2026 : Simulating the Final Ladder Before Round 0
/PREAMBLE
Welcome to the first season simulation for 2026.
As you might imagine, after more than a decade of simulating the remainder of men’s AFL seasons, the code required to do so has become … substantial. It now contains lines strewn throughout to handle this or that particular edge case — the introduction of a Round 0, the reintroduction of byes, including byes in the first round of a season, games that were supposedly from the same round but played months apart, and so on. It’s less a script now and more a living record of pandemics, bad weather, and competition rejiggery.
So when I heard that the League had confirmed that the 2026 season would involve a wildcard round, I looked forward to updating my code much as an American turkey looks forward to Thanksgiving.
This year, though, I’ve been working quite a bit with various LLMs — not least of all ChatGPT — and so I came up with the ultimate avoidance technique: I asked ChatGPT to make the change. It was surprisingly good at it, quickly forming a view about the intention behind different sections of the code and, more broadly, what the whole script was meant to be doing. At one point it even recognised that the Finals appeared to be played a week apart and quietly amended the code to ensure that they were. I’m not sure whether to be impressed by that or slightly unsettled.
I also gave ChatGPT free rein to improve the speed of the simulation. By the end of it, the code was producing 20,000 replicates in just over three hours — which, by historical standards, felt brisk.
Someone suggested that I might rewrite some of the especially slow sections in C++. An excellent suggestion, rendered problematic by one small detail: I’ve not written C code since 1984. I figured ChatGPT’s knowledge might be a little more recent than that, so I asked it to rewrite the entire R script in C++.
And so it is that I now have code that can generate 1,000,000 replicates in under nine minutes. The only material capability foregone is the maintenance of a game-by-game record for every simulated season — a feature that would require the creation of a modest 216-million-row table. That shouldn’t be too hard to reimplement, assuming I can find somewhere to store it.
One thing worth noting is that forecasts based on 1,000,000 replicates are not necessarily better than those based on far fewer. They are simply more precise estimates of the implications of the assumptions embedded in the code, because they reduce random sampling noise. It is entirely possible, however, that they are merely more precise estimates of incorrect values …
Anyway, let’s see what 1,000,000 simulations have to say about 2026.
RESULTS
This year’s post pre-round 0 simulations suggest that the teams can be roughly grouped as below:
(For anyone who is curious about the precise methodology I’ve employed, see this blog post over on the General Probability and Statistics journal)
Brisbane Lions: about 95% chance of being finalists (including the wildcard finals); roughly 70% chances of Top 4; roughly 30% chances of being Minor Premier
Geelong, Western Bulldogs, and Hawthorn: about 85% to 90% chances of being finalists; roughly 45% to 55% chances of Top 4; roughly 10% to 15% chances of being Minor Premier
Collingwood, GWS, Adelaide, Gold Coast, and Fremantle: about 65% to 75% chances of being finalists; roughly 20% to 30% chances of Top 4; roughly 5% chances of being Minor Premier
Sydney and Melbourne: about 55 to 60% chances of being finalists; about 15% chances of Top 4; about 2% chances of being Minor Premier
Carlton, Port Adelaide, and St Kilda: about 35% to 45% chances of being finalists; about 5% to 10% chances of Top 4; about 1% chances of being Minor Premier
North Melbourne, Richmond, Essendon, and West Coast: about 10 to 20% chances of being finalists; about 1% chances of Top 4; <0.5% chances of being Minor Premier
LADDER FINISHES
The home and away ladder and Finals projections appear below.
We have then 11 teams with about or better than 5-in-9 chances of finishing Top 10, eight with about or better than a 1-in-4 chance of finishing Top 4, and four with about or better than 1-in-8 chances of finishing as the Minor Premier.
As a broad commonsense check on the numbers here, a quick calculation of the correlation between teams’ expected wins and the combined team rating that they will take into their first game of 2026 is +0.98. A chart of the underlying data appears at right.
The line of best fit is:
Expected Wins = 11.5 + 6.15 x Rating
It explains about 96% of the variability of Expected Wins.
The teams that are furthest away from this line - presumably because their fixture is significantly different from average - are:
Brisbane Lions (more Expected Wins)
GWS (more Expected Wins)
Carlton (fewer Expected Wins)
Essendon (fewer Expected Wins)
This is consistent with the first table in the earlier Strength of Schedule analysis for 2026, which had Carlton and Essendon as having the hardest fixtures, Brisbane Lians having the easiest, and GWS having the third easiest.
FORECASTS FOR THE FINALS
Looking next at Finals prospects, we have:
Flag favouritism is currently, in order (with fair prices shown in brackets):
Brisbane Lions (about $4.25)
Geelong (about $7
Western Bulldogs (about $7.25)
Hawthorn (about $9)
Collingwood (about $14)
Adelaide (about $17.50)
Gold Coast (about $22.50)
GWS (about $24)
Fremantle (about $25)
Melbourne (about $35)
Sydney (about $36)
Carlton (about $55)
St Kilda ($100)
Port Adelaide ($150)
COMBINATIONS IN TOP X TEAMS
In the latest 10,000 simulation replicates there were:
1,000,000 different Top 18s
999.453 different Top 10s
34,659 different Top 4s
LIKELY IMPORTANCE OF PERCENTAGE
As the final table for this week, we’ll use the simulations to investigate the estimated likelihood that any given pair of ladder positions will be determined after Round 24 based on percentage.
So, our current estimates are:
MINOR PREMIER: There’s about a 1-in-6 chance that 1st will be decided on percentages
DOUBLE CHANCE: There’s about a 2-in-7 chance that 4th will be decided on percentages
AVOID WILDCARD ROUND: There’s about a 3-in-10 chance that 6th will be decided on percentages
MAKE FINALS: There’s about a 1-in-3 chance that 10th will be decided on percentages
