So, the season was too long? Deal with it!

There is still a proposal floating around league and team offices across the AFL to adjust the 22-game schedule to something that manifests more fairly for all eighteen teams.

The 17 game round-robin.

The genesis of the 22-game schedule was in 1970 when there were 12 teams in the league, and hence 11 opponents for each team to play twice a season, home and away. This seemed to be the perfect solution for people seeking balance and fairness in the schedule.

Until 1987, this situation was stable. Then the league added West Coast and the Brisbane Bears, upping the membership to fourteen teams. Of course, the 22-game season meant that each team played some opponents once and others twice (specifically, each team had four opponents which they faced only once).

Four years later, Adelaide made the league fifteen teams large, forcing a bye into every round (and seven into the first round!). We had 22 games, spread out over 24 rounds.

Then the league tried 20 games in 22 rounds in 1993, but went back to 22 in 24 the next year before Fremantle’s entrance as the sixteenth team balanced the bye-less 22-round schedule again in 1995.

With 16 teams, every team played seven opponents twice and eight once. In theory, each team would travel to almost every franchise at least every other year, and with rare exceptions every pairing got three meetings every two years.

Fifteen years later, similar adjustments were made with the addition of Gold Coast in 2011 and Greater Western Sydney in 2012. The result, still using that 22-game schedule left over from four decades earlier, was having 17 opponents slotted into 22 slots, leaving just five opponents to meet twice.

That’s not terrible in and of itself. The trick, though, is to balance each team’s schedule so that no team has a particularly difficult set of opponents one year and a much easier set the other year. The AFL, like every other such scheduling institution, has had mixed results doing this.

The imperfection of scheduling in advance, no matter who’s doing it, is the impossibility of knowing how an opponent will change from one year to the next. For the AFL, the methodology is to balance the opponents from the previous year’s ladder by dividing that ladder into thirds.

Those five doubled opponents are selected using approximately these criteria:

• Include “derby” opponents every year. (Adelaide and Port always play twice.)
• Balance the five doubled-up opponents by spreading them out over the three groups of six – ideally, taking at least one from each group and no more than two from any.
• If there’s an imbalance, it should lean towards the group of six which the team itself comes from. (Top six teams should play more top six teams; bottom six plays more bottom six.)
• Try not to repeat doubled pairings from the last year or two.

The difficulty is this: what if the teams you meet in 2017 from last year’s bottom six are, for example, Richmond and Essendon? And the team from the middle tier was Port Adelaide? And the teams from the top six weren’t West Coast or Hawthorn?

On the other hand, perhaps your closest rival had Hawthorn as the top six opponent, North Melbourne and early season Collingwood in the middle, and late season Gold Coast and Fremantle as your bottom six teams?

Not very fair, is it? That’s why there’s a push for something more naturally balanced.

So, what would a 17-game schedule have looked like this year? A schedule where every team simply played every other team once, and the five return games just… vanished?

I tracked that. Each of the five games where two teams met for the second time this season were ignored, leaving a 17-game schedule that (use your imagination) was spread out over 23 rounds. Placement ties were dealt with the old-fashioned way: by looking at the head-to-head match-ups.

Here are the results:

(Richmond was 3-1 against the four other teams which went 10-7; Port was 1-3. Sydney, Melbourne, and West Coast were all 2-2 against the other four, and were 1-1 among themselves, so I went to percentage in their three round-robin games. West Coast was 114 per cent, Sydney 106 per cent, and Melbourne 82 per cent – hence the order shown.)

(Western BD were 2-0 against the other two 9-8 teams; Essendon 1-1, and St Kilda lost to both. Finally, ties for second and 15th were decided by the head-to-head results.)

There’s not much difference between this ladder and the one in your newspaper or online this morning. Same top four, same top eight except for Melbourne replacing Essendon. Same bottom four, in the same order.

Most versions of the 17-game set-up refuse to reduce the length of the season (money talks), and extend the season with a five-game round robin among the top six, middle six, and bottom six teams.

To apply that premise to the 2017 season, we took the top six on the above ladder and have them compete again first in a five-game round-robin. (Note that where there was a second game played between two of these teams this season, I used that game for this round-robin, since this would be their second game this season.)

Here are the results given the games played this season:

1. Sydney (4-1, losing only to West Coast in R4.)
2. West Coast (3-2, defeating Adelaide in R23. Suddenly that game is important!)
3. Adelaide (3-2, losses to the Eagles and Swans.)
4. Geelong (2-3, def. Richmond 80-66 in R21.)
5. Richmond (2-3, wins against West Coast and GWS in the R18 rematch.)
6. GWS (1-4, with a victory over the Eagles in the R22 rematch.)

So, presumably, this 17-5 schedule gives West Coast the second seed in finals – home games, double chance, the works! Sydney’s 0-6 start becomes irrelevant: all they needed to do was make the top six and then slam through to the top seed in the final through those last five games.

If we’re taking seeds seven and eight from the middle tier (some versions suggest that the winner of the bottom tier round-robin should get a play-in chance against team number eight or something along those lines), here’s how their five-game round-robin would play out under the same rules:

7. Melbourne (4-1, and it’d be 5-0 except for losing the rematch to the Pies in R23!)
8. Essendon (3-2, winning the percentage tiebreaker with Port and the Bulldogs at 121 per cent.)
9. Western Bulldogs (3-2, with a percentage of 107 per cent against the Dons and Power.)
10. Port Adelaide (3-2, missing out on finals with a 75 per cent against Essendon and the Dogs.)
11. St Kilda (1-4, defeating Collingwood in R4.)
12. Collingwood (1-4, the Melbourne rematch their only victory.)

Ignoring the idea of Hawthorn getting a play-in chance by winning the bottom six round-robin (using the same rules, there were three teams at 3-2 and three at 2-3, with Hawthorn besting Fremantle and Carlton by percentage for the top spot), that would put the Demons and Bombers on the bottom two rungs of the finals ladder.

Depending on how you want to run your imaginary finals brackets, here are your two possible September finals fixtures:

Using the pure 17-game schedule
Richmond @ GWS and Adelaide @ Geelong, qualifying finals.
Port Adelaide @ West Coast and Melbourne @ Sydney, elimination finals.

Using the 17-5 schedule as played out above
Geelong @ Sydney and Adelaide @ West Coast, qualifying finals.
Essendon @ Richmond and Melbourne @ GWS, elimination finals.

For comparison, here’s what’s happening in reality:
GWS @ Adelaide and Richmond @ Geelong, qualifying finals.
West Coast @ Port Adelaide and Essendon @ Sydney, elimination finals.

I don’t care for the 17-5 schedule at all – it seems to unfairly weight the last five games. But the pure 17-game schedule lines up fairly closely with the AFL’s 22-game schedule; the differences in the results are negligible (unless you’re a Demon fan!), and I’ve yet to hear an argument for shortening the season that flies with me.

For me, the conclusion seems obvious – leave the schedules alone.

As an afterthought, if you prefer the 18-game model, where traditional rivals get a “home-and-home” series every year, the results are still about the same as the 17-game model: