Liverpool’s Chance Quality

There has been a lot of debate amongst the analytics community on Twitter recently, regarding how chances created vary in quality.

Whilst there is no surprise in that, when you look at a pitch map of them on Stats Zone, it’s clear that the quality varies greatly. Take a look at Liverpool’s chance creation map from the recent victory over Manchester United:

ImageAs an aside, it’s very interesting to note that the pair of set piece chances created by Steven Gerrard’s meet each other very close to the penalty spot; evidence of a deliberate tactic?

Anyway, back to the real issue here: chance quality. You can see on the above map a simple pass along the left touchline that doesn’t even reach the final third counted as a chance created.

This pass, by Jose Enrique, went to Raheem Sterling, who proceeded to dribble towards goal before having a shot. To my mind, the player creating the chance deserves little credit here, as there was so much to do before a decent shot could be taken. Just to clarify, as this was the pass to a player who had a shot, it has been correctly recorded as a chance created, as unusual as it may seem in this case.

I am therefore going to manually compile a record of the Reds’ chances throughout this season, using a ranking system that I have devised, so that we can then see who is really the best at creating quality chances for Liverpool.

Clear-cut (or ‘big’) chance data covers the very best ones already, but since EPLIndex are no longer able to provide this to subscribers, it is not available in the public domain (to the best of my knowledge). My system will enable us to look at all chances in greater detail.

I don’t want to have lots of categories, else they’ll all contain small samples, so I’m proposing six different types of chance. Assuming that Liverpool create a similar number of opportunities as they did last season (546) then I should have around ninety chances per category on average.

I have used Paul Riley’s SPAM model and Colin Trainor’s Heat Maps as a guide when deciding the different types of chance, though as I will have to compile this information myself manually, I’ve had to simplify it slightly.

As an example, whilst Colin divides the wide areas of the box in to two sections, I’m going to struggle to maintain consistency of recording on that front whilst using Stats Zone. I am therefore going to stick to using pitch markings to divide my zones.

It should also be noted here that no record is made of whether or not the chance resulted in an assist, as that is not down to the chance creator. The six categories, and the points I will be awarding, are as follows:

Central Box Open Play (CBO), 5 points – Passes in open play that are received in the central area of the penalty box, which I define as the width of the six yard box reaching out from the goal line to the edge of the box (so the same as in the SPAM model, in other words).

Central Box Set Play (CBS), 3 points – These passes are received in the same area as the above, but receive fewer points as it’s far easier to put the ball in to a specific area when striking a dead ball with all of your opponents at least ten yards away.

Wide Box (WB), 3 points – Again borrowed from SPAM, these are the areas within the penalty area that are wide of the six yard box.

Central Outside Box (COB), 2 points – This zone maintains the width of the six yard box, but covers the area from outside of the penalty box out to where the final and midfield thirds meet.

Final Third (FT), 1 point – The rest of the final third essentially, meaning outside of the penalty area, and beyond the width of the six yard box.

Outside Final Third (OFT), 0.5 points – The recipient of the pass will have to do remarkably well to score with one of these, as per the aforementioned pass by Jose Enrique against Manchester United. I was tempted to ignore these entirely, but felt I should give a minor acknowledgement to them as they will be recorded by Opta as chances created.

I’m not going to claim that this is a perfect system at all, and I gladly welcome any constructive feedback in the comments below, but hopefully it should help to move the ‘quality of chance’ debate along a little bit, certainly with regards to Liverpool at least.

I’ve already realised that crosses into the centre of the box will be ranked well even though they are usually hard to score from, but it is difficult to differentiate them from passes, and I am trying to keep this simple at present. This system will therefore make Stewart Downing look like a world beater, but that’s a price I’m willing to pay for the time being!

With the categories decided, I am now able to compile the figures for Liverpool’s first three league matches this season, to see both which Liverpool players created the best chances, and also how the Reds fared against their opponents overall on the day.

Let’s look at the players first. I should point out here that for some reason unknown to me, the figures on Stats Zone and WhoScored differ. The latter note that both of the Daniels (Agger and Sturridge) have created a chance, but the former has no record of that. As I will be using the former exclusively for this project, don’t be surprised to see more such discrepancies in future.

ImageIt’s important to remember that if a player creates a lot of chances in poor areas, then it’s almost certainly not their fault; they haven’t made their team mate who received their pass shoot at goal, after all.

But we can see that from the above table that whilst Iago Aspas has created the most chances for Liverpool this season, too many have been in poorer areas of the pitch. As I say, that’s not his fault entirely, but it does help to put his chance creation figures into better context for future reference.

Philippe Coutinho leads the chart here in terms of points per chance, though this research has revealed an interesting anomaly in terms of his creative powers this season.

Although his total of six chances is very respectable (with 2.0 per game currently joint eighteenth best in the Premier League), they were all created in the first sixty-five minutes of the match with Stoke City. In other words, the young Brazilian has now gone 205 minutes without fashioning a goal scoring opportunity. This will certainly be something to keep an eye on in the forthcoming weeks.

I have not compiled a player breakdown for Liverpool’s opponents, but I can reveal how the two teams have done in creativity terms in the three matches so far:

Image

For the teams, I have added a column for the percentage of all chances that find the centre of the box. As we can see with Stoke City, the majority of their best chances came from set pieces, and they won’t be the only team to visit Anfield this season where that is the case, so it makes sense to acknowledge it.

At this point, I obviously have no idea if any of these figures are particularly good or bad, though I endeavour to compile all of last season’s figures when I have the opportunity, in order to hopefully create some benchmarks. What I can say at this point is:

  • Liverpool created more chances against Stoke City than in the other two games combined by far, yet they were fractionally better chances on average in the Aston Villa game.
  • In their first game since Mark Hughes replaced Tony Pulis, the Potters were still very reliant upon set pieces to create their best chances, and have had the highest percentage of their chances in the centre of the box of any team here.
  • Although Manchester United had just one more chance than Liverpool at Anfield, overall their opportunities were 62% better. It should be noted though that Agger’s assist for the winner was logged as a shot on Stats Zone rather than a chance, and it would’ve been a CBO to boot. By the same token, the Dane was trying to score rather than assist Sturridge though, so perhaps Stats Zone are correct.

This is my first attempt at this sort of analysis, so as I mention above, all feedback is very gratefully received. Thanks for reading.

Recent and related posts you might like:

Liverpool 1 Manchester United 0: Stats Zone Analysis – The first in a new series, looking at the key issues from a match by utilising the Stats Zone chalkboards.

What A Difference A Year Makes – The gulf between Liverpool’s form in 2012 and 2013 is quite remarkable…

The Creative Impact of Coutinho – The Brazilian is outperforming the Premier League’s elite level creative talents right now.

The Coutinho Conundrum – There is an interesting quirk in the young man’s stats at present…

Please check out my other articles, and follow me on Twitter or via Facebook. Thanks.

About these ads

22 thoughts on “Liverpool’s Chance Quality

  1. Hi Andrew!

    This sounds like a really good idea. I have a couple of questions though and maybe suggestions:

    1) How is a clear-cut chance defined? I assume that the pass is receiver by, say, the attacker and then the ball is shot towards goal. Does it count as a clear-cut chance if the ball is blocked by a defender, caught by the keeper, over the bar or in?

    2) Why would this rating make Downing look like a world beater? You mention that crosses inside the box will award a lot of points but shouldn’t they be met by another player to be counted as a chance? As far as I am aware Downing’s crosses didn’t always find their targets.

    3) I feel as if Central Box Set Play awards too many points. 1.5 or 2 could be enough. With 3 points, corners for example, could gather quite a big amount of points but don’t really ‘tell a story’. And as Liverpool don’t put too much emphasis on setplays, I feel as if this stat somehow falsifies the ratings of a player. In my opinion, Liverpool are eager to score from open play with interchanging forwards and definitely suck at setpieces which are, also, kind of random if a ball is put high into the box. To award 3 points for such ‘randomness’ would be too much. It probably would make more sense for a team that capitalizes on setpieces.

    4) Central Outside Box: Wouldn’t it be better to increase the area? Such that this area is not delimited by the 6 yard box but by the penalty box. I feel as if a ball is received in this area, the chance might sometimes be even better if it is received not directly in the middle but more to the side of 6 yard box. But then your ‘Final Third’ area kind of becomes obsolete and nearly the same as ‘Outside Final Third’ since it is very hard to score from the flanks. Or you could introduce a mix between ‘Central Outside Box’ and ‘Final Third’ i.e. taking the space delimited by the penalty box and 6 yard box into account but that would increase your work by quite a bit.

    Other than that I think this is really good and shows that there is much more to a player, like for example Aspas, who seems to struggle but actually does quite well.

    Cheers,
    wish0r

    • In regards to your point 2: ‘as this was the pass to a player who had a shot, it has been correctly recorded as a chance created’ – therefore crosses by Downing which do not meet a player (who then has to shoot) does not count as a chance created.

    • Hi mate, thanks for reading and for the feedback. To answer your questions (where possible):
      1) – The Opta definition is “A situation where a player should reasonably be expected to score usually in a one-on-one scenario or from very close range”. It’s subjective, but they seem pretty consistent with it,
      2) – Crossing is not an efficient way of scoring. Downing could frequently fire in, say, 15 crosses a game, and on average, three would find a team mate. By their nature, these would all be in the centre of the box, so the best area, and so awarded the most points, even though they’re difficult to score. So he would look good, even though his chances would be harder to convert.
      3) I’m wondering that too – I’ve had lots of suggestions on Twitter regarding this, so will compile them all and have a think. This is very much a first draft of this system, so changes are likely.
      4) Again, another good suggestion, but as this is very time consuming to compile I’m trying to keep it simple. It might not be quite as good an analysis, but even I have my limits!
      Thanks again for the feedback, very much appreciated.

  2. Thanks for this. Looks promising. It will be interesting to see this in a few months, and notice whether we can actually see a process.

    I’m for instance, interested to compare the first matches from previous seasons to this one.

  3. Hi mate, I’m a big fan of your work!
    Have you thought about how a value of 0.5 for ‘Outside Final Third’ chances created could be unfair in cases like Coutinho for Sturridge vs Fulham or Gerrard for Suarez vs Sunderland?
    Cheers

    • Hi Ellroy, glad you like the blog, and thanks for commenting.

      I’ve just checked, and Coutinho to Sturridge at Fulham was received in the final third – remember, this system is based on where the pass is received, rather than where it was from. So the fact that Coutinho was around the half way line doesn’t matter.

  4. Interesting article as always, and some good comments already. It seems to me that there could maybe be some mileage in comparing where the chance was created to where the shot (if any) was taken. I’m thinking of Luis Enrique’s chance created for Raheem Sterling. Surely it wouldn’t have been a ‘chance’ if Raheem hadn’t done so well to get in position to shoot.

    Suppose I’m thinking out loud here, but it would surely have been a different kind of ‘chance’ had Raheem a) run on goal and dribbled round the keeper and scored into an empty net, or b) not taken the ball as far as he did in actuality and had a shot from farther out, that may or may not have gone in/elicited a save.

    Maybe what I’m suggesting is trying in some way to measure the level of contribution to the overall quality of the chance the original provider created. Hope that makes sense – I’m beginning to confuse myself!

    • Thanks for the comment, that does indeed make sense!

      As with most suggestions I’ve had, unfortunately yours falls into ‘good, but probably impossible to do’!

      I’d have to watch all of every game after the event to have a chance of what you are suggesting. I *might* do that for Liverpool, but as I want to assess the whole Premier League, that’s just too much work sadly.

      Thanks for reading.

  5. Pingback: PLCQ: The First Three Weeks | Bass Tuned To Red

  6. Pingback: Swansea 2 Liverpool 2: Stats Zone Analysis | Bass Tuned To Red

  7. Pingback: Shots On Target Against: A Closer Look | Bass Tuned To Red

  8. Pingback: PLCQ Week Five: Di Canio’s Demise | Bass Tuned To Red

  9. Pingback: Liverpool vs Fulham: Chance Quality Preview | Bass Tuned To Red

  10. Pingback: Merseyside Derby Stats Preview | Bass Tuned To Red

  11. Pingback: Spurs v Liverpool Preview | Bass Tuned To Red

  12. Pingback: 2013 Review (and a Thank You!) | Bass Tuned To Red

  13. Pingback: Liverpool’s Chance Champion | Bass Tuned To Red

  14. Pingback: Liverpool 2 Aston Villa 2: Stats Zone Analysis | Bass Tuned To Red

  15. Pingback: Henderson and Comolli | Bass Tuned To Red

  16. Pingback: Reds vs Blues: Shots On Target | Bass Tuned To Red

  17. Pingback: Liverpool vs Chelsea: Chance Quality Preview | Bass Tuned To Red

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s