Key Data Tables 2020/21

Compiled by Tony Attwood

In 2020/1 we began to realise that something curious was happening to Arsenal.  Something none of us had ever seen before.  So as ever we went searching for the statistics.

We started with the home/away statistics which show that the percentage of points won by away teams dramatically increased when there was no crowd, despite others saying this never happened.

Season Home percentage of points Away percentage of points
2018/19 57% 43%
2019/20 58% 42%
2020/21 44% 56%

After that we really set to work and this page contains some of the main data sets that we found and which have influenced our view of what happened in 2020/21 – and why it is important to make this public.

This data is not complete yet – we have some more to add, but this might interest you even without the next set of information.

Table 1: the first 14 games of the 2020/21 season.   14 games is an interesting dividing line since it took us up to Christmas Day.  Here is how the league table looked on entering the holiday season.

Team P W D L F A GD Pts
1 Liverpool 14 9 4 1 36 19 17 31
2 Everton 14 8 2 4 25 19 6 26
3 Tottenham Hotspur 13 7 4 2 25 12 13 25
4 Leicester City 13 8 0 5 24 17 7 24
5 Southampton 14 7 3 4 25 19 6 24
6 Manchester City 13 6 5 2 19 12 7 23
7 Manchester United 12 7 2 3 22 19 3 23
8 Chelsea 13 6 4 3 26 14 12 22
9 West Ham United 13 6 3 4 21 16 5 21
10 Wolverhampton Wanderers 13 6 2 5 13 17 -4 20
11 Aston Villa 11 6 1 4 21 13 8 19
12 Newcastle United 13 5 3 5 17 22 -5 18
13 Crystal Palace 14 5 3 6 19 25 -6 18
14 Leeds United 13 5 2 6 22 24 -2 17
15 Arsenal 14 4 2 8 12 18 -6 14
16 Brighton and Hove Albion 13 2 5 6 15 21 -6 11
17 Fulham 14 2 4 8 13 23 -10 10
18 Burnley 12 2 4 6 6 18 -12 10
19 West Bromwich Albion 13 1 4 8 10 26 -16 7
20 Sheffield United 13 0 1 12 7 24 -17 1

Arsenal were four point above the relegation places with four of the five teams below them having at least one game in hand over Arsenal. If Brighton and Burnley won their games in hand, Arsenal would be down to 17th.

The media made much of that – far more in fact than the commented upon what happened in the remaining 24 games of the season.  Here is the league table of last 24 games of the season.  For ease of reading I’ve just included the top eight clubs

P Team P W D L F A GD Pts
1 Manchester City 24 20 0 4 62 20 42 60
2 Arsenal 24 14 5 5 43 21 22 47
3 Manchester United 24 13 8 3 43 21 22 47
4 West Ham United 24 13 5 6 41 28 13 44
5 Chelsea 24 12 6 6 29 22 7 42
6 Leicester City 24 11 6 7 42 33 9 39
7 Liverpool 24 11 5 8 32 23 38
8 Tottenham Hots 24 11 4 9 43 31 12 37

So that’s part one: Arsenal were poor in the first third of 2020/21 but superb in the last two thirds of the season.

Part two: finding out what was going wrong

Even the briefest of looks at the profiles of referees who oversaw Arsenal games in 2020/21 will reveal two factors. 

First is the fact that while some referees just undertook one game involving Arsenal, other refs took control of Arsenal three, four or even five times.

This of course is not permissible in some leagues, where no referee can oversee any club more than twice – it is a basic way of stopping corruption.  But this doesn’t apply in the Premier League.

Arsenal got an average of 1.24 cards per game last season.  But as we can see, four of the five referees who oversaw Arsenal games the most (all having 3+ Arsenal games in the season) were referees who awarded above average yellow cards against Arsenal.

The only saving grace was that Mike Dean who gave out three yellows in his one Arsenal game, only had one Arsenal game.  Even for the lackadaisical PGMO, this was too obvious.

Referee Games Fouls awarded against Arsenal per game Yellow cards against Arsenal per game
1.Michael Oliver 5 11.60 1.40
2.Martin Atkinson 4 7.75 1.00
3.Chris Kavanagh 3 11.33 1.67
4.Craig Pawson 3 9.33 2.33
5.Andre Marriner 3 10.67 1.67
6.Paul Tierney 2 8.00 1.00
7.Anthony Taylor 1 8.00 0.00
8.Jonathan Moss 1 11.00 2.00
9.Kevin Friend 1 6.00 2.00
10.Graham Scott 1 5.00 2.00
11.David Coote 1 9.00 0.00
12.Stuart Attwell 1 8.00 0.00
13.Lee Mason 1 3.00 0.00
14.Mike Dean 1 12.00 3.00
Average 1.24 cards per game

The conclusion was clear.  PGMO would give Arsenal games to referees who gave out an above average number of yellow cards to Arsenal players while clubs are not allowed to question referees in any way.

So what could Arsenal do?

Arsenal’s management decided to stop having fouls given against them all the time, since it harmed the flow of their game.  To do this they had to reduce the number of tackles they deployed.  The logic was fewer tackles = fewer fouls against = fewer yellow cards.

These figures are for the whole season, although the bulk of the change came in the last two thirds of the season as the players got used to the new arrangement.

  • Tackles down 22%
  • Fouls down 18%
  • Yellow cards down 45%
Team Tackles 2019/0 Tackles 2020/1 +/- Fouls 2019/0 Fouls 2020/1 +/- Yellow 2019/0 Yellow 2020/1 +/-
Arsenal 584 456 -22% 421 345 -18% 86 47 -45%
Aston Villa 523 530 +1% 451 439 -3% 68 63 -7%
Brighton & HA 641 628 -2% 385 427 +10% 57 46 -19%
Burnley 547 507 -7% 412 384 -7% 67 48 -28%
Chelsea 638 612 -4% 386 433 +12% 60 50 -17%
Crystal Palace 686 636 -7% 425 409 -4% 62 54 -13%
Everton 671 640 -5% 472 386 -18% 70 59 -16%
Leicester City 742 681 -8% 418 416 0% 41 61 +49%
Liverpool 550 526 -4% 331 396 +19% 38 40 +5%
Manchester C 514 498 -3% 361 361 0% 60 46 -23%
Manchester U 580 551 -5% 433 452 +4% 73 64 -12%
Newcastle U 603 512 -17% 370 388 +5% 66 61 -8%
Sheffield United 611 634 +4% 410 459 +12% 61 73 +20%
Southampton 706 727 +3% 450 426 -5% 53 52 0%
Tottenham Hot 665 636 -4% 423 439 +4% 82 53 -35%
West Ham U 696 514 -26% 387 374 -3% 62 48 -23%
Wolverhampton 646 588 -9% 406 425 +5% 57 53 -7%

Of course Arsenal were not alone

  • 17 out of 20 clubs cut their tackling, but Arsenal did it the most.
  • 7 out of 20 clubs cut their fouling, but Arsenal cut theirs the most
  • 13 out of the 20 clubs reduced their yellow card rate, but Arsenal cut theirs the most.

As you may know, newspapers, TV and radio stations never comment on referee performances aside from the occasional “I think he’s got that one wrong” comments.  Statistics of refereeing are absolutely not allowed in media reports.

So Arsenal did what others were doing, but they did it more and they did it better, and in the end they cut the number of free kicks and yellow cards against them.  Which led to the rise up the league in the last two thirds of the season.

As for why no one else will run these figures and simple analyses – we will come back to that in another piece..

Next we look at fouls committed and fouls received.

Team Played Total fouls by team Fouls per game by team Total fouls by opposition Fouls per game by opposition Fouls committed / Fouls endured (%)
Arsenal 38 345 9.08 421 11.08 81.95
Aston Villa 38 439 11.55 567 14.92 77.43
Brighton and Hove 38 427 11.24 376 9.89 113.56
Burnley 38 384 10.11 403 10.61 95.29
Chelsea 38 433 11.39 377 9.92 114.85
Crystal Palace 38 409 10.76 438 11.53 93.38
Everton 38 386 10.16 409 10.76 94.38
Fulham 38 484 12.74 387 10.18 125.06
Leeds United 38 427 11.24 432 11.37 98.84
Leicester City 38 416 10.95 440 11.58 94.55
Liverpool 38 396 10.42 355 9.34 111.55
Manchester City 38 361 9.5 389 10.24 92.8
Manchester United 38 452 11.89 411 10.82 109.98
Newcastle United 38 388 10.21 432 11.37 89.81
Sheffield United 38 459 12.08 304 8 150.99
Southampton 38 426 11.21 438 11.53 97.26
Tottenham Hotspur 38 439 11.55 494 13 88.87
West Bromwich A 38 404 10.63 430 11.32 93.95
West Ham United 38 374 9.84 367 9.66 101.91
Wolverhampton W 38 425 11.18 404 10.63 105.2

Finally our figures from the Football Observatory

The CIES Football Observatory rates players on a match-by-match basis by “taking into account the pitch production and efficiency at both individual and collective level” (methodology) while using data from OptaPro.

It is interesting that in the past some British journalists have themselves proclaimed that Opta is the only usable source of reliable data, but now, oddly, they have suddenly shied away from Opta.  Just at the moment it shows how incredibly well Arsenal players are doing!

Their monthly analysis shows player rankings for games played at big-5 league level over four different time periods: last week, last month, last three months and last six months, obviously only including players who have played in enough games to have a suitable level of statistics available.

For the last month their ranking includes these players – rating the best players for the last month of the season.

  • 1. Nicolas Pépé: 94.8% 
  • 2. Granit Xhaka: 93.7% 
  • 7. Martin Ödegaard: 92.3%
  • 10.  Mohamed Elneny: 91.0% 
  • 13. Joe Willock 90.0% (Newcastle United FC)
  • 13. Gabriel Martinelli  90.0% 
  • 13. Bernd Leno.  90.0% 
  • 18.  Gabriel Magalhães  89.2%
  • 31. Calum Chambers  87.6%
  • 32. Rob Holding 87.5% 
  • 32.  Emile Smith-Rowe 87.5%
  • 35. Thomas Partey  87.4%