Presenting separate terrain & urban ranking lists dodes not require the removal of the current combined one.
The current combined should remain the master, but separately separate lists should be presented for terrain and urban. providing events are categorised correctly it's just another filter, like age class. In urban my ranking will rise, in terrain it will drop.
Ranking Scheme bias?
Moderators: [nope] cartel, team nopesport
Re: Ranking Scheme bias?
curro ergo sum
-
King Penguin - addict
- Posts: 1453
- Joined: Wed Sep 21, 2005 6:56 pm
- Location: Kendal
Re: Ranking Scheme bias?
It is rather more complicated than that.
To produce separate urban and forest lists the base score for each course would need to be calculated based on the previous scores of the entrants at events of that discipline.
To produce separate urban and forest lists the base score for each course would need to be calculated based on the previous scores of the entrants at events of that discipline.
- pete.owens
- diehard
- Posts: 752
- Joined: Tue Jan 15, 2008 12:25 am
Re: Ranking Scheme bias?
Reading with interest as the other half jogged around short brown yesterday making mistake after mistake, and I had a near perfect run on the green women, running apace that had I run it on short brown would have beaten him by 8 minutes (and earned me 60 more points), and yet he got 10 points more than me.
- housewife
- green
- Posts: 355
- Joined: Tue Nov 16, 2004 10:28 pm
- Location: probably at work
Re: Ranking Scheme bias?
pete.owens wrote:Which is exactly what you should expect.
If you had run green you should have been capable of running harder than you maintained over a brown course of twice the distance. So, the same m/km ought to score more ranking points over a longer course.
Well, the same still applied at my most recent experience of sprint, where I don't find the extra distance slows one down (my training times show that I run pretty much the same speed at these sorts of distances). I scored my lowest ranking points in several years at Cambridge on Saturday running the D course, MUV/WSV (with, admittedly, not a great run, but nowhere near as bad as the points might suggest), where I scored 1118. At the same m/km on C, I'd have scored 1147. Similarly, I'd have scored 1144 on the B course, and 1163 on A (comparison with previous races tells me that 1144-1167 is much more the range I'd have expected to score for a run like this).
It just seems to me that, time after time, as soon as I run a course with 'just' older people in it, my scores go down - it's been really marked as I've moved into MUV. As soon as I run up, my scores go up. Chatting in the car park, I think this is a widely held view - with several comments as blunt as 'the ranking scheme is not fit for purpose', and 'It's nonsense'. If this bias is true, it also suggests that it's easier to get ranking points at the smaller events, and harder at the bigger events, which is completely the opposite of what should be happening, if at all.
I don't have access to the data or the means, but I'd be really interested to see some analysis on this - taking a number of events where mixed ages have run a course (eg Black at middle-distance, when a lot of older competitors run up), and then looking at what happens when one includes and excludes younger runners' times in the calculations.
Last edited by awk on Thu Feb 23, 2023 9:00 pm, edited 1 time in total.
-
awk - god
- Posts: 3224
- Joined: Sun Mar 21, 2004 5:29 pm
- Location: Bradford
Re: Ranking Scheme bias?
Presenting separate terrain & urban ranking lists dodes not require the removal of the current combined one.
The current combined should remain the master, but separately separate lists should be presented for terrain and urban. providing events are categorised correctly it's just another filter, like age class. In urban my ranking will rise, in terrain it will drop.
Is that your best 6 urban events and your best 6 forest events and then your best 6 events overall? its no longer just a filter; many people further down the table won't have 6 of each to generate good rankings with.
- Atomic
- orange
- Posts: 140
- Joined: Thu Nov 25, 2021 11:56 am
Re: Ranking Scheme bias?
I'm with pete.owens on this. Most people (except perhaps awk! ) can't complete a Black course at the same mins/km as they can a Green course. If they could it would be reasonable to get more points. The difference in achievable mins/km will be less for 'adjacent' courses, but it must surely still exist.
So saying you could get more points for the same mins/km on a longer course is true, but is not an indication of bias.
How many people can run a marathon / 10K in no more than twice their half / 5K times?
So saying you could get more points for the same mins/km on a longer course is true, but is not an indication of bias.
How many people can run a marathon / 10K in no more than twice their half / 5K times?
- Snail
- diehard
- Posts: 709
- Joined: Wed Nov 30, 2005 8:37 pm
Re: Ranking Scheme bias?
I don't disagree with you over longer distances, to a lesser extent middle distance too, but we're talking sprint here. That's not the difference between running 5 and 10k but, in the examples I gave here, the difference between 2.5k and 3.1/4.1k actually run. I often find over these sorts of distances I am, if anything, slightly faster over the longer distance as mistakes cost proportionally less. (The mins/km difference between world records for M65 at 3000m and 5000m is 0.2 seconds).
As I said, experience tells me, and many others, that to score more points you need to run up, or, at least, against younger (faster) competition. Running a separated off MUV/WSV or equivalent means zilch points. What I'd like to see is some sort of hard, analytical evidence to show whether this is right or wrong. The most convincing argument I've read so far is of the skewed normal curve and where you sit on that.
As I said, experience tells me, and many others, that to score more points you need to run up, or, at least, against younger (faster) competition. Running a separated off MUV/WSV or equivalent means zilch points. What I'd like to see is some sort of hard, analytical evidence to show whether this is right or wrong. The most convincing argument I've read so far is of the skewed normal curve and where you sit on that.
-
awk - god
- Posts: 3224
- Joined: Sun Mar 21, 2004 5:29 pm
- Location: Bradford
Re: Ranking Scheme bias?
Snail wrote:So saying you could get more points for the same mins/km on a longer course is true, but is not an indication of bias.
Not always, but when e.g. at urban races a top M21 doing the M/WHV course for sprint practice would have to run 1min/km to score their usual 1400 points it suggests there might be something amiss...
-
greywolf - addict
- Posts: 1416
- Joined: Sat Apr 07, 2007 12:45 pm
- Location: far far away
Re: Ranking Scheme bias?
Snail wrote: The difference in achievable mins/km will be less for 'adjacent' courses, but it must surely still exist.
Some of the differences in the older course lengths at events seem negligible from that perspective, but I'm too young to comment. Does an extra 500m make so much difference?
(I feel for the M75s at this weekends Northern champs who get to run 200m less than M70s but with 40m more climb!)
Snail wrote:
How many people can run a marathon / 10K in no more than twice their half / 5K times?
I have a lot of respect for those who can run through a rough forest or over tussocky moorland at race pace, but I certainly can't. I wouldn't be orienteering much faster at all between 5k and 10k over that terrain. I'd have to be running over an hour for the distance to impact on pace usually.
- Len
- white
- Posts: 78
- Joined: Wed May 16, 2012 5:19 pm
Re: Ranking Scheme bias?
greywolf wrote:Not always, but when e.g. at urban races a top M21 doing the M/WHV course for sprint practice would have to run 1min/km to score their usual 1400 points it suggests there might be something amiss...
One should not expect the ranking system to work well for such extreme cases. There is not enough data to fit the model, and even if it could be fitted over such a broad range of conditions, it would necessarily be more complicated, with many seemingly arbitrary parameters, and people would be unhappy with it for that reason.
- MChub
- off string
- Posts: 36
- Joined: Mon Oct 08, 2018 7:43 pm
Re: Ranking Scheme bias?
Many years ago, when I was competing in traditional forest orienteering, I produced a spreadsheet for performance analysis that worked on the premise that, excluding mistakes, the time you took for each leg was made up of running time (proportional to distance) plus a fixed amount for punching, planning the next leg etc, and this seemed to work pretty well. (I believe I used straight-line distance, which made it pretty useless for urban events.) Could this extra time per control be a factor that's affecting things when comparing different course lengths? I'm sure I could run a course with 20 controls faster than one of the same length and difficulty with 40.
- roadrunner
- addict
- Posts: 1059
- Joined: Fri Sep 02, 2005 8:30 pm
Re: Ranking Scheme bias?
awk wrote:I don't disagree with you over longer distances, to a lesser extent middle distance too, but we're talking sprint here. That's not the difference between running 5 and 10k but, in the examples I gave here, the difference between 2.5k and 3.1/4.1k actually run. I often find over these sorts of distances I am, if anything, slightly faster over the longer distance as mistakes cost proportionally less. (The mins/km difference between world records for M65 at 3000m and 5000m is 0.2 seconds).
As I said, experience tells me, and many others, that to score more points you need to run up, or, at least, against younger (faster) competition. Running a separated off MUV/WSV or equivalent means zilch points. What I'd like to see is some sort of hard, analytical evidence to show whether this is right or wrong. The most convincing argument I've read so far is of the skewed normal curve and where you sit on that.
Out of curiosity, I tested this theory on the recent Cambridge sprint event (just the prologue, as that's what's on the BO site). For a fast runner (10:00 on the D course), indeed the longer the course, the higher the score. But for someone who would complete the D course in 15 minutes the order (highest score to lowest) is A, C, D, B, E, while for someone who would take 20 minutes on the D course, it would be D, E, C, A, B - so not straightforward!
(If you're wondering why I chose the D course time as a reference, it's because that's the course I ran.)
- roadrunner
- addict
- Posts: 1059
- Joined: Fri Sep 02, 2005 8:30 pm
Re: Ranking Scheme bias?
So does the formula for the ranking points use mins per km or are we just using that as a metric to compare performance with ranking score?
I ask because yesterday's event at Burbage Moor would be an example of where the comparison could not be made across courses: some visited the deep heather in the north east, others did not, this must have had a significant impact on mins pkm.
I ask because yesterday's event at Burbage Moor would be an example of where the comparison could not be made across courses: some visited the deep heather in the north east, others did not, this must have had a significant impact on mins pkm.
- Parkino
- red
- Posts: 176
- Joined: Fri Mar 30, 2012 9:37 am
Re: Ranking Scheme bias?
Parkino wrote:So does the formula for the ranking points use mins per km or are we just using that as a metric to compare performance with ranking score?
I ask because yesterday's event at Burbage Moor would be an example of where the comparison could not be made across courses: some visited the deep heather in the north east, others did not, this must have had a significant impact on mins pkm.
No - the scheme uses running times and current ranking points, scaled for both average time and spread of times so as to try to compensate for different courses. Each course is treated separately, so the fact that some have easier terrain won't affect it.
- roadrunner
- addict
- Posts: 1059
- Joined: Fri Sep 02, 2005 8:30 pm
Who is online
Users browsing this forum: Google [Bot] and 16 guests