of saying “the only stat that matters is the final score.” This is essentially what people are saying when they discount the 9 tackles Chambers made: none of them mattered really, it’s the one that he missed that really matters, because that’s the one that I remember seeing with my own eyes. But if that’s your argument, shouldn’t you just be happy that Arsenal won 2-1?

For me, the stats show that Chambers was targeted and he stood up to the challenge. Had he simply let guys go by him 14 times I suspect Arsenal would have lost that match. He did miss a few tackles (as you would) and whenever a fullback misses a tackle it’s panic stations (as it is because he’s last man!). But critically, he also made 9 tackles and the result is that Newcastle didn’t score on his side of the pitch. They scored on the other side.

I’ve been doing this stats thing for a long time and it’s always funny to me how people hate to have their biases challenged. Right now, the dominant narrative is that Chambers is slow or not a good right back — that Bellerin is the preferred choice. But Wenger selected Chambers intentionally for this match, probably because of his size and commitment. It was the right choice, even if he did miss a tackle.

In the end, Arsenal won their 6th straight match. It’s a tremendous run of form for the Gunners and so far we are closing the season on a real high note. It’s a huge turnaround from the start of the season where Arsenal looked, frankly, lost. In case you forgot, Arsenal started the season winning just 3 times in their first 10 matches — the worst start in Wenger’s career. This prompted some of us, myself included, to wonder if Wenger had run out of ideas.

What it looks like is that instead of being out of ideas, Wenger was trying to mold the team into something new and it was taking a long time to gel. I still don’t think this Arsenal side have a solid identity as a ball control team or as a counter attacking or counter pressing team — they look to be trying on many different styles at the moment. But they are playing some top football and are really clicking at the moment and it’s a lot of fun to watch.

Credit for that must go to the players and to the manager, one Arsène Wenger.


¹Little known fact: stats are collected with eyes and I trust a third party’s eyes, using widely agreed upon definitions, to collect those stats over your stupid eyes any day.


Just one stat proves Arsenal are better with Coquelin

I went through every Champions League and Premier League match this season looking for a reason why Ospina is preferred to Szczesny and I honestly can’t find one. But what I did find was a stat which I think proves beyond a doubt the impact that Coquelin is having on Arsenal’s defense.

First, Szczesny’s been in and out of Arsenal’s starting lineup 3 times this season. Once for a red card (Anderlecht), once for a hip injury (Dortmund), and once for smoking in the boys room¹. This gave me a chance to compare numbers for three different keepers this season; Szczesny, Martinez, and Ospina.

I looked at their stats and the only number that really popped out at me was the number of “big chances” that each keeper faced. Martinez is a bit unlucky in this regard, his sample size is small so it is artificially inflated by the terrible Stoke City match, where Stoke ran rampant over Arsenal’s midfield and created 4 big chances. But the telling stat was that Szczesny faced 1.67 big chances per game and Ospina faced a paltry 0.55.

Just in case you don’t remember, “big chances”² are the Opta stat which counts the number of clear-cut shots that a player has on a defense. These are those one-v-one moments and close range shots where we would normally expect to see the opposition score. Arsenal had 96 of these chances (in both Premier League and Champions League) and scored 45 (47% conversion) and the opposition had 49 and scored 21 (43% conversion).

In real numbers, Szczesny has played 21 matches in both Leagues and faced 35 big chances. He’s only saved 11, he’s allowed 14, and 10 were off target. That means he has a 44% save rate (shots on target that he saves) against big chances. Ospina has played 11 matches (both Leagues) but critically, he’s only faced 6 big chances, saved 2, and allowed 3 (a 40% save rate on a terribly small sample).

Just to put that into perspective, Ospina has faced 0.55 big chances per game. Szczesny was looking at 1.67 per game. Over 3 times as many!

If we look at this work rate from a saves perspective, Szczesny was saving Arsenal a big chance goal every other game and Ospina is only saving Arsenal a big chance goal every 5.5 games. One could actually argue that given the normal conversion rate of these shots and that Szczesny made 11 saves, he saved Arsenal more points than Ospina did. Ergo, one could say, Szczesny is the better keeper.

I’m not going that far. I think Ospina hasn’t been pressured enough for us to know if he is at the same level or better than Szczesny. Facing 6 big chances is, let’s just be honest, not a lot of work to judge someone by.

However, if we are interested more in the process than in the outcome³ we have to ask “why did Ospina face so many fewer big chances than Szczesny?” The answer is Coquelin.

Arsenal have used 4 different defensive midfielders this season: Arteta, Flamini, Chambers, and Coquelin. Coquelin has played 13 games this season and in those 13 games Arsenal have only faced 11 “big chances.” In the other remaining 24 games, Arsenal faced 38 big chances which is an average of 1.58 per game. In other words, just having Coquelin on the field has meant that Arsenal faced almost half the number of big chances than they did without him.

But wait… like a fine French , it gets better.

Coquelin started two matches with another defensive midfielder next to him: Flamini against West Ham and Chambers against Southampton. In those two matches alone, Arsenal faced 5 big chances. I suspect that this was down to the confusion of place which many Arsenal players feel. What is Coquelin’s job when he’s playing next to Chambers? Or Flamini?

If we remove those two games and only look at the 11 matches in which Coquelin has been Arsenal’s only outright defensive midfielder, Arsenal have faced just 6 big chances. 0.55 per game. The same number that Ospina has faced.

It looks to me like Coquelin is making it easy for Ospina and the rest of the defense.


¹This should be Szczesny’s song.
²”A situation where a player should reasonably be expected to score usually in a one-on-one scenario or from very close range. – See more at:”. This stat gets criticized for being too much of a judgement call. But for me, all stats are judgement calls. They are all created by human beings watching matches and applying agreed upon definitions. What is an assist? What’s an error? Even a completed pass is a judgement call. I am fine with their application of their definition for the purposes of this blog post.
³My new thing? I guess.