27 June 2014

Visualized: USA 0-1 Germany

As always, match data from Stats Zone and Squawka.

That was an emphatic 0-1 ass-kicking. And it was also the best 0-1 loss in American history.

If I were a German fan, I'd be at least a little worried by that result. Yes, they won. Yes, they did everything they needed to, ensuring qualification as the group winner. But to out-pass and out-possess your opponent by those insane margins, yet only take 13 shots, only put six shots on goal, only create eight chances, and only win 1-0 is a bit disconcerting.

That shot total is right in line with Germany's last two matches: with 13 against Portugal and 12 against Ghana. But against Portugal, Germany averaged 37 passes per shot. Against Ghana, 44 completed passes per shot. But against the USA, it was 53 passes per shot. Yesterday, Germany took one shot for every 17.7 passes in the attacking third. That's a lot of effort for little reward, no matter needing just a point to qualify as group winners.

In my two seasons of doing these match infographics, albeit almost totally for Liverpool, I've never seen such a large passing disparity, a passing total as high as Germany's, or passing accuracy as high as Gerrmany's. Germany had three players with more than 100 passes: Lahm, Kroos, and Mertesacker. Germany completed more passes than any other side has attempted in this World Cup. The only side I can remember bettering those passing totals were Spain in the two previous tournaments and Barcelona at its peak.

But almost all those passes came in midfield. Germany attempted 230 final third passes, completing 192, but few went into the US penalty area, forced to pass sideways 20 or more yards from goal and tentatively probe down the flanks.

Some credit where due. The USA defended quite well. Germany scored just once, a fortunate if incredibly well-placed rebound after the initial set play shot was blocked, after scoring four and two in the last two matches. The Americans sat deep by design, and Omar Gonzalez, starting for the first time, was especially impressive. The LA Galaxy defender made 10 clearances, both inside and outside the US penalty area, six more than the next closest player; won both of his aerial duels; and was successful with all three of his attempted tackles.

But once again, the USA attack looked insipid without Altidore. Sure, partly by design; as against Ghana, the USA dropped deep, willing to concede two-thirds of the pitch in order to protect the last third, and invited pressure. They were content to let Germany do whatever it wanted with the ball in midfield, which makes a certain amount of sense given that the US just needed the draw. But unlike against Ghana, they weren't ahead at any point yesterday, and there was also absolutely no sign of a counter-attack even after Müller gave Germany the lead. The USA took all of four shots in total, none on target. The two American shots from inside the penalty box weren't taken until the 90th minute, with both result and qualification already sealed for both sides.

It's difficult to discern just how much of Germany's inability to convert possession into attack and the USA's utter unwillingness to attack was down to the situation. Neither side needed any more than a draw, and the result in the other match – still in both the USA and Germany's favor, if very tentatively, even when Ghana drew level just three minutes after Müller scored – meant that the result in this match was meaningless in the end.

There was never going to be any 1982-style biscotto. German fans won't allow it after the embarrassment with Austria, and it's just not in the USA make-up. If they didn't play out a simple draw to eliminate Mexico in qualifying, they wouldn't do so here. Still, maybe I'm naive, but I still expected a little more going forward from the Americans, no matter how good Germany is.

The USA was pragmatic, and pragmatism worked, if only just. Just like in the previous two matches, changing the style to suit the opponents, to get just enough of out each contest. But pragmatism probably won't be enough to advance much further, with Belgium – and then possibly Argentina – to come.

23 June 2014

Visualized: USA 2-2 Portugal

As always, match data from Stats Zone and Squawka, as well as ESPN FC for the average position diagrams for both sides.

An FYI. Chances are, I won't often include substitutions in the passing networks, but it seemed important to display two of Portugal's: Eder, on in the 16th minute, and William, on at halftime. Both were more involved in the game than the players they replaced.

There were few surprises in the Portugal passing network. Ronaldo and Nani both further forward than Portugal's "striker"; a very tight midfield triangle, with Veloso's average position pushed left because of his switch to fullback at halftime; a lot of interplay between Moutinho, Pereira, and Nani down the USA left, although Nani struggled to keep the ball moving, completing few passes with Portugal's other attackers, creating only one chance and averaging just 66.7% pass accuracy. And, unsurprisingly, Ronaldo was a black hole – except, of course, for his infuriatingly perfect assist on the late late late equalizer (sigh) – taking more shots than any other player. And he clearly wasn't fit either, missing the target with six of those seven shots. The USA's numerous interceptions also helped disjoint Portugal's attack – 19 in their own half, 12 in the defensive third – demonstrating the Americans' smart defensive positioning.

The US passing network shows a reasonably compact midfield – at least between Beckerman, Bradley, and Bedoya, with Bradley very much the hub – and a clear desire to exploit Portugal's left flank. The US started its attacks on their own left, Howard mostly passing to Besler, working the ball into midfield, before trying to find Johnson, Zusi, or Jones (then Yedlin as a substitute) down the right, a tactic which demonstrably led to the second US goal. It's no coincidence that 12 of Portugal's 15 interceptions were clustered in that quadrant of the pitch.

Altogether, it was a see-saw match, one seemingly necessitated by the game state at the time. Portugal's early goal meant the USA needed to come out, while Portugal retreated into a shell similar to the USA's against Ghana, with the Americans averaging far more possession in the first half compared to the second. After the halftime changes, and the subsequent US equalizer, Portugal responded, and responded even more after the the Americans' go-ahead goal.

The "wave diagram" from Infostrada Sports excellently captures the ebb and flow:

The heat and humidity in Manaus probably had something to do with it as well, requiring each team to take the foot off the gas at times, most notably the USA just before the end of the first half, with the referee calling for a water break in the 39th minute which was very much needed.

Both sides also made crucial substitutions. Eder for Postiga was a wash – neither made much of an impact – but Portugal's halftime substitution was decisive. Veloso's move to left-back helped stem the tide on that flank, while William did well to break up the USA's decent midfield play, attempting more tackles than any other player despite only being on the pitch for 45 minutes, with only Veloso completing the same total. And Varela was the super sub with the late goal, just as he did against Denmark in Euro 2012: like yesterday's from a cross that defenders were in position to prevent.

But the USA's changes were just as important, especially DeAndre Yedlin, his pace responsible for the go-ahead goal, emulating Fabian Johnson's bursts into space in the final third, that pace much needed as the starters tired. The other two substitutes, coming on in the final few minutes, made less difference, although Wondolowski held play up well when given the chance.

Conceding with seconds to play was a kick in the shorts, but on the balance of play, the draw was probably merited by both sides. Portugal with more possession, taking more shots; the USA with better opportunities while matching Portugal in midfield and able to exploit the holes in Portugal's formation thanks to Coentrão's absence and Ronaldo's unwillingness to track back and ostensible lack of fitness. Of course, both of Portugal goals came from preventable mistakes: Cameron's wayward clearance, Bradley's inability to keep possession with the final whistle imminent. If only the USA had held on for 30 more seconds…

20 June 2014

Visualized: England 1-2 Uruguay

As always, match data from Stats Zone and Squawka.

Toothless, overrated, tactically predictable side meets counter-attacking side happy to concede possession and defend, knowing full well they have one of the most ruthless strikers in the world. And it went pretty much how you'd expect it to.

There's a new addition to the match infographics: the passing network, ripped off from a few places, but I remember seeing Will from The Tomkins Times doing it first for Liverpool on his blog. I will *probably* keep this up for Liverpool next season, but fair warning, it takes about as long as the rest of the graphic, and Liverpool usually complete about 100 more passes than England did yesterday. Although it will be easier doing it for just one team rather than both.

Anyway. Hodgson's game plan is remarkably clear, and unsurprising. Center-backs pass to midfielders or fullbacks, midfielders pass to fullbacks, fullbacks pass to wingers, attack breaks down on the flanks. Only once did it work, with England attacking at pace, in transition, for Rooney's goal. Although with better finishing (*glares at the hairy mutant potato in the #10 kit*), it could have worked better. Still, look at all those Uruguay tackles and interceptions on the flanks. And look at Alvaro Pereira's average position compared to Caceres. Caceres is as much a center-back as fullback, Pereira's more known for getting forward, but it was the latter who sat deep yesterday, fully aware of Sterling's threat and determined to nullify it. 23.7% of England's completed passes came in the attacking third, compared to 34.5% for Uruguay.

Meanwhile, there's Uruguay's passing network. No mucking about. Get the ball to Suarez, or get the ball to Lodeiro or maybe Cavani, who'll get the ball to Suarez. Counter-attacking football, to the letter. Liverpool fans have seen this movie before, the better version of this movie. Worth noting: England completed more than twice as many passes than Uruguay, but Uruguay's two strikers completed twice as many passes to each other as Sturridge and Rooney did.

In matches where Liverpool have much more possession, Suarez and Sturridge exchange around 12 passes, sometimes more, sometimes slightly less. When they have less possession, focusing on the counter-attack, it's usually right around six. Sturridge and Rooney completed three yesterday. Three! And that's with the majority of possession! There's the "but they don't play together often!" excuse, but England's club teammates didn't fare much better. Rooney and Welbeck exchanged all of one pass, Sterling and Sturridge two passes, Gerrard and Henderson four passes. It's like they didn't even know each other. Of course, I blame Hodgson.

The midfielders were almost as egregious as the strikers. Four passes between the two central midfielders, four passes from either Henderson or Gerrard to Rooney, Rooney often passing it back out wide to the wingers or fullbacks. Incidentally, England completed four of 21 open play crosses, with three of four leading to a chance (all three from the left).

Against Italy, Sterling was England's most dangerous player, with the added bonus of a much more balanced midfield. Rooney accommodated centrally nullified both of those facets, to the detriment of both midfield and attack. But he scored, took the second-most shots, and created two chances! Yeah, and he should have finished at least two more shots and created more chances. Meanwhile, both wingers did next to nothing, Sterling because of Pereira's tight-marking, Welbeck because he's Welbeck.

After a reasonably heartening performance against Italy, if in a losing effort, we got the full Hodgson against Uruguay, without even the determined deep flat back four. Lots of possession, unable to do a damned thing with it. Uruguay knew exactly how to nullify it. And had Luis Suarez.

18 June 2014

On Liverpool's 2014-15 Fixtures

See the full fixture list here. The Premier League threatens to sue me if I actually publish the whole list on my blog.

On first glance, August and February are the cruellest months.

Unlike in 2013-14, Liverpool don't have the luxury of an "easy" start: they host Southampton, a team that's given them fits in the past, swiftly followed by trips to City and Tottenham. There will be no building up confidence with narrow wins over Stoke, Villa, and Moyes. There will be baptism by fire and blood.

At least the midweek Champions League matches don't come before difficult league fixtures until November. The first three CL ties are followed by matches against West Ham (A), West Brom (H), and Hull (H), two of three at home. But then, Liverpool host Chelsea after a midweek European match, then host Stoke, and then travel to United, who won't have midweek European fixtures next season.

So yeah, November's fairly cruel as well.

Unlike last season, the festive season fixtures aren't as dismal: trips to City and Chelsea, followed by Hull at home replaced by Burnley away, Swansea at home, and Leicester at home. Thank heavens for small favors.

But then comes February: Everton and Southampton away, Tottenham and City at home. And if Liverpool progress from the Champions League group stage, they'll have the first leg of the Round of 16 either before Southampton away or Manchester City at home, with the second leg before either Swansea away or Manchester United at home. Gulp.

Like 2013-14, the run-in isn't especially difficult after traveling to Arsenal on April 4, with just one of seven matches against a Top 8 side, but once again, that match is against Chelsea in early May, this time a visit to Stamford Bridge rather than at Anfield.

Liverpool took nine points from the first three matches last season, a reasonably big factor in what was ultimately the beginning of an unlikely title challenge, as much for the confidence boost as the points on the board. If Liverpool can do similar again this season, it'll put them in even better stead to start the campaign.

04 June 2014

On Suarez and Sturridge's Shooting 2013-14

We got interrupted by LAMBERT WATCH, but I'd like to revisit this from last Friday, on Liverpool's shooting in 2013-14, with a focus on Liverpool's two primary goal scorers.

And here's a quick GIF comparing each's shot locations, just because.

As a reminder, this is how I've defined each area.

Liverpool took 651 shots in total last season. Combined, Suarez and Sturridge took 280 shots, which was 43% of Liverpool's shots in the Premier League. And they were responsible for 52 of Liverpool's 101 goals – 51.5%.

Via Ted Knutson, the "average forward/attacking midfielder" in the Big Five leagues (England, Spain, Germany, Italy, France) averages 2.35 shots per 90, 37.5% shot accuracy, 0.23 non-penalty goals per 90, and 11% goal conversion.

And then there's Suarez and Sturridge.


Suarez's shot chart makes Sturridge's pale in comparison, but there are a few caveats. Only eight players took more shots than Sturridge this season: Suarez, Giroud, Bony, Dzeko, Lambert, Jay Rodriguez, Rooney, and Lukaku, and all but Dzeko played more minutes than Sturridge. Only 10 players took more than 99 shots in 2012-13: Suarez (187, six more than 2013-14), Bale, van Persie, Ba, Cazorla, Michu, Defoe, Giroud, Tevez, and Benteke. Luis Suarez is just an inhuman freak of nature, wholly lacking in conscience, attempting 69 more shots than the next closest player (Giroud with 112) last season.

A quick bar chart of the 15 players with the most shots last season for emphasis…

Only nine players attempted half as many shots as Suarez last season. Only four players attempted more than five shots per 90 minutes: Suarez, Jovetic, Agüero, and Harry Kane. Suarez played 2963 minutes last season; the other three played 367, 1527, and 500 respectively.

And that's the overwhelming narrative in these charts. Suarez will shoot from anywhere, anytime, early and often, yet converted an incredible amount this season. His shooting accuracy was 50% or better in five of the six box locations, only below average on the left side of the box, but he also struggled to get shots inside the six-yard box. And 77 of Suarez's 181 shots (42.5%) came from outside the box, leading to seven goals, but accurate with only 28.57% of his shots.

Meanwhile, Sturridge is a much more orthodox striker, not quite judicious, but often shooting from more sensible locations. He was a conversion machine in the danger zone (scoring from 14 of his 19 on-target shots), even if surprisingly inaccurate in the center of the 18-yard box; more than decent from the wide areas of the box; and his outside the box shots were clustered within 30 yards from goal. Sturridge only attempted two shots from the deep or wide positions outside the box. Suarez attempted 21.

That's an insane level of Danger Zone accuracy from Suarez, but his overall accuracy was dragged down by the number of shots from distance. Sturridge was more consistent zone-to-zone.

The percentage of each's total shots by locations, compared to Liverpool's in total, makes this more emphatic.

Compared to Liverpool's team-wide totals, a higher percentage of Suarez's shots came from the wide areas of the box and from deep, while a higher percentage of Sturridge's shots came from the six-yard box and the area just outside the box. Six-yard box shots are almost always the best shots, and 12.1% of Sturridge's shots came from the six-yard box, compared to 3.3% for Suarez and 6.3% for Liverpool in total. Sturridge scored on eight of his 12 six-yard box shots.

It's also worth noting how each played when Liverpool were winning, tied, or losing, which Ben Pugsley helpfully looked at earlier today. I highly encourage you to read the whole piece but the short version: Suarez is a flat-track bully whose accuracy and efficiency drops if Liverpool are tied or behind, while Sturridge is puts up fairly consistent numbers no matter the game state. Which again fits into the 'Suarez will shoot from anywhere, so it's no surprise when he's even less accurate when Liverpool are more desperate; Sturridge is more deliberate, more consistent' narrative that we see in each's shot location.

Had just one player been so potent, improved so dramatically, it's probably unrepeatable coincidence. That two players did it makes it less likely to be a chance occurrence. And when you factor Sterling into the mix – nine goals on 45 shots, 48.9% overall accuracy, 60.9% Danger Zone accuracy (eight goals, 14 on-target shots from 23 total shots) – three makes a crowd. And gives me hope that Suarez and Sturridge can perform at similar, even if not quite as amazing, levels next season.

Suarez probably still needs to be pickier about where he shoots from, Sturridge needs to improve his accuracy in the center of the 18-yard box. But we just saw two outstanding seasons from two Premier league strikers, including one that's arguably in the pantheon of great seasons.

02 June 2014

On Rickie Lambert

Rickie Lambert's now officially a Liverpool player, for somewhere around £4m plus a variety of rumored add-ons, most likely appearance and Liverpool-performance based. So what are Liverpool getting?

Lambert's shot chart is slightly worrisome. 39.8% shot accuracy is above average but not outstanding; it's 0.18% higher than Liverpool's team-wide accuracy last season. For comparison, Suarez's was 44.75%, Sturridge's 42.42%. There are an awful lot of shots from distance, and while I don't have the data, it's probably safe to assume the majority were from free kicks. 15.5% of Lambert's shots last season came from deep central positions (more than 30 yards from goal), which is a higher percentage than Suarez, who we know has absolutely no conscience about where he shoots from and is Liverpool's primary free kick shot-taker. Five of Lambert's 13 goals came from dead ball situations: three penalties, two direct free kicks.

If you remove penalties from the equation, Lambert had eight goals from 45 shots in prime positions (the six-yard box and center of the 18-yard box), scoring once every 5.63 shots. Which is massively inferior to Suarez (once every 2.85 shots) and Sturridge (once every 2.93 shots). Who are, to be fair, two of the best scorers in the league.

There's also a big hole in that shot chart, taking incredibly few shots from the right side, whether inside the box or just outside it. But that also fits with being very right-footed; all of Lambert's Premier League goals in the last two seasons were either right-footed (22) or headers (6).

I was also surprised how few of his shots came from inside the six-yard box; just six, only 5.8% of his shots in total. However, also surprisingly, Suarez also took just six shots from inside the six-yard box this season, and he attempted 78 more than Lambert in total.

Regardless of these concerns, Lambert looks a more-than-competent goal scorer, especially when you consider what Liverpool are supposedly paying Southampton.

The chances created chart is much more reassuring. The only strikers to create more last season were Suarez and Rooney; the only players with more assists were Gerrard and Suarez (and unlike Gerrard, none of Lambert's assists came from set plays). And he was even more prolific in 2012-13, creating 80 chances, but only leading to five assists. Lambert averaged 2.30 KPp90 in 2012-13, which dropped to 1.73 KPp90 last season. And, interestingly, a lot of his chances both inside the box and just outside come from the right, the area where he's seemingly reticent to shoot.

The chances he creates are often excellent as well. Via @DanKennett:

Andrew Beasley of Bass Tuned to Red came to similar conclusions using his Chance Created Quality metric, highlighting how many of Lambert's chances were created from open play into the center of the 18-yard box.

Only four of Lambert's 54 created chances came from crosses, including one assist, which is a minuscule amount. And I'd quibble with calling that assist a cross; Opta classified it as one, but it was a low centering pass to a wide-open Rodriguez, with both players through on Newcastle's keeper in Southampton's 4-0 victory last March. And Liverpool is not a team that plays many crosses. The fewest in the league, in fact.

Liverpool, however, does like a through ball; six Liverpool players – Suarez, Coutinho, Sturridge, Gerrard, Henderson, and Sterling – were among the top 20 for accurate through ball last season. Rickie Lambert also likes a through ball, although he completed far fewer than the aforementioned Liverpool players.

In the above article, Bass Tuned to Red also mentions a few of the worrying factors. Ted Knutson's radar chart helps highlight them fairly obviously:

Despite playing for a team that pressed so high, so often, Lambert contributed little defensively, averaging just 0.48 tackles+interceptions per 90 in 2013-14. Similar was the case the season before, averaging 0.60 per 90. Those numbers are far below Southampton's other attackers, as well as Suarez (1.4 p90) and Borini (2.3 p90, although he spent a lot of time as a wide midfielder with Sunderland). They are, however, very similar to Sturridge's 0.56 tackles + interceptions per 90 last season.

70% pass accuracy obviously isn't great, but also isn't out of the ordinary for an out-and-out striker. Sturridge and Suarez completed 79% and 75% of theirs respectively, but it's higher than Giroud (69%), Lukaku (67%), Torres (67%), Crouch (64%), Ba (64%), Benteke (61%), and Carroll (60%), among others.

Also unlike Suarez and Sturridge (and Sterling and Coutinho…), he can't dribble himself open, attempting just 20 take-ons last season, completing just nine. Completing nine of 20 take-ons would be about normal in two games for Luis Suarez, attempting 237 and completing 93 last season.

Lambert was also hit and miss as a substitute last season, a role he'll play far more often for Liverpool. It is an incredibly small sample size; he made just six substitute appearances last season. In two of those games, the last two, he was excellent: a goal and assist against Norwich, a late winner at Swansea. But in the other four – which, to be fair, were against Chelsea (twice), City, and United – he failed to register a shot or a chance created three times.

And yes, Lambert's 32, to be 33 in February. His Liverpool contract is supposedly for two years, and I'd be surprised if Liverpool get much more than that.

But he's seemingly in fantastic shape, and durable, missing just one match in the last two seasons, a slight hamstring injury that kept him out of the 2-2 draw against Arsenal last January. Yes, he's only played 5945 of a possible 6840 minutes, substituted in 10 of his 31 starts last season, but he's probably not going to be asked to play a lot of 90 minutes at Liverpool.

Like others, I can't help but mention Gary McAllister a little more than a decade ago at Liverpool, or Teddy Sheringham for Manchester United before that, as similar signings. Liverpool under Rodgers and FSG has notably gotten younger, and rightfully so, but at the same time, experienced players can often play crucial roles both on and off the pitch. Bellamy was 32 when he returned to Liverpool, with far more injury concerns than Lambert, and no one's complaining about his contributions in 2011-12. Like Bellamy, Lambert is a boyhood Liverpool fan, but also from Liverpool, a member of Liverpool's youth set-up from ages 10 to 15.

Squad depth and experience, two things Liverpool very much needed last season.

And Rickie Lambert is very much a depth singing, a replacement for Iago Aspas – trusted to play all of 132 minutes last season – backup for Suarez and Sturridge, competition for the most-likely-returning Fabio Borini. Lambert is a different option, much more a prototypical target-man, very good at holding up play (something Sturridge is underrated at) and vastly better in the air than any of Liverpool's current strikers, but still an excellent all-around footballer. Unlike some former Liverpool strikers (*glares at £35m currently on fire in a dumpster somewhere*).

Lambert averaged slightly more than two successful aerial duels per 90 last season, winning 41% of his duels, far better than Suarez (0.58 per 90, winning 26%), Sturridge (0.28 per 90, winning 18%), or Borini (0.97 per 90, winning 33%). Neither Southampton nor Liverpool play a lot of high balls, but it's a handy option to have. As long as the player can do other things as well (*again glares at the dumpster fire*).

Like Gerrard, Lambert's excellent from the penalty spot – having never missed, in 34 attempts, since joining Southampton in 2009 – and like both Gerrard and Suarez, not too shabby from direct free kicks (you know, such as this one).

It's not surprising to see Liverpool trying to raid Southampton's pantry so thoroughly. It's a side that plays like Liverpool play, a side that gave Liverpool multiple problems in two of the last three meetings before being bumrushed by a team on fire last March.

Depth, experience, a reliable backup, a different option. An England international, a Liverpool-born player who'll run through brick walls for the club. For all of £4m. Barely more than Assaidi, less than Poulsen or Konchesky (if you include the players going the other way). Wrapped up by June 2.

It's hard to argue against that kind of business. And it assuredly won't be Liverpool's last business.