When Scores Don't End In 0.033 Or 0.066

Okay, so I didn't quite continue with my amateur evaluation of the code changes with this post (don't worry, there's a four-day weekend coming up for me :P), but in that time I still learned something new by looking through the full results book released by FIG which includes all results, start lists, individual judge scores for each score, tie breaks, etc. from the London Olympics here courtesy of Full Twist.

While looking through to find the judge breakdowns of certain routines (such as the highest a judge awarded to Maroney's TF Amanar vault was a 9.8), I ctrl+f'd Afanaseva (because they didn't use Afanasyeva or Ksenia but instead Kseniia Afanaseva) trying to find her E scores for her floor routine that scored so high with a fall and got to the beam QF results where Afan had a 15.066 and Kyla a 15.075.

(Now, I never really looked to see who came in with what score since I never watched the entire qualifications until well after the games were over and that I didn't want to spoil myself looking at the results, FYI the qualification scores and qualifiers to finals are shown on scoreboards on the screen during broadcast at the end of the last subdivision)

So then I wondered "so how did they go from thirds of a tenth to quarters of a tenth?"

Now, further down where it goes to the individual judge scores, there is an asterisk next to Kyla's score while at the bottom of the result sheets shows the legend, stating that reference scores were used as so:

Something I didn't know, I found this PDF on the FIG site about reference judges:

Basically, it's another pair of judges independent from the five (sometimes six) judges (the E-jury as described) that coughs up most of the scores that we see that are separated by thirds-of-a-tenth increments.  Depending how high (or low) the E score is, there is a certain allowance of how the main five/six judges score average can differ from the reference average (R-score).  When the difference is too large given the general E score from both, the main judge average and reference average scores are then averaged again to bring in a new number, is it happened with Kyla's beam and so many other routines.

In Kyla's case, the initial execution score given by the E-jury was 8.8, but the R-score given was 9.150.

An R-score of 9.150 falls in the range between 9.0 to ~9.4, the E-jury score can't have more than a 0.150 deviation in order to be credited.  Instead, both the 8.8 and 9.150 were taken and averaged to produce the 8.975 score given.  This can also happen when scores end with 0.016 (Aly's 2010 WCH EF FX score), 0.041 (Ponor's 2011 WCH EF BB score), 0.083 (Izbasa's 2010 WCH EF FX score), etc., and has been happening for a while, apparently, but personally, I had no idea what was making that happen.

Now I (maybe you?) know :P?  Just goes to show there are more than the 5 or 6 judges we typically see!

Of course, I happily welcome any corrections, since I'm definitely no expert or professional on gymnastics judging, let alone a certified judge like some people in the gymnastics community are fortunate to be.