2017 : Simulating the Final Ladder and Final Series After Round 22

In this week's 50,000 simulations of the remaining round of the 2017 AFL season, four teams saw material changes in their Top 8 chances, and four more saw their Top 4 chances change by more than 5% points.

(For details on the methodology I've used in the simulations, please see this blog.)

RESULTS

The charts below summarise the distribution of ladder finishes for the 18 teams from the simulations run after the current round, Round 22, and Round 21. Note that the ordering of the teams is slightly different in the two charts, reflecting improvement or deterioration in some team's expected ladder finish.

Most striking this week is the reduction in the number of feasible finishes for most teams, and the concentration of probability in one or two ladder finishes.

AFTER ROUND 22

AFTER ROUND 21

In terms of Final 8 chances, the big winners this week were:

  • Essendon: +40% points
  • Melbourne: +10% points (which makes it +42% points in the past two weeks)

The big losers were:

  • Western Bulldogs: -33% points (which makes it -59% points in the past two weeks)
  • West Coast: -15% points

In terms of Top 4 chances, we had as winners:

  • Geelong: +12% points (which makes it +26% points in the past two weeks)
  • GWS: +7% points (up by 42% points now in the last four weeks)

... and, as losers:

  • Sydney: -8% points
  • Port Adelaide: -7% points

In the heat map below we see, as noted earlier, yet more narrowing in the range of ladder finishes for every team this week. There still remains, however, non-trivial uncertainty around the positions from 7th to 11th, and some uncertainty to be resolved both further up and lower down the ladder.

AFTER ROUND 22

AFTER ROUND 21

That uncertainty can, yet again, be quantified by calculating Gini coefficients for teams and ladder positions (see this blog for some detail on what Gini coefficients are and how to interpret them).

We see that only two teams became less certain about their final ladder positions as a result of the weekend games (Adelaide and GWS), neither of them because their hopes of playing Finals altered materially, but both of them because their chances of finishes somewhere in the Top 4 changed significantly.

Most uncertain about their final ladder positions now are West Coast and GWS, with West Coast capable of finishing anywhere from 7th to 11th, and GWS anywhere from 1st to 4th.

The teams enjoying the largest increase in the level of certainty surrounding their final ladder position were:

  • Western Bulldogs - now over 90% likely to finish 9th to 11th
  • Hawthorn - now over 70% likely to finish 12th (and, if not that, then 13th is the only other option)
  • Essendon - now over 75% likely to finish 7th or 8th
  • Richmond - now almost 60% likely to finish 3rd (and, barring that, almost certain to finish 5th or 6th)

Of the positions in the Top 8, there is most uncertainty now about 4th and 8th, with five teams capable of finishing 4th and five other teams capable of finishing 8th.

There is most certainty about 14th on the ladder, with Fremantle about an 86% chance for that position. Next most certain is 1st spot, which Adelaide is now an estimated 82% chance of occupying.

The positions seeing the largest increase in the level of certainty this week were:

  • 7th, with two teams now representing over 90% of the probability
  • 12th, with two teams now representing 100% of the probability
  • 11th, with two teams now representing over 80% of the probability
  • 8th, with two teams now representing almost 60% of the probability
  • 10th, with three teams now representing almost 90% of the probability

GAME IMPORTANCE

Next, let's update the game importance table. It shows, you'll recall, how each team's chances of making the finals varies depending on the result of each of the remaining games. The overall Weighted Ave Importance figure tells us the average amount by which the teams' finals chances are expected to change on the basis of the outcome of a particular game. A larger figure means that the game is expected to change teams' finals chances, on average, by a larger amount and hence is a more important game.

(For more details about the methodology and the rationale for it, again please see this blog.)

To no-one's surprise, the analysis is showing that the most important games in Round 23 are, in order

  • Essendon v Fremantle, which materially affects the chances of five teams
  • Western Coast v Adelaide, which also materially affects the chances of five teams

The Collingwood v Melbourne game is also relatively important, though less so than the other two because it mostly affects only the chances of Melbourne and West Coast.

Of the remaining games, despite what the media are saying, none are likely to have any significant import in the context of the Top 8, though the Hawthorn v Western Bulldogs, and Richmond v St Kilda games might have some significance, contingent on other results.

MOST LIKELY TOP 2s, top 4s, AND 5th to 8ths

This week we'll again look at the team orderings most likely for the Top 2 and Top 4 spots, as well as for positions 5th to 8th. But, with the increasing levels of certainty afforded by the brevity of the remainder of the season, we can also, for the first time, sensibly look at the entire Top 8.

Based on the latest simulations, the most likely final Top 2 is now Adelaide/Geelong, which occurred in over 50% of the replicates. The previous most-likely Adelaide/GWS finish is now only about a 30% chance.

Only three pairings now represent the entirety of the feasible space, according to the simulations.

The most likely Top 4 is now Adelaide/Geelong/Richmond/GWS, which occurred in about 30% of replicates. An Adelaide/GWS/Richmond/Geelong finish is now only second-most likely, having previously been most-likely and occurring in a little over 1 replicate in 6.

Our third most-likely quartet involves the same four teams, but in a different order. Together, these three most-likely orderings account for almost 60% of replicates.

Looking more broadly, the ten most likely Top 4s shown here account for about 99% of replicates.

Looking next at the lower half of the Top 8, we find that two orderings have now become somewhat more likely than the rest, though none has emerged as a clear favourite.

Most common is a Port Adelaide /Sydney/Melbourne/Essendon finish, which occurred in about 12% of replicates, while next-most common is a Sydney/Port Adelaide/Melbourne/Essendon finish, which occurred almost as often.

The eight other orderings shown here each account for about 3% to 8% of replicates each, and all 10 account for almost 60%. 

Lastly, and for the first time this season, let's take a look at the entire Top 8.

Such is the remaining uncertainty about the final ladder that even the most-likely Top 8 ordering carries an estimated probability of just under 6%, while the second most-likely ordering carries a similar probability.

Together, the Top 10 most-likely orderings account for only a little more than one-third of the replicates.

It's hard to recall a recent season where there remained this much uncertainty about the Final 8 with just one round to play.

SIMULATIONS OF THE FINALS SERIES

Applying the methodology as described in this blog to the 50,000 final ladders produced by our original simulations yields the following results.

According to the latest Finals simulations, it remains a race in six for the Flag, though only five teams - Adelaide, GWS, Geelong, Richmond and Sydney - have better than about a 1 in 8 shot. Six other teams have chances ranging from 0.2% to 6% for the Flag.

Compared to last week, Adelaide remain about 35% chances for the Flag even though their probability of taking out the Minor Premiership has dropped to a bit over 80%. GWS' chances for the Minor Premiership have increased to almost 20%, though their Flag chances remain about 15%. Richmond, Geelong and Sydney still all have about 12 to 15% chances for the Flag, while Port Adelaide's Flag chances have remained steady at around 6%.

We see from the heatmap above that, if we define the season in terms of the six events listed there, the most-likely finishes for each team are estimated to be:

  • Win Grand Final: Adelaide
  • Lose in a Preliminary Final: GWS, Richmond and Geelong
  • Lose in a Semi Final: Sydney
  • Lose in an Elimination Final: Port Adelaide, Melbourne and Essendon.
  • Miss the Finals: all other teams

We also see that Adelaide are still estimated to be about 57% chances for making the Grand Final, GWS and Geelong now about 33%, Richmond about 27%, Sydney about 24%, and Port Adelaide about 14%. Melbourne and Essendon are the only other teams with chances of about 5% or more.