All animals are equal, but some animals are more equal than others.
When I think of the Super Rugby conference system George Orwell’s Animal Farm comes to mind. How equal are the different conferences, and also how do the different systems used compare to each other?
To most people the fairest system is a single table format, no conferences, playing everyone else twice home-and-away and with byes before significant travel and thus negating jet lag and travel wear and tear.
Of course this ideal is impossible. The closest compromise that a lot of Roar posters yearn for is the golden age of Super 14 where everyone played each other once.
The reasons are that the best teams make the finals and they finish in a strict order of merit for finals rather than by placement within unequal conferences. Let me label this the traditional system.
I thought I would add some quantitative measures to this emotional subjective discussion. I thought I would turn to data as opposed to emotional about that the new conference system. I wanted to compare 2015 results under the new four conference system to 2015 results under the traditional system.
The details and outcomes of converting the 2015 results to the new four-conference format is in my previous article.
This would allow a quantitative indication of how much the new system advantages or disadvantages teams from a traditional system. Also as an extra benefit it provides a quantitative indication of the fairness or otherwise of the previous conference system. Was the old conference system so bad? This comparison provides insight into answering that question.
I took 2015 year’s results. I applied the new bonus points system with the calculations as supplied by Brett McKay in his article. Then I removed the duplicated intra conference games. The 2016 draw had supplied two of the four to be removed. By maintaining two home and two away games it fell out simply which other two games had to be removed. I added in the missing inter conference games from the other conferences using 2014 results. Thus 14 games instead the 16 from 2015.
2015 with new bonus system
The first is just a summary of Brett’s McKay’s applying the new bonus point system.
2015 using 2016 conference system
The next shows the results of the changes wrought by the new conference format. If the four conference system had been used in 2015.
2015 using the traditional system
The next shows the results of the changes wrought by reverting back to a traditional system of a single table and no conferences.
Reverting back to a traditional system for 2015 would have impacted the South African teams the most. In fact no South African teams would have made the finals.
Their place was taken by the Crusaders. What is the bias in the old three conference system? This can be seen by summing up the points difference you can see the gain per conference and adjusting for the number of games. The traditional method would involve 14 games while the three conference system involved 16 games.
I believe the results are contrary to popular belief visa a vis the three conference system.
The Australian conference has negligible bias. On a per team basis a detriment of -0.6 points.
The New Zealand conference has a minor bias of -1.6 points per team i.e. approximately one draw in a season against it.
The South African conference has a small bias towards it of 3.2 points. Almost one game per team.
So all in all there was not much bias in 2015 with the three conference system in terms of points, with the main impact actually being the determination of finals placings by conference leader.
The difference between the traditional and 2016 conference system using 2015 data
The next table shows the results of the changes wrought by the new conference format compared to a traditional system on 2015 results.
There is a very significant difference for one conference between the traditional system and the new four conference system.
The Australian conference has a small bias towards it of 2.6 points per team. Approximately one extra drawn game per team.
The South African conference two has a small bias towards it of five points per team. That’s one bonus game win per team.
It is as expected for these two conferences to be so similar since they have the same draw, play every New Zealand team once, and play the other South African conference once.
The New Zealand conference has a negligible bias against it of -0.2 points per team. This is better than the three conference system because it has less games within it’s own conference and one more against the Australian conference.
The South African conference one has a major bias towards it of 16.3 points per team. This is an advantage of an extra four wins per team.
The traditional system and the three conference system are quite equal but the four conference system is quite skewed and it is easy to see who the pigs are in this animal farm and who are the sheep, cows and horses.
For an Australian or New Zealand Team to top the table they need a very good year. For a South African team to top the table they need to be in the conference not playing New Zealand teams and only have a reasonable year. South African teams do well overall against Australian teams but poorly against New Zealand teams, in fact not as well as Australian teams. So one South African conference not playing any New Zealand teams is a significant bonus.
Additionally the other significant advantage gained is the top team in the South African conference that plays the New Zealand teams (and hence has a very similar draw to the Australian teams) has a poor year yet can jump from tenth to fourth due to the system of home team finals for conference winners.
It seems the SARFU is calling the tune, but who are we to complain? The decisions they made did allow for an expansion to add in an Argentinian and a Japanese team. It also allows for further expansion quite simply.
This is neatly summed up by another line from Animal Farm.
No one believes more firmly than Comrade Napoleon that all animals are equal. He would be only too happy to let you make your decisions for yourselves. But sometimes you might make the wrong decisions, comrades, and then where should we be?