ranking list

[Adventure Racer]

I don't think anybody is suggesting otherwise Eddie - the only point of the graph being to show the O and longO discrepancy. I don't think there was any intention of proving anything else with it.

[Snail]

I thought that the interesting point of the graph was that the mean min/km rate(edited for Martin's benefit!)for Long Orange / Orange are not much out of line with the 'adjacent' courses. That might imply that using the average to determine which runners are discarded could still potentially produce sensible rankings(?)

If the problem arises when a few highly ranked runners "run down" and effectively set ranking points for others on the course, would a way of addressing the issue be to calculate rankings for everyone, including juniors, but only to publish those rankings for over 18s (or whatever is the current policy from time to time)? That should dilute the impact of having only a few ranked runners per course.

It could also address the seeding issue that NeilC mentioned - the rankings for say first year 18s could potentially be made available solely to the organiser of that race.

[martin]

Perhaps Spookster's graph and point would have been clearer if he had plotted speed (eg metres/second) rather than the inverse speed of mins/km that orienteers seem to confuse with speed.

[Miner]

I'm not really inclined to try to apply the ranking formula - too much effort to render meaningless results. Suffice to say I've noticed I got 906 for coming second on Green at a recent regional event while the only adult on orange got 1208 for finishing halfway down the results. I'll echo others views on the software - I used to be an analyst/programmer and there would be no dificulty in including the functionality required - perhaps it was never mentioned in the original spec?

My other bugbear is the exclusion of night events from the ranking list - night O seems to be increasingly marginalised, especially in the classic cross-country format. There are several local leagues of night score events but few regionals and the SENiLe format seems now to be deceased. Excluding night O while including TD3 events, sprint O and trail races seems like just another nail in the coffin.

[NeilC]

My other bugbear is the exclusion of night events from the ranking list - night O seems to be increasingly marginalised

When the draft Appendix was circulated I did ask whether there was any evidence that night events were any more anomalous than say sprint races. The response was that historically night events had given anomalous results.

[Big Jon]

"historically night events had given anomalous results" what this means is that the speed merchants who win by day don't always win by night. You could probably say that races in tough technical terrain give anomalous results and exclude them too! Night-O is the most technical form of O, shame the rankers don't like it!

[Scott]

There was a lot of discussion in the workgroup about whether or not to include night events - I think it would be fair to say that we were pretty evenly divided. The responses to the consultation document were what swung it in favour of not including night events (although both sides of the debate were well represented in those as well).

[graeme]

Perhaps Spookster's graph and point would have been clearer if he had plotted average velocity (i.e. in the required direction) rather than the inverse velocity of mins/km that orienteers seem to confuse with speed.

[Adventure Racer]

Unsurprisingly for those who know me, I'm another one disappointed at the exclusion of night events. Not that it really makes that much difference to rankings - despite doing lots of night events (typically at least as many night as day events during the night-O season), I probably only ever did one or two a year which would have attracted ranking points. As others have said, there's certainly no reason why the results would be any more anomalous than urban events or others which are making their way onto the rankings list.

[mappingmum]

. . . .historically night events had given anomalous results.

Night events have a greater justification for inclusion than urban. If urban is now included, it's a no-brainer that night-O should be too.

[SIman]

I have just returned to this thread after a few days and I can shed a little light on the results from the Edinburgh City Race.

The raw results file was sent to BOF and the uploading to the results and ranking lists was done there. Because of this there was no further checking by anyone involved in the event to see if it was sensible.

For example, the winner of the Yellow Course was Alex Carcas M12 of INT, but due to an aparently minor mixup during registration, he has entered together with Jason Inman's BOF number. As has been explained, Jason himself ran the Short course first. He then ran the Long course but was marked n/c and that was included in the file sent for processing but the processing has ignored this. I have also noticed that his BOF number was missing from the data for the Short course, but he has been given points.

Obviously there are a number of teething problems.

[andypat]

[quote="Miner"]Suffice to say I've noticed I got 906 for coming second on Green at a recent regional event while the only adult on orange got 1208 for finishing halfway down the results. quote]

I suppose thats down to the person on orange being given their own average points because there was noone to beat. Personally I think the unified ranking list should at least stick to TD5 events only (wasnt that in the spec?).

Otherwise it would be a bit like me demanding inclusiuon in the Scottish football sqaud because I've been scoring a barrowload of goals in the local Pub league. Its not comparing like with like.

And while I'm on it - of course night events should be included! (Although I believe the Scottish Night champs is maybe now a level 3 event for purposes of levies maybe so would be excluded anyway)

[greywolf]

"historically night events had given anomalous results" what this means is that the speed merchants who win by day don't always win by night.

Nope it was more to do with the perception that night events were more likely to have very small fields e.g. last year's Scottish Night Champs had 5 courses, with 4, 24, 16, 7 & 4 finishers respectively - which increases the chance of anomalous scores - if one person in a field of 4 has a poor run they really do gift loads of points to everyone else - in a field of 150 e.g. the Edinburgh city race the effect is trivial.

But as Scott says, the discussion about night events was fairly evenly balanced. If the night-O enthusiasts want L2 night events to be ranking then make the case to BOF - unlike some of the other tweaks proposed, the inclusion of (future) night events requires no fiddling with the algorithm or the interface and should therefore be straightforward and easy to implement

Personally I think the unified ranking list should at least stick to TD5 events only (wasnt that in the spec?).

The problem with that is you'd effectively exclude many orienteers from the ranking list, as much of the country doesn't have TD5 terrain. As many (traditional, cross-country) L2 and some L1 events are run at TD4 (regardless of what the courses are called) the working group recommended ranking TD4 & TD5 courses at traditional cross-country style events.

[NeilC]

Nope it was more to do with the perception that night events were more likely to have very small fields e.g. last year's Scottish Night Champs had 5 courses, with 4, 24, 16, 7 & 4 finishers respectively - which increases the chance of anomalous scores

Then the logical solution would have been to not allocate ranking points to courses with less than n runners rather than ban a whole discipline. And of course the ranking of one or two seniors on orange courses demonstrates the pointlessness of banning night events by this criterion.

I've gotten nowhere attempting to lobby a board member on this issue. Who else in BOF would you recommend approaching?

[andypat]

I see what you mean re the terrain. Maybe the argument is more to do with the algorithm then. Something along the lines of tennis where its worth less ranking points to win a lesser tournament. Maybe a penalty applied to courses that are a lower TD than the highest TD for the event.

Say at a colour coded, Black, Brown, Blue, Green are 100%, light green get ranking less 10% and orange less 20%?

MAybe the algorithm already does this - I'm not a mathematician .....