Friday, February 22, 2013

Second half stats

It's been a busy couple of months for me which has unfortunately meant that posts around these parts have been minimal and basically limited to weekly stat updates. Hopefully my proverbial calender is now a bit clearer so for the first time we can get back to actually providing some analysis.

I'm going to start with a quick one - just something that came together while looking at some other data. I've calculated the expected goals (xG) for every player in the league (regressed using league conversion rates) for the first 19 weeks of the season and then the gameweeks since. I'm not generally a fan of arbitrary cut off points but the one advantage of using the 19 game mark is that, generally, this will mean that each team has played each other once. Individual circumstances may apply to specific players (such as a change in club, or change in position) but as a general point I believe the 19:19 split is useful.

With that then, here's the data plotting players' xG for GW1-19 on the x axis against their xG for GW20+ on the y axis. Those on the left of the line have 'performed' better in the second half of the year, while those on the right have seen their production decline. It should be noted here that these numbers are player time adjusted.

I guess the first thing that jumps out is that no one has really exploded since the halfway point in the year (or vice versa) which suggests that their is a strong element of predictability to these stats - a point which should give us faith for future projections. We do however have a couple of players to highlight:

Demba Ba - it's hardly mind blowing to suggest that Ba's value has fallen since his move to Chelsea, mostly due to the lack of guaranteed playing time, but we can see hear that his average production has also fallen; from an elite forward to just another warm body. I can only think that his vast ownership (22%) is due to a high number of dead teams, but if you're among that group, please act accordingly.

Marouane Fellani - I didn't expect to see the Belgian's name on this list, but looking closer at the stats, it's possible that he is one of those dangerous players who we collectively assume can be plugged into a team for the remainder of the year, when in reality better value can be found elsewhere. His data is still fine (10 SiB since GW20) but in context with other midfield options (Walcott 22, Michu 15, Puncheon 14 and Nolan 12) he doesn't quite look like the no-brainer option he once was. It would likely be foolish to part ways now, with Norwich and Reading on deck, but after that Everton face Arsenal, City, Stoke and Spurs, which together with this decline in stat production should make Fellaini owners at least browse the other available options.

Wayne Rooney - Rooney was one the reasons I ran these numbers, and having planned to bring him in this week, it's somewhat concerning to see him fall on the wrong side of the line. Saying that, I think we have a real issue of small sample size here, given that the English front man has only played 3.5 games during the "second half" of the season (scoring three goals in the process). Still, seven SiB in that period doesn't compare favorably with the likes of Suarez (22), van Persie (17) or even someone like Benteke (20), even when adjusting for playing time. As a stand alone factor this wouldn't be enough to stop me investing here yet together with his doubtful status for this week, followed by the off week in GW29, this doesn't look a great investment to make right now. Instead, he should probably be targeted as someone to look to whenever United's double gameweek falls.

Sergio Aguero - The Argentine was the other player I was targeting this week and wanted to check on, and he fares better here than his Manchester rival. It's been a stop-start season for Aguero, whose 9 goal haul feels like a disappointment, despite only racking up 16 starts to date. Mancini gave a typically - and somewhat tired - "changes will be made" speech after the poor performance against Southampton, though he followed that with a ringing endorsement of Aguero after his brace against Leeds so I don't think there's any legitimate worry that playing time is going to be a major concern here. With Aguero also set to miss out in GW29 there's an argument to be made for holding off here too, though I'd be okay plugging the Argentine in for two weeks then assessing if you can afford to bench him later.

Gareth Bale - Bale has been simply outstanding of late, leading all midfielders in shots with 36 in the second half, despite playing one or two games less than his rivals. His first half efforts were good enough to push him close to the top of the scoring charts anyway, and with his increased statistical production, there's good reason to believe he may well end up at the very top come season end. The upcoming games are good rather than great, but given his successful run of late has come during a number of away games and a game against the Champions-in-waiting, we should be fairly confident he can succeed in all but the toughest of fixtures.

Frank Lampard - The chart supports Lampard's return to fantasy relevance quite nicely, though it also serves as a reminder that while Lampard's points production has been elite, the underlying stats - while much improved - aren't sufficient to suddenly catapult him back to elite status. That said, with a steep, though manageable price tag, he remains worth a look over the aforementioned Fellaini or even Michu, given the differential potential on offer.


cmscot said...

I hold RVP, but am starting to consider options. Man U will have a lot of upcoming fixtures, and the league will soon be a formality. So I can see him being saved for cup games especially if they beat Madrid. Rooney too.

It would be interesting to know how the stats treat the end of the season - surely returns become much more unpredictable for some players for reasons as above. Is there any analysis for how returns alter at end of season for
1) teams with nothing to play for (eg Fulham)
2) teams desperate to improve league position (eg Arsenal)
3) teams saving players for other comps (eg ManU)
The simplest analysis would suggest a Buy weighting for Arsenal, a Sell weighting for ManU, and a no-change for Fulham.

What do you think?

Marke said...

Off topic Chris, but are you aware Liverpool and Swansea players are missing from the player 8GW forcasts.