Expected goals plus-minus

As the season ticks on, we're starting to get a bit of data to work with, albeit sometimes in small samples. This season I am not working with a relatively complex player model but am really interested in trying to "play the fixtures" as much as I can. A couple of sources of frustration, therefore, have been as below: 1. Sites which provide the difficulty of fixtures often seem a bit simplistic. For example, the Premier League site shows Wolves as an average "3-rated" opponent, both at home and away yet the reality is more complicated. While at home they have been very solid defensively, with an Expected Goals Conceded (xGC) of just 4.71 in five games, which ranks third best, yet on their travels they have surrendered 7.94 xGC, which is the third worst . Even more confusing is that most sites ( though not all ) don't distinguish between attacking and defensive fixtures, so facing Leeds at home is presented as an easy "2-rated" fixture when in reali

Selecting defenders within the same team

We talked about this earlier in the season, but with a couple more games in the bag, I've updated the viz which attempts to help you choose your defenders. The "All Players" tab should in theory help you do this coming with a blank slate, but the r aison d'etre for the viz is to help you choose between players in the same side, once you've decided that, for example, you want a City defender. The viz can be found linked from the main menu , and here and should be opened separately to best read this piece. A few notes as of GW5 below. Arsenal I'm not sure how much value there is with this defense right now, but if you are planning to play the fixtures and take a chance with one of Emery's men, this would suggest it should be Mustafi rather than the 10% owner Bellerin who you should target. Bournemouth Cook is the man here for now, though if a clear winner emerges from the developing Daniels vs Rico battle then one of those players could topple him b

Selling Salah

Given his exploits last season, and a pretty healthy start to this one, it's hard to believe the thought of selling Salah is even worth exploring. But, international weeks give us time to think through less obvious solutions so let's make a case for it. But first, a few opening notes: This is a short term measure. Despite Mane's hot start and promising form from the likes of Hazard, I still believe Salah is the best fantasy option for the season. His underlying numbers are excellent, his team looks as good as ever and his playing time is assured (other than the odd rest of course). This move make more sense if you haven't yet wildcarded, but plan to do so in the next 4-6 weeks, potentially before GW9 during the international break. If you have already played that chip then using three or four transfers in 4 gameweeks to rotate one player (along with a corresponding move) might be tricky (although if you have wildcarded the rest of your team is presumably settled?) T

Comparing like-for-like fixtures and underlying data

As noted in other recent posts, it can be hard to know what data to rely upon in the early gameweeks. We all know that shots inside the box or those on target tend to correlate well to goals, but with only a few of those events happening each week, it can be easy to overreact to unstable data . One data point I am interested in, is how well teams are doing compared to same fixtures from prior season. This data is still fraught with issues - teams might have changed personnel or have injuries or simply suffered a bad game - but we at least have a baseline to compare against which gives a little more context to what we're looking at. The below visualization shows the Goals (G), Expected Goals (xG), Shots Inside the Box (SiB), Created Chances (CC) and Successful Passes in the Final Third (P3rd) for each team's fixtures for 2018-19 (CY) compared to the same slate of games from the 2017-18 season (PY). We'll highlight a couple of the more interesting observations below the vi

Looking at early conversion rates (to justify your transfers or crush your dreams)

There seems to be an emerging narrative in the early week of the season that fantasy teams are starting to converge and a “template” has emerged for people to follow. The strong start of the elite full backs, the apparent “must own” status of pricey Salah and Aguero and a couple of budget midfield enablers (Neves and Kante) make it seem like this is the case. If we look at the prior year ownership data for GW3 though, we see a broadly similar pattern of ownership: Overall we had 14 players with 20%+ ownership where as this season we have 17, with one clear leader over 50% and a handful more over 30%. This season I think we have a couple more budget options with the likes of Wan-Bissaka, Richarlison, Neves, Kante and to an extent Zaha and Mkhitaryan, where as last year we saw just Hernandez, Hegazi and Rooney. I wonder if the the presence of these budget options here might well be what is leading folks to conclude that teams are starting to all look the same. Differentiation is a

How much extra value do premium defenders need to offer? Part II

Previously we looked at some data from last season and tried to set a rough benchmark as to how much value you need from premium defenders in order to be worth their higher price tag. If you didn't catch that piece then I'd encourage you to do so, but if not, the TL;DR is that we estimated that you should be aiming to earn somewhere around 26 points attacking points (goals, assists and bonus points) for every 1.0m extra invested over a player in the same team i.e. Alonso over Luiz or Walker over Stones. It's also worth re-empahasising that this would exclude any playing time concerns, which further complicate things somewhat. The below graphic starts to plot the defenders from each team to quickly highlight the potential areas where the more expensive players are starting to show their value. I must note here that two gameweeks is too soon to draw any strong conclusions based on the underlying stats, but then again, we need to start making transfer decisions now so like i

How much extra value do premium defenders need to offer? Part I

At this stage of the season, those of us who are statistically inclined face a bit of a dilemma. We know that too much weight should not be placed on the limited data that's available, but yet the alternative is to rely on no data and simply make your picks based on the one or two games of action you've seen, or worse, some voodoo about world class players not being able to score in August. With this in mind, this time of year offers a good opportunity to explore a couple of other broader topics before the games start coming thick and fast and we get lost in expected goal regressions. Though data in a wider sense may not yet be stable , we can start to take some reasonable confidence from starting lineups and some of the lessons we can draw from there. Specifically here, I am interested in (a) which players offer a discount among their teammates and (b) when would we be willing to pay for the more expensive option. This is typically most useful when it comes to defenders an