Another negative topic!

[andypat]

I do like the current ranking system. I would have been in favour of weighting bigger events but can see why it hasnt been done now. I'm currently concerned that despite my gradual increase in fitness following a period of injury last year, my ranking points acheived are going down! I'm hoping this is due to something mathematical and not my own self delusion that i am getting a bit fitter.

[229]

.

[IanD]

Also part of the idea behind including more events in the rankings was to get BOF members who don't travel much into their appropriate position in the ranking list. Almost everybody should be able to get to six ranking events per year. Many people don't get to six 'major' events. If you weight the big events, these people will once again be under-ranked.

Not saying that this is necessarily a show-stopper for weighting, just pointing out that weighting would mean abandoning one of the objectives of the present unweighted method.

[awk]

Me too. I'm making less mistakes, getting higher placings and yet lower scores.

But does this really matter if it is happening to everyone. You will still stay in the same place in the ranking list as it will never reach the point where we all score 1,000 points at every event.

Yes it does matter if the ranking list is to provide a current ranking list as opposed to an historic one. An improving orienteer getting better as the number of ranking points s/he can gain declines, is trying to move up a list and get in front of people who have benefited from 'inflated' scores, and will continue to benefit for 12 months afterwards. The improver will only overtake them once their scores are dropped due to time - but that's heavily delayed, by which time our improver is scoring even fewer ranking points, whilst the incumbents continue to benefit from comparatively inflated scores. There will be a crossover, but only a long time later.

And, as I said in my previous post, I think there is a north-south split too, where it's easier to score more highly in the north than in the south.

[AlanB]

I do like the current ranking system. I would have been in favour of weighting bigger events but can see why it hasnt been done now. I'm currently concerned that despite my gradual increase in fitness following a period of injury last year, my ranking points acheived are going down! I'm hoping this is due to something mathematical and not my own self delusion that i am getting a bit fitter.

Yes, I have a similar experience. I'm sure I'm fitter and better than I was last (I should bloomin' hope so!) yet I've struggled to match last years points.

Except that I scored much better in the South of Scotland O League events last winter (in which I enter Green) than I tend to do in age-related events (short blue).

[buzz]

If you honestly think there's nothing arbitrary about weighting, then you'll be able tell us what the non-arbitrary weighting factor is, and why. Wont you?

Sorry I misread gnitworps comment - I thought he meant choosing to weight level A events was arbitrary.

Of course the weighting itself would be arbitrary but there are plenty of arbitrary weightings in mathematical models.

The anecdotal evidence seems to suggest that running smaller events, shorter courses and sprint/urban events seem offer better chances of scoring than running your age class in championship events. It would be very easy for the so called mathematical purists to test this by analysing past data and if this is the case perhaps they can explain why and proposed a solution (perhaps by adjusting weightings until the anomoly disappears ?)

[greywolf]

I thought he meant choosing to weight level A events was arbitrary.

So you think weighting a closed event (CSC Final) and the Midlands Champs in a year when it's a poorly attended TD4 bramble fest, but not weighting the WOC selection races (usually best quality elite field of the year) and the Scottish 6 Days (best and biggest fields for other classes) isn't arbitrary?

The anecdotal evidence seems to suggest that running smaller events, shorter courses and sprint/urban events seem offer better chances of scoring than running your age class in championship events.

You're conflating two (alleged) issues: whether at any given event points are easier to come by on shorter courses; and whether small / sprint / urban events are easier for points scoring.

The former has already been acknowledged as an effect of the initial points distribution which was based on mins/km and had the effect of, amongst other things, not putting the very best sufficiently far ahead of the mass. As has been noted here, many of those who were initially over-rated are now finding their scores gradually decline as this effect is eroded over time.

The latter is harder to quantify: perhaps you haven't worked it out, but the ranking system is basically zero-sum - at any course at any event the mean of the points scored is the mean of the points of ranked runners before the race. Crudely (becuase the bottom 10% rounded up are trimmed as outliers), if there are 10 competitors on a course with an average points score of 1000 before the event, then the mean of the points awarded at that event will be 1000. What the rankings system does is allocate those 10,000 points according to an algorithm which attempts to compare the actual spread of times against the expected spread of times (based on runners previous mean scores).

So it's hard to see how one kind of event could offer better chances of scoring than another... with two caveats: 1) obviously the system is less robust with smaller fields, so there is more chance of freak high (and correspondingly low) scores in smaller fields - but some courses in Level A events have small fields too 2) there may be some small effect as a result of the inclusion of runners who have not competed in a ranking event in the last 12 months, and who finish high enough to be trimmed as outliers: this probably happens more often in small events, but could equally occur at BOC or the JK.

It's difficult to see how weighting Level A events (or a subset of them) would address any of these effects - as has been stated many times previously it would simply reduce the list to a "how many level As have you been to exercise", because whilst what you think you want to do is give more points to the winner of BOC and the JK, what you will actually do under almost all weighting scenarios is give more points to all the people who had crap runs at the Midlands Champs, so that turning up at a Level A is worth more than a great run somewhere else.

PS If you were insistent on weighting then the most reasonable weighting scenario was suggested by Graeme (I think) - that scores from e.g. British Champs, JK, etc, could be counted twice in your best 6 - this would only benefit those who for whom these events were already amongst their best scores

[Ant W]

Not sure if this is relevant, but is the mean score per orienteer calculated on top 6 results or all results in the previous 12 months?

[buzz]

So if I understand you correctly Greywolf you seem to be suggesting that its OK to score more points by running shorter courses, at smaller events peferably sprint or urban races?

If not perhaps you could suggest a solution?

[greywolf]

So if I understand you correctly Greywolf you seem to be suggesting that its OK to score more points by running shorter courses, at smaller events peferably sprint or urban races? If not perhaps you could suggest a solution?

No ffs, evidently you don't understand

I'm suggesting that the "running shorter courses" or more precisely "running against competitors who may have been initially been over-ranked" effect is real but diminshing anyway (and is at least part of the reason why people on here are moaning about not scoring as highly as before)

And that I can't see any reason why the sprint /urban races (or any other type of race) would generate more points for the whole field than any other type of race. There are some anecdotes based on a few individuals scores, often at the front of the field. Until someone shows definitively that this effect is real then I don't see a need for a solution. Did you read the previous post?

Results from small fields are always going to be less reliable than results from large fields. The simple answer is to up the threshold to 20 ranked runners. Or how about 50? But the higher the threshold, the fewer the races that qualify, so fewer folk have qualifying events...so you introduce more unreliablity. 10 is a compromise. Feel free to argue for a different threshold.

Ant W: all (non-zero) results in the previous 12 months

[pete.owens]

There is another, more straight-forward, explanation why some orienteers will score more points at urban or sprint O events:

That their particular skills are suited to those events.
A strong runner will tend to score well on a long physically demanding course on deep heather moorland. One adept at contour interpretation will do well in sandunes. Someone who's strength is in fine navigation will do well on a middle distance course in a technical pathless forest.

Similarly someone who can run fast on smooth surfaces, read a map at speed and make good route choice decisions quickly will tend to excel on urban courses.

[maprun]

How about showing the median (middlemost ranked score)ranking score of each runner alongside their 6 best scores. There are plenty of people whose ranking benefits from running dozens of ranking events a year, but their typical score would place them lower. It avoids the luck that comes with small fields and poor planning. Sitting alongside the 6 best ranked scores you'd at least see the best scores and the typical scores of runners.

[AndyO]

Alternatively - just ignore the ranking lists. Don't know what purpose they serve anyway

[madmike]

Alternatively - just ignore the ranking lists. Don't know what purpose they serve anyway

goal setting

[awk]

So if I understand you correctly Greywolf you seem to be suggesting that its OK to score more points by running shorter courses, at smaller events peferably sprint or urban races?

Goes to show how much of this discussion (including my own contributions) are based on anecdotal evidence. You have several times cited sprint/urban events as easier to get points at. Personally, I think the opposite.