Saturday, August 16, 2014

Fantasy Football - 7: How Much Does the Opposition Matter?




Lab Record Type Section

Introduction

This post is part 7 of a series of number crunching posts I intend on doing on some EPL Fantasy Football data I have. I focus on the importance of the opposition in this one.

Part 1 has the main introduction and some other relevant details and assumptions and things like that. It also has the Dream Team of the Year and Weekly Dream Teams which I shall be referring to frequently.




Relevant Summary of Earlier Parts

In part 6, I recommended people to avoid picking players from teams which had tough fixtures lined up. This sparked a FB comment thread where my friend, Shreyas Honrao, argued that -
  • If you were one of the 35% owning Ramsey at this point in the season, then you were probably safer than one of the 65% not owning Ramsey. The logic being that if you didn't pick Ramsey then for every rich haul that Ramsey bought in, 35% of the competition would be ahead of you by rich haul much points.
  • The good attackers are oblivious to the quality of the opponent and do well against any team
I, on the other hand, argued that -
  • The 35% are wrong for picking Ramsey since he has a tough set of fixtures. They should be investing in other high quality players that have easier fixtures
  • The fixtures matter to the attackers


Is it Safer to Have Ramsey?

Yes.



Is it Wrong to Have Ramsey?

We could use ANOVA or some other fancy shmancy statistical technique but the intentions is to keep it as simple as possible so here's a chart showing the distribution of points and a table showing the mean and median points for various groups of players. Players with a price of 7.5 or more are defined as high price players, below 7.5 are low price players. If the price varied across 7.5 then the player is considered high price during the weeks it was above the cutoff, and considered low price during the weeks it was below. The top 7 teams were clubbed into one group. In alphabetical order, these teams are Arsenal, Chelsea, Everton, Liverpool, Manchester City, Manchester United, and Tottenham Hotspurs. Only players that played at least 60 minutes in that game were considered.

How to read this - eg. box on the top right covers all high priced forwards playing for one of the top 7 teams,
and offers a comparison between their performance against a top 7, and a non-top 7 team



Opposition Nature Team Nature Player Status Position Mean Points Median Points
Not Top 7 Not Top 7 High Price Midfielder 4.42 3.00
Not Top 7 Not Top 7 High Price Forward 4.18 2.00
Not Top 7 Not Top 7 Low Price Midfielder 3.40 2.00
Not Top 7 Not Top 7 Low Price Forward 3.85 2.00
Not Top 7 Top 7 High Price Midfielder 5.44 3.00
Not Top 7 Top 7 High Price Forward 6.51 5.50
Not Top 7 Top 7 Low Price Midfielder 3.85 3.00
Not Top 7 Top 7 Low Price Forward 5.61 5.00
Top 7 Not Top 7 High Price Midfielder 4.20 2.50
Top 7 Not Top 7 High Price Forward 3.46 2.00
Top 7 Not Top 7 Low Price Midfielder 2.67 2.00
Top 7 Not Top 7 Low Price Forward 3.18 2.00
Top 7 Top 7 High Price Midfielder 4.59 3.00
Top 7 Top 7 High Price Forward 5.04 2.00
Top 7 Top 7 Low Price Midfielder 2.97 2.00
Top 7 Top 7 Low Price Forward 5.30 3.50

There does seem to be a it of a difference. In the high priced, top 7 team midfielders, and forwards boxes, you can see a slightly heavier bottom indicating that they score slightly lesser points against top 7 teams, as against bottom 13 teams. In the above table, you will also notice that the mean points drops from 5.44 to 4.59 for the same group of midfielders, and 6.51 to 5.04 for the forwards. The median drops far more drastically for forwards, going from 5.5 to 2. Midfielders are more stable at a median of 3 irrespective of the quality of the opponent. Note that 3 points is what a midfielder is awarded for having a clean sheet and playing 60+ minutes and 2 points is what a forward is avoided in the same scenario but I don't remember that many 0-0s played out amongst the top tier to affect the median to be the values they are.

I tried another way of classifying them, and classified them as players whose points per minute over the season were higher than 80%ile within players in the same position -

Midfielder 80-20 cutoff was 0.055 points per minute played. Forward 80-20 cutoff was 0.078 points per minute played

This now indicates that the high performing forwards from the top 7 teams aren't as drastically affected in their performance while midfielders show a similar drop as they did when split by price.

Two similar charts are also drawn below for their defensive colleagues -

The price cutoff has no been changed to 5, since defense is usually value lower that offense


Goalkeeper 80-20 cutoff was 0.048 points per minute played. Defender 80-20 cutoff was 0.045 points per minute played

The top 7 Goalkeepers seem to be far less affected by the quality of the opposition than anyone else on the pitch and if you actually look at the performance cutoff chart then you would notice, in what would seem like a paradox, goalkeepers seem to be stepping their games up a bit when they play a tougher opposition. Defenders take a bit of a dip either which way.


Conclusion

Just being a good player should not merit inclusion in the squad. A look at the fixture list is also recommended.

If there are tough fixtures, then the inclusion criteria for forwards and goalkeepers, should be based on points per minute played.




Other Posts in the Series


Part 1: General Team Composition
Part 2: Picking the Goalkeeper
Post 3: Picking the Defensive Line
Part 4: The Value of Money
Part 5: The Value of Wildcards
Part 6:  Getting Ready for 2014-15
> Part 7: How Much Does the Opposition Matter?

No comments:

Post a Comment