Who Won Round 19, 2021? Updated

An analysis of the ripple effect of the round's results, according to the internet's finest computer models. [ Details ]

How It Works: A team has a good round if it performs well, but also if other teams competing for the same ladder spots perform poorly, since this reduces competition for those spots. This is an algorithmic ranking of the change in each team's likely final ladder position due to the round's results.

Matter of Stats
Live Ladders
Glicko Ratings
ROUND INCOMPLETE: 8 games remaining
Port Adelaide
Acceptable Expected Win
defeated Collingwood by 28 points at Docklands
Top 4 55% +10
4th +0.3
15.3 wins +0.4
119% +1.1
Gained ground against:
Brisbane Lions +0.3 Sydney +0.3
vs Richmond at M.C.G.
Top 4 95% -1
2nd (no change)
16.7 wins (no change)
132% +0.2
Lost ground to:
Port Adelaide -0.3
West Coast
vs St Kilda at Perth Stadium
Top 8 39% +2
9th +0.1
10.9 wins (no change)
95% +0.1
Brisbane Lions
vs Gold Coast at Gabba
Top 4 47% -5
5th -0.1
14.6 wins (no change)
127% +0.3
Lost ground to:
Port Adelaide -0.3
vs Adelaide at Docklands
Bottom 2 44% -4
16th +0.1
6.2 wins (no change)
80% +0.2
Gained ground against:
Collingwood +0.5
vs Fremantle at Carrara
Top 4 14% -2
6th -0.1
14.0 wins (no change)
116% +0.1
Lost ground to:
Port Adelaide -0.3
Greater Western Sydney
vs Essendon at Carrara
Bottom 5 13% -3
13th +0.1
9.3 wins (no change)
93% (no change)
Gained ground against:
Collingwood +0.4
Gold Coast
vs Brisbane Lions at Gabba
Bottom 4 45% -5
14th +0.1
7.8 wins (no change)
87% (no change)
Gained ground against:
Collingwood +0.4
Western Bulldogs
vs Melbourne at M.C.G.
Top 2 68% -2
1st -0.1
16.6 wins (no change)
140% -0.1
Lost ground to:
Port Adelaide -0.4
St Kilda
vs West Coast at Perth Stadium
Top 8 20% (no change)
11th (no change)
10.3 wins (no change)
89% (no change)
vs North Melbourne at Docklands
Bottom 5 9% -1
12th (no change)
9.8 wins (no change)
96% -0.1
Gained ground against:
Collingwood +0.4
vs Greater Western Sydney at Carrara
Top 8 44% (no change)
8th (no change)
10.5 wins (no change)
104% -0.1
vs Sydney at Carrara
Top 8 23% (no change)
10th (no change)
10.3 wins (no change)
93% -0.1
North Melbourne
vs Carlton at Docklands
Spoon 76% -2
18th (no change)
4.7 wins (no change)
69% (no change)
vs Hawthorn at Docklands
Bottom 3 77% -2
17th (no change)
6.4 wins (no change)
78% -0.2
Gained ground against:
Collingwood +0.4
vs Geelong at M.C.G.
Top 8 50% (no change)
7th (no change)
10.7 wins (no change)
102% -0.2
vs Western Bulldogs at M.C.G.
Top 2 57% -2
3rd -0.1
16.4 wins (no change)
125% -0.1
Lost ground to:
Port Adelaide -0.4
Acceptable Expected Loss
lost to Port Adelaide by 28 points at Docklands
Bottom 4 70% +11
15th -0.4
6.8 wins -0.4
90% -0.9
Lost ground to:
Hawthorn -0.5 Adelaide -0.4 Gold Coast -0.4


HONOURABLE Margin better than expected
ACCEPTABLE Margin around expectation
DISHONOURABLE Margin worse than expected
EXPECTED Game won by favourite
CONTESTED Neither team heavily favoured pre-match
UPSET Game won by underdog


To be ranked highly, a team should:

  • Win — ideally in an upset. It's always better to win, but computer models factor in the chances of each win already, and their ladder projections don't change much when teams win games they're expected to. The numbers do change when teams pull off an upset, since this accrues a win that wasn't previously accounted for.

  • Have other teams fighting for the same spots on the ladder lose, ideally in upsets. This is just as important, since it doesn't actually matter how many games you win — you just need more than your competition. When other teams projected to finish around the same area of the ladder lose, it makes it easier to take their spot (or stop them from taking yours). When they lose in upsets, it's especially good, since this means they are likely to exert less competitive pressure on higher ladder spots than previously believed.

  • Be more competitive than expected. Regardless of whether you win or lose, when a team exceeds expectations, it will be considered more likely to win in the future, and its projected numbers will shift upward accordingly. This is most visible in projected percentages, which almost always rise after a competitive performance, regardless of whether the game was won or lost.

Everything is relative! If a team pulls off a terrific upset win but so do other teams competing for the same ladder positions, no-one has gained much.

In general, you always want a team to lose if it is likely to finish near your team on the ladder, and to drop form if your team is fixtured to play them in the future. You don't care whether teams win or lose if they aren't competing for the same ladder positions, although it is helpful if they gain form ahead of playing your competitors.

Gaining/Losing Ground

Numbers beside each team indicate the change in relative projected wins. For example, if Brisbane Lions show "Gained ground against: St Kilda -1.1," this means that the Lions have effectively gained 1.1 projected wins this round compared to the Saints — most likely a combination of the Lions now being projected to win more games than previously believed plus the Saints being projected to win fewer.

These lists are in approximate order of significance, which is a combined estimate of both how large the change is and how likely it is to affect this team's finishing position.

Algorithm v0.6

This analysis uses an algorithm to rate the importance of different movements, so that a team moving from 3rd to 2nd is considered more significant than a team moving from 13th to 12th. In particular, it considers:

  • Distance between teams before and after the round

  • Team rank before and after the round

  • Whether the team has moved into or out of the Top 2/4/8

  • Rounds remaining for teams to close gaps