We keep hearing that the 2017 November tests are mainly for building depth of national squads and that the results are irrelevant. So, which results are most important?
To answer this question, we will use 2012, 2013 and 2014 test match results separately to predict the 2015 Rugby World Cup (obviously by pretending we don’t know the World Cup results).
This analysis is broken down into two steps before reaching the conclusion.
Step one, we use the world rankings table straight after the 2011 World Cup as the “base table”. And we update this base table with each of the 2012, 2013 and 2014 match results to create the three tables below.
The detailed rankings system explanation can be found on the IRB website here.
For example, the “Post world cup + 2013” table below is the IRB world rankings table straight after the 2011 Rugby World Cup updated with the 2013 test match results (i.e. 2012 and 2014 test match results are ignored). The same logic applies to the other tables.
We see that Ireland had 80.2 and 78.7 ranking points on “Post world cup + 2012” and “Post world cup + 2013” tables respectively but as high as 86.2 on the “Post world cup + 2014” table. This will be obvious for some because they won the 2014 Six Nations but had a disappointing 2012 and 2013 Six Nations (3rd and 5th respectively).
Step two, we investigate how predictable the 2015 Rugby World Cup results were for the three tables. This is done by analysing how much ranking points were exchanged for each match.
For example, on all three tables below, there were zero ranking points exchanged when England beat Fiji in the opening match because this was what all three tables (and almost all neutral fans) expected. However, when Japan beat the Springboks, all three tables exchanged 2.0 points from South Africa to the Brave Blossoms because this was obviously a surprising result.
However, “Post world cup + 2013” table exchanged 1.7 ranking points from France to Ireland when Ireland beat France but “Post world cup + 2014” table exchanged only 0.7 because of the Irish Six Nations success in 2014 unlike their results in 2013.
Therefore, the more predictable the outcomes were for the table, the less ranking points are exchanged and vice-versa.
The table below shows the ranking points exchanged for each match (upsets highlighted) on each table. “Post world cup + 2012”, “Post world cup + 2013” and “Post world cup + 2014” tables’ ranking points exchanges are shown on column 1 (i.e. on the “[,1]” column), 2 and 3 respectively.
The 2013 and 2014 results were almost as useful for predicting the 2015 world cup results as each other, the 2012 results were about 14% less useful.
A detailed reasoning is given below:
A total of 24.6, 21.7 and 21.5 points were exchanged on the 2012, 2013 and 2014 tables respectively. Therefore, the 2015 World Cup results were most predictable based on the 2014 test match results (as we would expect since they are the most recent results). But the 2013 results made the World Cup results as almost as predictable with the “Post world cup + 2013” table exchanging around 1% more points than the “Post world cup + 2014” table.
This is a lot better than the 2012 results with the “Post world cup + 2012” table exchanging around 14% more ranking points than the other two tables.
The lens applied to these results is an extension of the discussions that have previously taken place here.
Author: Kaito Goto