Tuesday, December 4, 2012

Pushing the Panic Button: Gameweek 15

The above visualisation is going to form the centrepiece for a new regular post (probably bi-weekly) which is kind of the opposite of the fanning the flames piece. Rather than focus on players who did well this week, we're going to look at players who didn't, and particularly those who haven't for a number of weeks and identify whether, yes you guessed it, the panic button should be pushed.

Before we get into specific players, a quick review of what the table above shows:

xG - This is the expected goals a player should have scored in a given week based on the number of shots he took (inside/outside the box), his historic skill level regarding hitting the target and the league average rate of converting shots on target (SoT) to goals. This metric gives an overview of the chances a player is receiving, along with the quality of them and the way he is converting them.

Pts - simply the points scored by the player over the last five gameweeks

SiB% - the percentage of a player's shots that have been taken inside the box, as a cumulative total over the season. A sustained decline here would indicate that a player has perhaps been deployed deeper or is taking more reckless shots and thus might see his goals dry up.

SoT% - the rate at which the player is hitting the target, as a cumulative total over the season. As with SiB%, a decline here would suggest that the player is possibly taking ill advised shots or is getting into worse positions and thus his production may suffer.

pSiB% - the player's share of his team's shots inside the box, as a cumulative total over the season. A sustained decline could indicate he is becoming less influential in his team's success (it could also just mean his team are creating more chances, but either way it would need to be investigated).

CC - simply the number of chances created per game over the last five gameweeks

pCC% - the player's share of his team's chances created. A decline here would suggest a player is becoming less influential in his team's play and thus assists may be harder to come by in the future.

Player Focus
Let's focus on a couple of players. You can play along at home by selecting the player from the menu on the right hand side of the visualisation:

Eden Hazard - in terms of goal potential Hazard hasn't really seen any decline of late, though we do need to keep an eye on that declining SiB rate, especially given the very real possibility that player roles might change under new management. The assist potential is more troubling as Hazard's actual chances created and his share of Chelsea's total have both seen a fairly sharp decline. With away games at NOR, EVE and STK to come after the off week, I wouldn't say we're at the panic stage yet, but I would be tempted to pull the trigger here.

Santi Cazorla - Cazorla was one of the few players who could offer genuinely elite returns in both goals and assists and while that's still true to an extent, the gloss has come off of late, particularly on the assist front. That said, though declining, he still accounts for over 20% of Arsenal's created chances and just enjoyed one of his best xG performances of the year. The upcoming fixtures are too good to panic here, though I wouldn't blame anyone for at least sweating this one a little. With Bale injured and the Chelsea boys struggling s bit, Cazorla looks no more risky than most of the elite options so I'd stay the course for now.

Dimitar Berbatov - four straight two-pointers have seen close to 100k managers abandon the Berb-wagon, and the data shows that their panic is not totally unfounded. In his defense, the last three games have seen Fulham face @STK, @CHE and TOT and while Berbatov's xG has clearly suffered, he's still getting his shots in good positions and has only seen a slight decline in his share of Fulham's chances. This isn't a panic situation yet, and honestly it's not one I'm even that concerned about given Fulham's upcoming run of games.

Nikica Jelavic - Fire! Panic! Jelavic's xG has declined each of the past five weeks and his share of Everton's SiB has been creeping down by the game. Throw in two appearances without a single created chance and we have a player whose once promising 8.2m price tag suddenly looks massively inflated. With 50k sales in the past few weeks the general public are clearly tired of Jelavic's decline and based on the above there isn't much evidence to suggest he will turn it around in the immediate future. With 6% ownership there's still further to fall and if you own him you'll probably want to bail now or risk losing more money.

Kevin Nolan - the former Bolton man has gone four games without any fantasy joy but the data shows he hasn't really shown any decline and actually enjoyed a couple of his better games in the last two gameweeks. Considering these four games have been against some of the league's better sides (STK, @TOT, @MUN and CHE) there's no reason to worry about Nolan.

Steven Pienaar - There's reason for genuine concern here. Pienaar has always been more of an assist threat than a goal threat but two blanks in the xG column are still concerning. He is getting more shots inside the box but is hitting the target less which suggests the quality of those shots isn't really improving. His assist numbers remain solid and his share of Everton's total has steadied, but there's a chance that Everton's offensive success might flatten after a dynamite start to the year and thus Pienaar's going to be getting a piece of a smaller pie. In a vacuum I wouldn't panic here but given his price tag, the plethora of other options available for less money and his relatively high price tag, I'm pushing the button.

So that's the basic idea of this new feature. Hopefully the data will spark some thoughts if nothing else, and I welcome any comments below, @plfantasy or on Facebook.


Ton the Don said...

I've had Jelavic for awhile now, but have been looking to cut bait for some time. What are some strikers worth 8.2 or below that I should consider a switch to? I have no extra budget at the moment.

Snorre said...

Weird that you chose to exclude Sessegnon, otherwise very cool. Sterling still looks great here.

Ton the Don:
Obvious answer is Berba, has great fixtures upcoming and Ruiz' return should benefit him.

Ton the Don said...

Unfortunately, already have Berba.

Henry Darana said...

Oh My God..
Brilliant Post, Chris..

Can you do this for the last 10 gameweeks instead of the 5.. Or atleast 7 gameweeks...

Henry Darana said...

Ton : Le Fondre and Long..

Unknown said...

First time I'm seeing this blog, just wanted to say this is really good. Thanks.

CDI said...

Thanks for this Chris. Might stick with my RVP, Suarez and Ba frontline as no one else really jumps out at me.

stooshermadness said...

Chris - this is a really interesting way to look at things and provides much food for thought. I am one of the those holding Santi Cazorla beginning to panic. I think sometimes perhaps the underlying numbers are just that - underlying numbers for an individual. I think as we look at these underlying numbers, managers will need to consider an individual player's good underlying numbers in conjunction with the consistency of his team's perfomances. Over their last 8 GWs Arsenal have scored - 0,1,1,3,5,0,1,0. Take away the 5-goal outlier against 10 man Tottenham (Arsenal were being bossed til Adebayor's moment of stupidity) and there is not a lot there to suggest - even with Cazorla's good underlying numbers - that the Arsenal attack is worth betting on. No one is consistently scoring Cazorla's CC. And this wasn't a tough stretch of games either. Given Cazorla's price, I wonder if a sideways move to Mata after his blank GW or Bale after he returns isn't a good idea. Even Gerrard now that Lucas is back. Or even a dropdown in price to a 'keeps on ticking' Kevin Nolan to use the funds elsewhere. I am hanging on to Cazorla for a week or 2, hoping the numbers pan out into FF results, but I am not particularly sanguine about my expectations.

mischty said...

your blog just gets more and more awesomer.

@shots_on_target said...


Chris Glover said...

Ton - Ba is the obvious candidate but given his injury status and tough upcoming games I wouldn't be too confident in pulling the trigger on him. The model loves Lambert but he has the week off next week, so that really leaves Berbatov, Defoe or someone on a double this week. Of that group I like Berbatov the most by some distance but to maxmise the DGW I would look strongly at Le Fondre too.

Chris Glover said...

Snorre - I'm not sure anyone was ever really in on Sessegnon to be pushing the panic button. I tried to focus on widely held players so Sess didn't make the list.

Ton - sorry just noticed you have Berb. I might be tempted by Le Fondre in the short term, then maybe move back to Ba when the fixtures ease up a bit.

Chris Glover said...

Henry - I could add more weeks though I'm worried that might be information overload. I wanted to try and capture players whose form is starting to tip and might need attention. However, I do agree that a longer sample size is needed for a fuller analysis. So, I might try and put together a similar data set with perhaps 10 weeks, or even just the season to date for each player which would be available all the time.

Chris Glover said...

CDI - it's funny, it seems like just a week or two ago there were lots of forwards I wanted and struggled to settle on three. Now I feel like the only two guys I really love are Rooney, Berbatov and possibly Suarez. After that we have RvP (price), Ba (fixtures), Lambert (off week), Le Fondre (lack of proven ability) etc etc who all bring at least one question mark to the table. Amazing how quickly things change.

Chris Glover said...

Stoosher - excellent points and I totally agree. Not all chances created are of equal value and to use the Cazorla example, if he slips in Gervinho who takes a rash shot from an acute angle while Mata slips in Torres 6 yards from goal with a nice reverse pass, the latter is always going to be much more valuable. I've struggled with how to account for this fact, and while the model does use team averages rather than league averages for converted chances, it still doesn't account for the type of pass. I am currently working on something to try and categorise players by looking at the number of crosses, through balls, touches in the area etc they get to create profiles and then look at how their chances are created. This would then also help account for wing backs like Baines and Shorey who rack up great CC numbers but struggle to get more than the odd assist with any consistency.

On watching him play one would think that Cazorla's CC would be quality as he clearly has a good eye for a pass and is playing alongside talented (though not elite) finishers. I think it's that reason I still have some faith in him to have success once again, but I too would categorise myself as nervous and have flirted with Mata these past couple of weeks.

Chris Glover said...

Thanks for the kind words everyone. My simplistic brain likes these visualisations so I'm always pleased when others get something from them too.

Colin Gahan said...

'kin good blog this. I'm a futures trader, could do with someone like you to run stuff by

DF said...

I. Love. This. Well done Chris!

That is all.