By Gordon Haverland, with additional commentary within the article by Tony Attwood
The question has been raised: why do teams generally do better at home than when they play away? There are many possible answers, but one is that the referee treats away teams differently from home teams, perhaps being more lenient with home teams in terms of the tackles that are allowed.
For every Game Day of the 2018/19 EPL season, I have collected the number of treatments a referee has overseen, the number of minutes teams played short-handed due to a need for treatment, and the number of substitutions required.
As of Game Day 15, 17 referees have data of some kind. As such, I will consider the top 4 (approximately the square root of 17).
Game Day 15 – Rank
In these tables in columns 3, 4 and 5 a minus figure reflects a greater number for the away team, while a positive number reflects a greater number for the home teams.
|
Rank |
Referee |
Treatments |
Difference between home and away in numbers |
Difference between home and away in minutes |
Difference in number of subs home and away |
Home Subs |
Away Subs |
|---|---|---|---|---|---|---|---|
|
1 |
Andre Marriner |
28 |
-6 |
-15 |
-3 |
1 |
4 |
|
2 |
Martin Atkinson |
26 |
0 |
5 |
0 |
4 |
4 |
|
3 |
Anthony Taylor |
25 |
-3 |
0 |
0 |
2 |
2 |
|
4 |
Jonathon Moss |
22 |
-8 |
-18 |
-5 |
0 |
5 |
|
4 |
Lee Probert |
22 |
-2 |
-3 |
-4 |
0 |
4 |
C0lumn 4: Difference between home and away treatments in numbers
Of the Top-4, only Martin Atkinson is “even” in terms of the number of treatments between home and away teams. Lee Probert is close at -2 meaning there were more treatments for away players, as is Anthony Taylor at -3. I would suggest that Andre Marriner and Jonathon Moss are unacceptable at -6 and -8 respectively. Away players are clearly in danger when playing under these referees.
Column 5: Difference between home and away treatments in minutes
Part of overseeing a large difference in the number of treatments needed by home and away teams, is a likelihood that the amount of time the home or away team had to play shorthanded is likely to be quite different from zero, and again Marriner and Moss are at -15 and -18 minutes respectively meaning the away teams are being placed in a much more difficult situation playing a man short.
Column 6: Difference between home and away substitutions
If a referee is overseeing a lot of treatments, it is possible there will be a lot of substitutions. Marriner, Moss and Probert are all showing a tendency to oversee more substitutions by the away team due to the need for treatment as a result of how they manage the game.
Column 7 and 8: Home and away substitutions
I added the Home::Away substitutions to the end. By and large, Moss and Probert are only seeing the need for the Away team to substitute players that need treatments, while Atkinson is equally hard on Home and Away.
The number of treatments (home plus away) was used to rank the referees, and the top plot shows how the rankings have evolved over the season.
The next plot down (#2) is the difference in the number of treatments between Home and Away. Mike_Dean and Lee_Mason cause the most treatments to the Home team, while Jonathon_Moss and Andre_Marriner are hardest on the Away team (Game Day 15).
Plot #3 is the difference in short-handed playing time due to treatments, Roger_East, Kevin_Friend and Lee_Mason are causing the most short-handed play for Home, while Jonathon_Moss, Stuart_Attwell and Andre_Marriner are doing the job on Away.
Plot #4 is the difference in the number of substitutions. Chris_Kavanagh, Mike_Dean and Lee_Mason cause the most Home substitutions, while Jonathon_Moss, Lee_Probert, Michael_Oliver and Roger_East are causing Away to substitute.
Technical note
The plot itself was generated in R using the plotly library. This is likely the first and last time I will be using plotly, I don’t like the idea of having pay for a commercial version in order to generate a graph that I can edit to improve how it looks.
Amongst the well known 4 major anti-Arsenal match centre referees who often are appointed by the PGMOL to referee Arsenal in big games in the PL like: Mike Dean, Anthony Taylor, Martin Atkinson and Andrie Marriner but who try to conceal their being anti-Arsenal match referee in nature or on orders in the PL do operate in cladinetine mode but not so openly against Arsenal but still make sure Arsenal are negatively impacted especially in the big games in the PL, Michael Oliver is a kind of anti-Arsenal match referee referring in the PL who does his own anti-Arsenal match referee referring in the PL openly on the field of play to the seeing of football fans viewing without Hiding it when referring in Arsenal big game in the PL I’ve observed. Arsenal vs Man City in the opening season day match at the Ems in the PL this season is an example.
Amongst the well known 5 major anti-Arsenal match centre referees who often are appointed by the PGMOL to referee in Arsenal match in big games in the PL like: Mike Dean, Anthony Taylor, Martin Atkinson, Jonathan Moss and Andrie Marriner who try to conceal their being anti-Arsenal match referee in nature or on orders in the PL but operate in cladinetine mode and not so openly against Arsenal but make sure Arsenal are negatively impacted in the match especially in the big games in the PL, Michael Oliver is a kind of anti-Arsenal match referee referring in the PL who does his own anti-Arsenal match referee referring in the PL openly on the field of play against Arsenal to the seeing of the public football fans viewing in the stadium and on TV without hiding it when referring in any Arsenal big games in the PL I’ve observed. Arsenal vs Man City in the opening season day match at the Ems in the PL this season is an example.
Gord, I think there is some merit in this analysis, the problem is most of the time the treatment happens on the pitch irrespective of the gravity of the injury.
Walter, I understand the rule to be other than in the case of a head injury or the injury being the result of a foul, the player should leave the field for treatment.
Most treatments involve the players leaving the pitch for seconds and bounding back on like spring lambs.
Meanwhile we supporters are seeing the rhythm of the game, broken, time being wasted at “the most expensive seats in Europe”.
Why don’t referees apply the laws of the game?
At one time, I was finding data on all treatments, the ones where the player didn’t leave the field and the ones where the player did. Starting with the 3rd (?) game of this season, the only treatment data became the ones where the player left the field. So there is a commentary entry which says when the injury was noticed, andalmost always some kind of mention as to when the player is allowed to return to the field (or is substituted).
How I look at Referees, is I first look at how many treatments a referee has “managed”.
If two referees have managed the same number of treatments, I look to how many minutes of short-handed play was needed (home plus away). And if that it tied, how many substitutions were needed (home plus away). And if that it tied, the referees are tied.
Which is fine, I have “rules” for determining the degree of worseness of referees.
Once you sort a list of referees for worseness, you need to assign a rank to them. There are multiple ways of assigning rank.
I think this is my data from GameDay 2.
0 _ _ _ Chris Kavanagh 00:00 00:00 00:00 _000 000 000
1 _ _ _ _ Paul Tierney 01:00 02:00 00:00 _001 002 000
2 _ _ _Martin Atkinson 01:01 00:00 00:01 _000 000 -01
2 _ _ _ _Jonathon Moss 00:02 00:01 00:01 _-02 -01 -01
2 _ _ _ Stuart Attwell 01:01 01:02 00:00 _000 -01 000
2 _ _ _ _ Graham Scott 02:00 05:00 01:00 _002 005 001
3 _ _ _ Andre Marriner 01:02 00:00 00:00 _-01 000 000
3 _ _ _ Michael Oliver 01:02 05:03 01:01 _-01 002 000
4 _ _ _ _ _ _Lee Mason 04:00 03:00 01:00 _004 003 001
4 _ _ _ _ _ _Mike Dean 02:02 03:05 01:00 _000 -02 001
6 _ _ _ Anthony Taylor 00:06 00:04 00:00 _-06 -04 000
6 _ _ _ _ Kevin Friend 03:03 08:01 01:00 _000 007 001
7 _ _ _ _ Craig Pawson 04:03 09:03 01:01 _001 006 000
So, the number on the left, is the number of treatments. Where number of treatments are tied, I then look to the total number of short-handed minutes (which is enough) to sort.
There are many different rule systems for assigning rank. Below I’ve prepended 3 different rankings to the data (and deleted trailing data).
13 _13 7 0 _ _ _ Chris Kavanagh
12 _12 6 1 _ _ _ _ Paul Tierney
9.5 08 5 2 _ _ _Martin Atkinson
9.5 08 5 2 _ _ _ _Jonathon Moss
9.5 08 5 2 _ _ _ Stuart Attwell
9.5 08 5 2 _ _ _ _ Graham Scott
5.5 06 4 3 _ _ _ Andre Marriner
5.5 06 4 3 _ _ _ Michael Oliver
3.5 04 3 4 _ _ _ _ _ _Lee Mason
3.5 04 3 4 _ _ _ _ _ _Mike Dean
2.5 02 2 6 _ _ _ Anthony Taylor
2.5 02 2 6 _ _ _ _ Kevin Friend
1 _ 01 1 7 _ _ _ _ Craig Pawson
There are other ways to assign rank. I just thought it would be useful to show you that assigning rank to data is slightly ambiguous.
With all of these systems, if you have a number of data points which are all “tied”, they all get the same rank.
I suppose at some point, I will produce more “trajectories” of referee worseness and present something. I won’t use plotly, so it means learning yet another system and while doing so, I will play with different means of assigning rank.
If you have comments on the stacked graphs in one image, by all means submit them.
Concerns I have, I would like bigger markers. I suspect some colour choices confuse people with some kind of colour blindness. I would like to use colour and dash/dot/line style, so that if there is a colour blindness conflict, the line style makes it (less) unambiguous.
Great work and effort , Gord .
Respect !