Discussion: Long O Results

Long O Results from Fair Hill are posted on the DVOA event website. Well, sort of. They're there but the hour field got chopped off of the results. Working on getting that fixed....

They look better now.

splits loaded here too. don't forget to claim yours if that didn't happen automagically.

You guys are amazing. Is this going to be the new program USOF adopts? I hope so because it rocks.

Kenny, this is an incredible service to the o-geek community. Thanks so much for your superior site and for putting this stuff up so quickly!

As per Boris request the map of blue course with my routes - part 1 , and part 2

By the way, this is an absolutely, totally excellent Long O' champs event. Outstanding map detail and design that made for a huge number of route choices that needed to be made and executed well. Not as hard as a Long O' at Harriman, but a lot more fun for a 'Long' event in my mind because you _could_ plausibly try to optimize between up to a dozen route choice variations made possible by this map on each leg - which is probably getting near capacity for reading while running. (When it gets near infinite choices on long legs at Harriman, it's mostly triage to find something that looks okay while zoomed out, then you can micro-optimize the details as you go, which I find fun too, hard to do with the precision I felt was possible on this map!)

Plus the great weather (okay, that was good fortune) allowed for a pleasant run, and the fields/trail sections made for good footing that made it fun to read ahead and think about the route choices and pre-plan execution for the forest bits. Never a moment to spare mentally, even with a few ~2km legs to run.

Awesome courses, awesome choice of terrain for a Long O', awesome mapping of that terrain, great finish stadium, and of course what started this thread, great results! Even AP results!

----

Switching gears as to JF's question of whether USOF is switching to AP-like rankings/results reporting. Short answer: that's largely up to K.W.

I've heard USOF will be soliciting bids shortly for a system to hold/calculate/display the results/USOF-rankings online. By it's nature, I think that system would have to be independent of AttackPoint even if KW created it. With USOF rankings reporting:
- users don't control their own results reporting, non-epunch things have to be loaded by event director/rankings coordinator, name spelling issues need to be dealt with & cleaned up, splits are not needed (nor available in many cases)..., things like DNF/DQ/SPW & even 'credit days' need special handling... USOF membership should be checked) So it's not AttackPoint.

Still, it's looking like the USOF rankings system will become pseudo-real-time on-line if Clare's momentum continues. The remain question is whether it will look kinda like AP rankings, or like CF rankings, or DVOA rankings, or some other thing.

I wonder if the points promotion at http://www.attackpoint.org/rankings.jsp
has something to do with all the extra results? Doesn't look like it happened on purpose - or did it?

points promotion?

I had noticed that the rankings page said my ranking points exceeded 100 while the overall AttackPoint ranking system still thought my ranking points were in the nineties. Of course, I also noticed that a few races were somehow giving 0 points to everyone who ran them (now fixed) so it seemed clear that the software plumbing behind the rankings was in flux. Now everwhere agrees with over 100 - maybe Ken has decided to abandon normalisation of the top three runners around 100. Just imagine what'll happen to Eddie and Nadim's scores if the same soon applies to the points for discussion posts and leg comments.

I'm a little puzzled as to the points awarded for the Green-X Long course -- they seem to be lower than what I might have expected. Is this because the distance isn't given?

Yep, by points promotion, I meant that many at the top went over 100 points, vs. the top 3 averaging 100 before.

The bottom of the rankings has some issues too - I noticed a few people (e.g. Gail G.) have an average of zero points, despite having a few non-zero runs. And the zero's are probably not ideal to begin with.

Tricky getting this all working right, I'm sure. Thanks KW.

Thanks for the explanation. You know how you get a word in your mind? By promotion, I was thinking putting it forward, promoting it as in advertising it...

I didn't change anything on purpose, looks like a new bug related to unclaimed SI splits.

gail's brown course shouldn't be ranked at all because there are only 3 AP runners, but it's probably counting the SI people too which is throwing it out-of-whack.

yep, all good now. at least I know you guys are paying attention ;-)

Looks better now. At least back to what it was with the low-scoring-Canadians bias :) And there's even been some intermixing. Odd that doesn't clear itself up.

Well, not that odd. These rankings are implemented in a fairly drifty way, with the input score for a race being the one throws out your bad runs. Since most of the main group has extra runs, they creep up each iteration as they keep having only to top 60-80% of their scores count. But the Canadians in the bottom group don't have any races to throw out, so they don't drift much - and with normalization that pushes them down. Still above zero because some interaction has kept them in place.

It's a good thing they don't have too many races or they might drift up faster than the rest of us in the iterations, and the rest of us would fall below the Canadian group. (That might have actually happened at one point a few years ago.)

The ridiculously low rankings for Western Canadians probably would clear up if Louise and Marie-Cat would or could enter all their splits. I say "could" since when I check the US Champs results I see that Cristina is the only Red X runner to have entered her splits - even if Louise, Marie-Cat and at least one other F21 entered their splits for Day 1, it might have the effect of dragging down Cristina's ranking rather than significantly boosting theirs. Perhaps the new ability to upload splits for an entire race will lower the effort required enough to solve that problem. It may also help that in the last proposal I saw, the US Team Trials were to be one of the WOC selection opportunities for Canadians next year.

Adrian, on the other hand, could enter his US Champs splits on Red Y and get two reasonable ranking numbers anytime he chose.

Jon, when was that proposal to have US Trials as a selection opportunity?

It was in a proposal Ted circulated for comment via Osquad on October 5. Now that I look at the message, he addressed it to you separately at a Hotmail address as well, which must have had some problem. I've just forwarded you a copy at the address associated with your AttackPoint moniker.

I have no idea how the AP ranking system works (is there any more explanation than "a variation of the USOF formula"?), I just assumed it was some kind of mystical formula that had nothing to do with how well I *thought* I had run. But, as for the Canadians and the US Champs, there are at least 3 other ranked American AP users who ran Red X and haven't entered their splits, either. So yeah, I think the split upload feature would probably help the problem quite a bit.