Lies, damn lies, and statistics: the faulty reasoning behind why Athletics Australia wants to send only half of our marathoners to the London Olympics
- Created: Wednesday, 08 February 2012 20:03
When Lauren Shelley ran an IAAF A-qualifying performance at the Osaka marathon at the end of January, the marathon became the first event that Australia has qualified a full contingent of male and female athletes to compete at the London Olympics.
Yet Athletics Australia would prefer to send half of them.
Martin Dent (2:12:23), Jeff Hunt (2:13:14 and 2:13:19), Shawn Forrest (2:14:37) and Shelley (2:35:57) exist in a twilight zone where they have bettered the IAAF A-qualifiers of 2:15:00 and 2:42:00, but haven’t met the stricter 2:12:00 and 2:32:00 standards set by Athletics Australia; Michael Shelley (2:11:23 and 2:11:38), Benita Willis (2:28:24) and Lisa Weightman (2:29:23) have all met the AA mark.
The mystery – unknown to senior figures of the distance running community until now – as to why Athletics Australia has chosen to derogate from the qualification standards set by the IAAF for the marathon, were revealed to Inside Athletics by Athletics Australia High Performance Manager, Eric Hollingsworth.
"Well first of all it's worth remembering that across the board our entry standards are set in order to ensure we take the most competitive team possible to major championships, it's about quality not quantity,” said Hollingsworth via email.
“And the marathon is no different.”
Yet, it the marathon is the only event which Athletics Australia sets a standard different to the IAAF’s benchmarks. And the rationale for doing so demonstrates a fundamental lack of knowledge of championship marathon running.
According to Hollingsworth, the duration of the qualification period, depth of performances in non-championship marathons and the number of athletes who reach the IAAF standard are reasons for AA tightening the mark.
“Adhering to an Athletics Australia standard, as opposed to the IAAF 'A' standard was in place prior to my arrival at the organisation, but the premise was good, which is why we have continued it.
“Comparatively across all men's and women's events at the world champs the total number of athletes to achieve the IAAF standard in the marathon would be disproportionally a lot higher than an 'A' standard in any other event, by quite a large margin.
“However, the stats show that the AA standard would be in line with the other events and therefore suggest it's a competitive mark.
“It sounds confusing, but the stats don't lie!”
The statistics may be true, but are being used to support a faulty premise: that prior performances are a good predicator of championship marathons. A more thorough statistical analysis shows this not to be the case.
Compare the graphs below, which plot the seed performances (based on personal bests) against the finishing place in the two most universal of Olympic events, the 100m and marathon, from the Beijing Olympics.
As expected, the graph for the 100m shows that there is a high correlation between the personal best of athletes prior to the Olympics and their finishing place (the trend line explains 87% of the data, an extremely high correlation). However, the graph for the marathon demonstrates virtually no correlation between personal best and finishing position, with the trend line explaining only 15% of the data.
Further, the data for the Beijing marathon highlights a number of interesting examples regarding competitiveness: Zimbabwe’s Mike Fokoroni, who started the Beijing marathon with the 74th fastest personal best at 2:14:01. He improved his personal best to 2:13:17 and finished 11th; Sixteenth place, the equivalent of making a semi-final in many track events, ran 2:14:00.
Last year’s world championships, witnessed by Hollingsworth as the national team manager, demonstrated a similar dynamic, with 2:15:56 required to finish in the top 16; Only the top 8 bettered Athletics Australia’s qualifying standard of 2:12:00.
Looking closer to home, one needs only to consider the performance of Martin Dent at the 2009 world championships in Berlin. By Athletics Australia’s own selection rationale he wasn’t deserving of individual selection, but picked as part of the World Marathon Cup team. With a 2:14:46 personal best at the time Dent finished a creditable 21st in 2:16:05. His personal best is now 2 minutes, 23 seconds faster, which if applied to the Berlin results would have seen him finish 8th.
Likewise, Hunt has been consistent at the 2:13 mark over the past two years, was the only Australian to toe the line at last year’s world championships and has a personal best of 2:11:00 with a come from behind performance on his debut. In fact, he is the fastest Australian marathoner during this Olympiad, ahead of Shelley, who has won a Commowealth Games silver medal. Surely the potential is there as much as any other who has met the IAAF’s A-qualifier for success at a championship.
Perhaps the most faulty premise regarding selection is Hollingsworth’s assertion that he prefers quality over quantity. Picking additional athletes does nothing to diminish the quality of performance of those who would be selected with a minimalist team, and only provides the opportunity for others to achieve quality performances. And that is particularly true when it comes to the marathon.