1580 Ranking Points!!!

[mharky]

This is ridiculous...

http://www.britishorienteering.org.uk/index.php?pg=results&eday=59273&course=9002&

[Big Jon]

Come to Deeside and get great orienteering and loadsa points!

next level C event is noted here:

http://www.marocscotland.org.uk/home/fixtures/

[LostAgain]

1167 for 2nd on the Blue when beaten by nearly 5 minutes by a 14 year old also looks a little excessive. I know the 14 year old is good but still seems excessive.

[greywolf]

Might be worth double checking the BOF numbers - the freak points distribution* on brown is almost entirely due to the inclusion of Mr Wesley Musall, M60 from SUFFOC - previous average points score 327** - was it really him, running way better than might have been expected on past performance? (my apologies if so!)

This is a systemic weakness [edit: vulnerability would be a better word] in the algorithm - it could (and almost certainly did) happen in the old rankings lists (but they had so many anomalies no-one would have noticed) - and could equally also happen if you restricted the list to Level A events.

*Oleg got many more points than he should, but most of the rest got fewer than normal.
I reckon if he's omitted, Oleg would get c. 1360 points, and Pete Lawrence & Bob Daly c. 1090, rather than 870

** The explanation would take some time - but crudely, his inclusion greatly increases (more than doubles) the SD of previous competitors points, and therefore more than doubles the spread of points awarded at this event - those at the front get more than they should, those behind get less...

[Scott]

According to the results on the Maroc website, Wesley Musall was actually Dennis MacDonald of GRAMP. I would guess that somebody on the entries desk made a mistake with the BOF number.

[andy]

If this is the case I find it quite amazing that one runner in 12th/18 could make such a massive difference in points

[Neil M40]

A quick scan of the rest of the results and I have found another anomoly - 14th on Light Green is Abigail Mason W12 according to BOF, and Althea Dicken W50 according to MAROC. I know this one doesn't affect rankings, but it shows how many mistakes there are.

This must be the reason I am not ranked higher than I am.

What can BO do to prevent it? Any cross checking of name and age against BOF number when results are submitted will throw out lots of spelling errors and typos when there is no problem with the BOF number, and give the results official a pile of extra hassle.

You could introduce a check-digit on the BOF number but that would mean piles of work for BOF, and I struggle to remember six digits never mind seven.

[King Penguin]

At the Maroc event Brown had time range 48 mins to 1:23 and points range 1580 to 648.
One week later, also in Scotland at ELO, Brown had time range 42 mins to 2:13 (1:15 excluding last place) and points range 1218 to 336 (899 excluding last place).
(http://www.britishorienteering.org.uk/i ... eday=59599)
Similar time range (ignoring last place) but significantly different points range.
I guess it must be due to significant differences in standard deviation.

[graeme]

! srebbid uoy egnar elitrauq-retni esU

[greywolf]

I guess it must be due to significant differences in standard deviation.

No, it was because the results sent up on the BOF website included an incorrect BOF number. Did you not read the rest of the thread?

[graeme]

Did you not read the rest of the thread?

Maybe he read Appendix K...

2.2.3 At ranking events, the mean of the current scores of the ranked runners who complete a course is awarded to a runner matching the mean time of those ranked runners. The standard deviation of the current scores of those ranked runners is also calculated along with the standard deviation of their times. For these four calculations, the final 10% (rounded up) are ignored e.g. if there are 38 ranked runners on a course, only the first 34 are used for the calculation.


I think this means that your explanation can only be right (as it surely is) if they haven't coded the algorithm properly...

[DJM]

mharky wrote:
just shows how jobbie the ranking system is, and
This is ridiculous...

Good to see such constructive comments - they really help

Actually they do, as we need eagle-eyed competitors to spot and report anomalous results. The changes we've recently made to the algorithm have been instigated largely through our being made aware of previous anomalies through nopesport posts, so ... many thanks

Also, on this occasion I tend to agree with Big Jon's previous "garbage in garbage out" comment since, if BOF numbers have indeed been incorrectly keyed in, we do have a real instance of "GIGO".

Does anyone know if MAROC will be checking BOF numbers and re-entering the results to the British Orienteering website so we can correct the anomalies?

[Scott]

I think this means that your explanation can only be right (as it surely is) if they haven't coded the algorithm properly...

It's a bit ambiguous, though - it could be interpreted as "ignore the slowest 10% on that course for all calculations", rather than "ignore the slowest 10% for calculating the mean and s.d. of time, and ignore the lowest 10% of current scores for calculating the mean and s.d. of current scores" (which I assume is what you are suggesting, although this is all a bit beyond my feeble arts-student brain ).

I believe the first interpretation was the intended one anyway - I think the idea was to discard people who had got hopelessly lost on their course.

[DJM]

graeme wrote:
! srebbid uoy egnar elitrauq-retni esU

To which my reply is "!oot snoc rieht evah segnar elitrauq-retni"

graeme also wrote:
I think this means that your explanation can only be right (as it surely is) if they haven't coded the algorithm properly...

If by this you mean that "Wesley Musall"'s points shouldn't have been included in the points calculation, then you're not correct as it's the last 10% of the ranked runners measured by the time they took who are ignored, not those with the lowest points.

I've checked the new algorithm by independently calculating points on several specimen courses and I agree with those on the website every time. I have no evidence to suggest the coding has been done incorrectly.

[buzz]

[quote="DJM] Also, on this occasion I tend to agree with Big Jon's previous "garbage in garbage out" comment since, if BOF numbers have indeed been incorrectly keyed in, we do have a real instance of "GIGO".[/quote]

Blaming an error in a BOF number seems a lame excuse. Surely even the most basic of statistical algorithms could spot anomolies of this nature. If someone suddenly gets thousands more points than they have previously managed just leave them out of the calculations until they are validated. Or is that just too simplistic?