Posts

Showing posts from December, 2012

Gameweek 21 Preview Data

Image
Clean Sheet Data Powered by Tableau Attacking Data Powered by Tableau Individual Data Powered by Tableau Note that the individual rankings are based on the assumption that all players will play 90 minutes. Injured/suspended players are generally excluded but borderline cases are left in and I make no assessment as to whether individual's will be rested/rotated.

Lineup Lessons: Gameweek 20

Goalscorers are highlighted in bold , players with an assist are underlined Arsenal Szczesny, Sagna, Koscielny, Vermaelen, Gibbs , Arteta, Wilshere , Cazorla , Oxlade-Chamberlain , Podolski , Walcott Subs: Mannone, Rosicky, Giroud , Ramsey, Djourou, Coquelin, Gervinho Pretty standard lineup here from Arsene Wenger, with Walcott once again given the chance to shine up top. While I was somewhat hesitant to crown Walcott as the next Thierry Henry after his garbage time goal at Reading, he's racked up 10 shots over his last two games and looked very comfortable this week, taking up advanced positions rather than drifting out wide and only looking to feed off balls put behind the defense, which can sometimes make Arsenal a bit blunt going forward. At 8.6m he isn't a steal and I still believe that Giroud will get some  minutes up top, but Walcott is certainly a very viable option and I'd be happy to grab him from GW23 onwards after which Arsenal enjoy a very favourable fi

Gameweek 20 Preview Data

Image
Attacking data Powered by Tableau Clean sheet data Powered by Tableau Individual data Powered by Tableau

Lineup Lessons: Gameweek 19

Given the fact the games are coming so quickly at the moment, I'm going to keep the below very brief. Normal, more in depth analysis will resume in GW21, just in time for the opening of the all important transfer window. Aston Villa Guzan, Lowton, Herd, Clark, Baker, Bennett, Westwood, El Ahmadi, Delph, Holman, Benteke Subs: Given, Ireland, Albrighton, Bowery, Bannan, Lichaj, Carruthers Nothing to really note here, and to be honest, any lineup changes other than Benteke suddenly not playing would be pretty irrelevant given how this team is playing right now. Chelsea Cech, Azpilicueta, Ivanovic, Cahill, Cole, Mikel, David Luiz, Moses, Oscar, Mata, Torres Subs: Turnbull, Lampard, Hazard, Ferreira, Marin, Piazon, Ake. Luiz's move to midfield is interesting and as noted last week, it probably makes him the best pickup in this back line. Terry is supposed to be back in the next couple of games which I would suggest will shift Ivanovic to right back and thus I don't

Goalkeeper Rotation Data

Image
Disappointed with another pair of socks? Grandma got you another gift card for the Body Shop? Well, unwrap the below - a fully updated guide to the best goalkeeper/defender combos, starting from GW21 - when the transfer window opens. Given that the same rotation logic can be applied to defenders as 'keepers I have included teams who offer budget options anywhere in the back five (e.g. Dawson at Spurs) even if their 'keeper would be excluded from a true budget rotation strategy. Powered by Tableau For reference, using this same logic, playing Joe Hart every week is going to net you the lowest average goals per game forecast, with a number of budget sidekicks giving you a 0.85 average. In my opinion I find little data to backup paying the 2.0m premium for those slightly improved forecast numbers though and thus I'm happy to stick with a rotation strategy for the second half of the year. For reference, Arsenal just miss out on the sub-5.0m bracket, with their che

Christmas Bonus (Points)

Image
Powered by Tableau

Gameweek Preview Data: Gameweek 19

Image
Clean Sheet Rankings Powered by Tableau Attacking Rankings Powered by Tableau Individual Rankings Powered by Tableau

Dousing the fire, fanning the flames: Gameweek 18

Image
In the interest of time given the busy festive schedule, I'm posting the fanning the flames  data now and will hopefully get a chance to flesh it our with some narrative in the next day or so. I thought it was more a priority to get GW19 rankings up though, which will be posted in the next few hours. Powered by Tableau

Lineup Lessons: Gameweek 18

Goalscorers are highlighted in bold , players with an assist are underlined Arsenal Szczesny, Sagna, Mertesacker, Vermaelen, Gibbs, Arteta , Wilshere, Oxlade-Chamberlain, Cazorla, Podolski, Walcott Subs: Mannone, Jenkinson, Koscielny, Coquelin, Ramsey, Arshavin, Gervinho An unchanged side here for Arsene Wenger with Walcott getting another start up front in place of Giroud (who didn't even feature on the bench due to illness). For what it's worth I'd suggest that Oxlade-Chamberlain has done just okay since coming into the side and I'd be surprised if Walcott wasn't restored back to a wide position soon with Giroud retaking his spot up top. I'd suggest that the "Arsenal overplay it" narrative is sometimes an oversimplified crutch used to explain their lack of success which is often simply attributable to not having as good players as City or United rather than some philosophical shortcoming. Nevertheless, there is an element of truth in that crit

Model Review: A Comparison

Image
After posting a review of the latest model , reader Agnar had the excellent suggestion of not just comparing the model to actual results, but benchmarking it against another simpler forecast system, likely employed by large sections of fantasy managers. As suggested, a simple and useful method here is to simply take the points per game (actually points per 90 minutes to avoid the odd sub appearance skewing things) accumulated in gameweeks 1-7 and then comparing that rate to the actual rate delivered in gameweeks 8-17. We can then plot that side by side with the model analysis and see (a) which is better, and (b) if there are any specific areas in which the model succeeds/fails. The first graph shows the P90 for gameweeks 1-7 plotted against the P90 for gameweeks 8-17 (filtered to only include players who racked up 600+ minutes). The second graph shows the forecast points for gameweeks 8-17 from the model plotted against the actual P90 for gameweeks 8-17 (again, limited to players

Gameweek 18 Preview Data

Image
Clean Sheet Rankings Powered by Tableau Attacking Rankings Powered by Tableau Individual Rankings Powered by Tableau

Model review: Gameweeks 8-17

Image
As promised, it's time to take a step back from forecasting and projecting and look at how the model has performed since I rolled it out in Gameweek 8. There are three ways to look at this: Look at each player's weekly forecast score versus their actual weekly score Look at each player's aggregated forecast score versus their actual aggregated score Look at each player's average forecast score (per 90 minutes) versus their average actual score Each measure has it's advantages and disadvantages but I will throw the first option out as any model is simply not going to be accurate enough on a weekly basis enjoy great success. Of course, we can use its outputs to forecast the probability of different players' chance to succeed but even if it was perfect we'd still see massive fluctuations. So that leaves options two and three which each have some advantages: option two is the truest comparator of how the model performed over the period while option thre

Regression and (arguably) the gambler's fallacy

Image
Reader Tony asked a good question this week, and after writing a lengthy response in the comment section, I decided it was worth posting here for others to see. Tony's question: I know we always say that things even out over the season but do you think it takes that long? I wonder if you measured a player's expected points over a 1 or 2 month period would it match up to his actual points? We could therefore start to predict when a player is 'due a big win'. Being 'due' is a tricky proposition. Although some fans don't like to reduce their beloved heroes to a statistical probability, that is all we're really doing on this site and once you make that leap, the idea of being 'due' kinda falls into the gambler's fallacy (the notion that after rolling 'black' five straight times, the roulette wheel is bound to come up red). I say kinda because it depends whether you believe that (a) each player has little or no control over how his sho

Dousing the fire, fanning the flames: Gameweek 17

Image
We've added new wrinkle this week, first with the data table below, and second with the players we're going to highlight. First, let's look at the new table. It shows the key data points for every player who scored six or more points for the week, along with their actual (P) and expected points (xP) and then the difference between the two. xP is calculated based on the actual  underlying stats for the week and does not  relate to any forecasts associated with the model. What we're trying to do is identify legitimate big performances rather than those where a player, for example, notches one shot, one on target and one goal yet makes out with a nine point haul. Those points still count of course but we'd be more suspicious about their sustainability and thus they'd need more discussion (and in most causes some dousing of the fire). Where a player's points are backed up by his xP I will be more inclined to gloss over them, particularly where said player is an