When you look through the RMTs, you see a lot of comments long the lines of "x and y have opposite home/away games" or "z has good fixtures to start". However, this is somewhat done on a "finger in the air basis". I thought it would be interesting to ask: exactly how big are these edges, and how consistant are they between positions.
All of the following is based on the 08/09 results, which is the only year for which I have match by match FPL point totals. If anybody has such stats for other years, especially last year, please let me know.
It is also based on only players who played more than 60 minutes in a game (i.e. qualified for bonus points/clean sheets). This removes the impact of multiple subs who scored 1 point; to be honest though, it makes little difference to the results, as such subs have a similar impact in most games, on average.
(1) Home advantage
Unsurpringly, home advantage exists and is significant. On average, players at home score 3.81 points per game, while away they score 3.21.
What's more interesting is how that changes by position:
Code:
Pos A H Tot H Ad
Goal 3.68 3.99 3.84 0.31
Def 2.91 3.67 3.29 0.76
Mid 3.21 3.74 3.48 0.52
Att 3.65 4.17 3.92 0.53
Goalkeeper are by far the smallest - presumably what is happening here is that goalkeepers are making more saves away from home, which makes up for the goals they conceed. Defenders don't have that advantage; they get penalised by the lack of clean sheet with any compensation.
Conclusion: Being able to rotate your defenders to ensure home fixtures is a huge advantage.
(2) Quality of opposition
Again, we all know that you would rather be playing Blackpool than Chelsea. But again, how big is the differential?
I have ranked the teams seperately by their defensive and attacking qualities, based purely on average FPL points conceeded (the defensive quality is based on average points scored by the opposition midfielders and attackers, while the attacking quality is based on points scored by defenders and goalkeepers). Clearly this ignores goals scored by defenders etc., but it will do as a broad brush and is probably a better measure than purely goals scored/goals conceeded (since it is important whether a team tends to score every game, or whether they tend to score a bunch is one and none at all in others).
Code:
Rank Goal Def Mid Att Tot
1 3.16 1.86 2.43 3.23 2.40
2 3.26 2.13 2.62 2.80 2.51
3 3.26 2.42 2.67 3.23 2.71
4 3.49 2.45 2.90 3.13 2.82
5 3.34 2.70 3.45 3.17 3.10
6 3.61 2.66 3.20 3.73 3.12
7 3.41 2.94 3.35 3.52 3.22
8 3.76 3.24 3.77 3.41 3.50
9 3.21 3.40 3.57 3.56 3.47
10 4.18 3.42 3.58 3.79 3.61
11 3.45 3.52 3.49 4.58 3.66
12 3.81 3.67 3.84 3.82 3.77
13 4.08 3.57 3.82 4.60 3.87
14 4.18 3.66 3.85 4.31 3.88
15 4.18 3.75 3.87 4.28 3.92
16 4.53 3.82 4.22 3.60 3.99
17 4.08 4.11 3.53 5.03 4.04
18 4.53 4.02 3.70 4.42 4.02
19 4.74 4.23 3.74 4.83 4.21
20 4.45 4.27 3.92 5.13 4.30
Total 3.84 3.29 3.48 3.92 3.51
Again, what is interesting here is differential. Note that these stats are based on a smallish sample of events, and hence are clearly not exact. To get a better estimate, we can use regression to work out how many points, on average, each rank is worth ('PPR', points per rank):
Code:
Goal Def Mid Att
PPR 0.077 0.118 0.068 0.102
In other words, very roughly speaking, you would expect a goalkeeper to score 0.385 points more against a 10th ranked attack than against a 5th ranked attack (on average).
Again, goalkeeper selection is far less important than defender selection. Midfielder selection is again the least important. Striker selection is, perhaps surprisingly given the home/away result, almost as important as defender selection. I'd like more data to be able to confirm this.
Conclusion: Again, being able to rotate defenders is very important.
(3) Home advantage vs Opposition Strength
Bringing those two points together, we can look at relative importance of reach factor. To do this, we calculate how big the opposition ranking difference needs to be to offset home advantage:
Code:
Goal Def Mid Att
H Ad 0.31 0.76 0.52 0.53
Ranks 4.0 6.5 7.7 5.2
In other words, we are saying that a goalkeeper being at home against a 10th ranked team is equivalent to being away to a 14th ranked teams.
Conclusion: Though both home advantage and opposition stength are important, you should be giving home advantage a stronger weighting for defenders and midfields, and opposition strength a stronger weighting for goalkeepers and strikers.
Overall conclusions:- Goalkeepers are by far the most consistant players in terms of how much their average points vary, closely followed by midfielders - this suggests that it is more useful to have backups in defence and up front, because your GK/mids lose less than your def/str when they are playing Chelsea away;
- Being able to rotate defenders to take advantage of both home fixtures and weak opponents is more useful than being able to rotate your midfielders;
- Being able to rotate attackers by opponent is also useful; being able to rotate by home fixtures is less important.
