it does because we do not know the true difference. for example, say the true difference between Liukin's routine and He's routine is a 9.1 and a 9.0 respectively, then the true difference would be 0.1 in favor of Liukin. However, if one judge gives Liukin a 9.2 while giving He a 9.0, the difference only deviates 0.1. However, if the judge gives He a 9.2 and Liukin a 9.0, then the diviation is not 0.1 from the "true value" but 0.3. since we do not know the "true value" in this case, we cannot determine the "true deviation". However, the main point is...it matters whether the difference lies in favor of Liukin or in favor of He. Therefore, the 0.15 average deviation you suggest MUST indicate who it is in favor of.
inaccurate statistical models...yes. you cannot claim statistical significance if you model contains too many possible inaccuracies in the first place. it's been a while since i've taken a stats class, but a puny pool of 6 samples is way too small givin the possible variations.
just out of curiosity, which model were you using?
the IOC must check before the final scores are released to the public. That's standard procedure regardless of this incident. For example, in the women's all around, the judges had to recheck Nastia's balance beam score.
now you are just pushing it. the australian judge is psychic and knows what the max and min scores will be? didn't you just say earlier that each judge judges differently? how does the australian judge know that other judges won't go super easy on both girls?
you know how to deduct for each mistake or exactly which mistakes there are? sorry, i didn't know you were an professional gymnastic judge.
it all comes down to whether 0.3 is a significant deviation. 6 is clearly a puny sample size given the possible variations. why don't you look at more data? take the pommel horse tie for example. the "average" difference was 0.3. while the romanian judge thought that the two performances only differed by 0.1, the bulgarian judge believed they were different by 0.7. what now bulgarian judges also suck at their job?
http://results.beijing2008.cn/WRM/EN...html#GAW007101