Liverpool vs Manchester United: Expected Goals Preview

This is a very quick one. United have won the last four league meetings between the two teams, but did they deserve to? I’ve used my expected goals system (which is explained in full here) and Danny Page’s match simulator to have a look.

Here are the shot maps and stats from the four matches:

Continue reading

Advertisements

Liverpool vs Chelsea: League Cup Semi-Final Review

As I was lucky enough to attend the second leg of the League Cup semi-final between Liverpool and Chelsea, I thought I’d post a few thoughts on the match, the tie overall, and of course throw in a few stats.

Continue reading

Hull City 3 Liverpool 1: Stats Zone Analysis

On the face of it, the headline stats suggested that this was a fairly even contest. You wouldn’t expect to lose a match 3-1 when you’ve had 61.5% of the possession, created seven chances to your opponents’ six, had an equal number of shots on target (four), and only three fewer shots in total. Yet looking deeper into the numbers reveals where the issues were.

Continue reading

Liverpool 1 Manchester United 0: Stats Zone Analysis

As an owner of a shiny new smartphone, I’ve recently become acquainted with FourFourTwo’s excellent Stats Zone application. Typically, they’ve now made it available on the web too anyway, meaning that I didn’t need the new phone to access it after all.

Anyway, now that I can use Stats Zone, I’m going to post interesting things from there that I spot from the latest Liverpool match, to try to get to the bottom of why the result turned out as it did; in this case, spectacularly well.

Continue reading

Carroll and Crosses

In Michael Cox’s weekly chalkboard round up for The Guardian this week, he pointed out how Liverpool’s crossing had been poor in the draw with Manchester United at the weekend.

The suggestion was that without Andy Carroll on the pitch, there was often no-one for the wingers to aim for in the box.

Using the statistics for each of Liverpool’s league games this season, I thought I’d see if there was any correlation between the amount of time Carroll has played in a match, and how accurate the Reds’ crossing has been.

Broadly speaking there has been. It’s important to remember that the crossing accuracy figures are for the whole match, and not just the time that Carroll has been on the pitch though.

Similarly, the data is not available to show what percentage of the crosses the Geordie striker got himself on the end of, so it’s impossible to state definitively what Carroll’s influence has been.

But it does seem that if Carroll is not on the pitch, then Liverpool need to find a different way to feed chances to their other strikers.

Statistics sourced from EPLIndex. Please take a look at my other articles, a list of which can be found here.

Arsenal 0 Liverpool 2 – How important was the sending off?

Liverpool beat Arsenal 2-0 last weekend, and staked an early claim for a place in the top four. But until Emmanuel Frimpong was sent off in the 70th minute, the match was fairly even, with the young Gunner who was making his debut playing well (a few hot-headed moments aside).

But what impact did the red card have on the match? The statistics show that Liverpool’s passing was markedly more accurate against the ten men; nothing unusual in that, with them having more space to play in and less opposition players trying to stop them, but I took a look at the Guardian’s Chalkboards to try to quantify the difference.

Here’s a comparison of where on the pitch Liverpool attempted their passes before and after the dismissal of Frimpong:

The big difference between the two chalkboards above is that Liverpool actually attempted some passes in the Arsenal box following the sending off, something they’d not managed earlier in the match.

It’s important to remember that immediately following the sending off Kenny Dalglish made a double substitution: Luis Suárez and Raul Meireles replaced Andy Carroll and Dirk Kuyt. The impact of this can be seen on the chalkboards.

The two most attacking wide squares accounted for 7% of the teams attempted passes whilst Carroll was on the pitch. Safe to assume that quite a few of these passes may have been crosses towards the big man.

The figure for the same areas dropped to just 2% after his substitution. Liverpool also increased the percentage of their passes in the two squares directly in front of the home team’s penalty box; from 7% to 9% after the substitution.

This was demonstrated in the central play in the build up to both of the team’s goals. Frimpong’s self imposed absence lead to space in the middle of the pitch, and Dalglish exploited that superbly with a tactical and personnel switch.

I have also taken a look at where on the pitch the completed passes originated from:

As you would expect, these figures mirror the changes seen with the attempted passes – less out wide and more through the middle. Perhaps the key thing here is the 1% figure in the right half of Arsenal’s penalty box on the second chalkboard, as this is Meireles’ assist for the second goal which wrapped up the three points.

Liverpool’s passing accuracy stats are the best way to sum up the difference the sending off had on the outcome of the match:

Up to 70 minutes – 377 passes attempted, 281 completed –

74.54%

70 mins to end – 168 passes attempted, 144 completed –

85.71%

Quite a difference. In 22% of the match (the final 20 minutes), the Reds attempted 31% of their total passes, and logged 34% of their total completed passes.

Sending off + double substitution = massive difference to result. Kudos to Kenny for making Frimpong’s foolishness benefit his team.

Please take a look at my other articles, a list of which can be found here.

Pass and Move vs Pass and Hoof: Liverpool’s Passing Statistics 2010/11

This article first appeared on The Tomkins Times on 12th July 2011. Stats sourced from Anfield Index unless stated. All figures quoted are average per Premier League match.

One of the most surprising statistics from the season just ended is that Liverpool passed the ball more accurately under Roy Hodgson than they did under Kenny Dalglish.

Sure, it was a 74.78% success rate for Roy compared to 74.08% for Kenny, so not a lot in it, but Hodgson still takes the honours.

The Reds were far more successful under Dalglish (1.83 points per game) than they were under Hodgson (1.25) though, so how can we explain these passing statistics?

I will be looking at various aspects of passing to try and explain this slightly curious phenomenon.

No doubt Roy Hodgson, king of the hoof, instructed his team to make more long passes than Kenny Dalglish did, right?

Well yes, but not by much at all. Unfortunately the stats for unsuccessful long passes are not available, but I’m sure you can guess who I think would come out on ‘top’ there.

One very fascinating discovery from the passing statistics is the lop-sidedness of Liverpool’s play last season. If ever you want proof that the Reds need a decent left sided midfielder, and hopefully new signing Stewart Downing will fill that gap, then this should help:

I can only assume that a lack of quality options on the left side of the pitch lead Liverpool to switch play to the right side rather than the left around 18 times more every match. This must surely have lead to a degree of predictability that the opposition could benefit from? That said, if I had the choice of passing to Konchesky or Johnson, I know who I’d pick.

As Andy Carroll has been brought on board at great expense to be a target man, it’s logical that Liverpool will need to make the most of their crossing. How did they get on in this respect last season?

Worryingly, the Reds were more successful at crossing under Roy Hodgson, before Carroll had even come to the club. I’m sure Damien Comolli is aware of this fact, and also that Stewart Downing and Charlie Adam, both of whom have recently been acquired by Liverpool, found teammates with 24.38% and 23.47% of their crosses respectively last season, above the average for the Liverpool squad.

As I mentioned here, Pepe Reina’s form in goal improved under Dalglish, but how about his kicking accuracy?

It did improve once Roy Hodgson’s goalkeeping coach Mike Kelly had been removed from the premises, but not by a massive amount. I suspect quite a lot of these kicks would have been shorter under Dalglish’s management, and so more likely to reach their intended target by default.

In terms of the direction of the team’s passing, I was quite surprised by the following figures:

Roy Hodgson’s team played a higher percentage of their passes forward, and a lower percentage backwards than Dalglish’s. Clearly, the differences weren’t huge, but I’m sure most of you reading this would have assumed it was the other way round, much like I did.

One theory I’ve heard (and, in principle, agreed with) regarding Roy’s higher passing success rate was that his team passed it around at the back under little pressure (which improved their statistics) before hitting it long.

Whilst that may be the case, it’s surprising to see that Roy was slightly ahead in both halves of the pitch, and not just the defensive end. These figures do not include goal kicks and throw-ins, though I wouldn’t expect them to alter the percentages hugely.

To try and break it down further, I used The Guardian’s chalkboard data and divided the pitch up into six sections as per this example:

Section 1 includes Liverpool’s goal area, through to section 6 where the opposition’s goal is to be found. Here’s a breakdown of where on the pitch the team attempted passes under the two managers, and likewise for completed passes. Please bear in mind that these figures will not match the above defensive or attacking half ones, as they also include goal-kicks and throw-ins.

These statistics tie-in a bit more with what we saw from the pitch itself; Roy Hodgson’s team were making significantly more passes than Kenny Dalglish’s in the defensive half of the pitch. As a lot of these would have been attempted under relatively little pressure, Roy’s side were able to complete more passes too.

Perhaps the most revealing information of this whole study can be found by looking at the average pass completion rate for each of the six zones:

As you can see, the figures are fairly similar for every zone apart from the most attacking one (area 6), where Kenny’s team completed almost 7.5% more passes. This probably goes a long way to explaining why they scored an extra 0.74 goals per game on average.

Of course, this great a variety of statistics can be used to prove pretty much whatever you like. But in virtually all of these comparisons, the differences between the two managers are fairly small.

Except the passing success rate for the final sixth of the pitch where Kenny romps home, and for me that’s the key fact here. Roy’s team could knock around the back all they liked; Kenny’s team put it in the opposition net.

Advantage Dalglish.

Please take a look at my other articles, a list of which can be found here.

Christian Poulsen: Defending The Indefensible?

This article first appeared on The Tomkins Times on 25th January 2011.

‘Lies, damned lies and statistics’ is a phrase used to try and undermine research, by effectively saying that you can prove anything with numbers. But can you use numbers to go so far as proving Christian Poulsen has done in any way well at Liverpool?

Poulsen is a fairly rare beast: a player that no fan seems to defend or have any time for at all. In recent years the Liverpool fanbase has been hugely divided over the merits of Lucas Leiva for example (though admittedly not so much in this parish), but I haven’t as yet come across anyone strongly defending the Dane.

That’s not to say that I personally think he’s any kind of world beater, or that he should be a long term fixture in the team’s midfield, but I thought it would be interesting to examine his statistics to see how he has got on.

As a holding midfielder, the key abilities required are primarily tackling and passing, so that’s where I’ll look at how he measures up. I have the stats on how Lucas and Mascherano did in these particular aspects last season, so we can compare like-for-like.

Last season Lucas was on the field for 2,845 minutes and Mascherano 2,820, so essentially the same length of time (assuming you ignore twenty-five minutes across a whole season).

Christian Poulsen has so far appeared in ten league games, spending a total of 656 minutes on the pitch. In order to compare his statistics with his predecessor and current colleague, I have therefore divided the figures by 656 and multiplied them by 2,845 to extrapolate them in line with the amount of time Lucas spent on the pitch last season.

Passing

In 2009-10, Javier Mascherano was Liverpool’s most successful passer of the ball with 2,115 passes attempted. His accuracy rate of 83% means that 1,755 of them found a teammate.  Lucas Leiva came second with 1,832 passes tried, and an accuracy rating of 84% (or 1,539 successful passes).

The above figures show that, whilst not quite at Mascherano’s level, Poulsen’s passing figures are virtually on a par with those of Lucas. He also passed the ball more successfully than the team overall in virtually every match. The stand out figure for me is a passing success rate of 93.42% at Old Trafford. Before you say they were all five yard backwards passes, take a look at this:

A lot of them were short, defensive passes to be fair, but there were two from around half-way which met their intended target inside the United box. Is Poulsen the new Alonso? No, of course not, but there’s clearly some passing ability there.

Tackling

As with passing, Javier Mascherano was the club’s top performer in 2009-10 for tackling with 178 attempts, at a success rate of 81% (144 tackles). Lucas was next in line with 102 successful tackles from 148 tried (69%).

This is certainly a bit more of a mixed bag than the generally positive passing stats. What’s interesting to me is that Poulsen is on target (pro-rata) to attempt 38 more tackles than Mascherano did. His 60% success rate puts him significantly below the efforts of the Argentine however. That said, winning just five tackles more would have put him above Lucas’ percentage from last season.

What is of concern is the fact that he did not attempt a single tackle in ninety minutes away at Birmingham, despite Liverpool being under the cosh for large spells of that game – how is that possible for a central midfielder? To only succeed in 46.67% of tackles against West Ham (in arguably the worst performance by any visiting team in many a year) at Anfield is surely of concern too.

Conclusion

In researching this article, I have started to wonder if Poulsen is judged by association, though not just because he was bought to the club by the much maligned Roy Hodgson.

What I am actually referring to is association of result – he was widely accepted to have had a decent game against Wolves at the weekend, although this was where he registered his lowing passing percentage so far (discounting his one minute cameo at The Reebok), but the team won convincingly.

I doubt many Liverpool players will ever register a 93.42% passing success rate at Old Trafford (plus three out of four tackles won), and I can’t recall anyone saying anything positive about his performance after the team lost. To give his performance a little context, when we won 4-1 in 2009, the passing and tackling stats of Gerrard (66% and 43% respectively), Lucas (83% and 78%) and Mascherano (81% and 60%) show Poulsen in a largely favourable light.

Maybe tackling sticks in the mind more, and so is viewed more favourably by fans? His best tackling performance came against Wolves, though interestingly his next best percentage came in the home defeat to Blackpool (where he also succeeded with 80.56% of his passing), but did anyone think he played well in that match?

Of course there’s far more to whether or not he’s been a success than these statistics. For example, if he’s not getting near opposition players to even attempt a tackle, then clearly there’s an issue. Where the tackles occur also plays a part.

But let me ask you this – if a non-Liverpool fan was having at go at Lucas (as may likely occur, alas), but you could say “he’s won 60% of the tackles he’s gone in for and succeeded with a whopping 82% of his passes” then the chances are you would. So why not for Poulsen?

Lies, damned lies, or statistics?

You can follow me on Twitter here. Please take a look at my other articles, a list of which can be found here