|
Dec 2017
10:28am, 7 Dec 2017
6,135 posts
|
The_Saint
That isn't a bad explanation but given a close look at the layout (flat yes but no flatter than many and with 180 degree turns), it would be interesting to see what a Jones Counter made of it.
Cardiff Blackweir for example is this flat and with one bend that maybe costs a second being generous
|
|
Dec 2017
10:32am, 7 Dec 2017
62,216 posts
|
Gobi
Lol - that bend is terrible 
|
|
Dec 2017
10:58am, 7 Dec 2017
3,498 posts
|
larkim
Spideog - I'm not sure that's correct. Ignoring the fact that I think the SSS score wouldn't be calculated with only 5 runners (see below), if their individual performances compared to their own personal standards were considerably quicker the overall average would be a strong negative (e.g. -1.0)
With 100 runners getting only slight PBs, the overall average would only be a slight negative (e.g -0.2)
I think from what Tim G says he tries to get good representative data, so for example SSS scores are not derived from runners who only run parkruns - there needs to be a deeper data set for a runner to be included in the scoring of the race/parkrun itself. But if the runners are in scope for the calculation, and they all show a consistent pattern, the size of that population of runners doesn't matter.
Of course, in the real world the size of the populations does matter as too few runners in the calc will tend to skew things through the low likelihood of them being a representative sample, and that's why he only runs the calc once some minimum thresholds of volumes of runners is established. If there are too few, a score of 1.0 is recorded instead.
|
|
Dec 2017
11:18am, 7 Dec 2017
14,901 posts
|
Dvorak
I think some people are confusing the two parts of the run distance rankings. As I understand it, the SSS scores are derived for each event individually (and by event I do not mean course, each running of the course is considered individually; so there is not a rating for XYZ parkrun which is then varied week to week, there is a rating calculated from scratch each week for XYZ parkrun.) And for this rating it does not matter who ran it, just that some group of people ran it.
Derived from this is a vSSS score for each runner, which then feeds into their individual handicap. However, what any individuals vSSS score would be DOES NOT influence the SSS score of an event.
My vSSS score (except for parkruns, which are allocated) only appears when I log in and claim a performance. I'm not sure if this is the same for everyone or if Power of 10 athletes have it allocated automatically. (I can't have a look just now, RBR is not bringing up race results.) If the calculation of SSS scores relied on vSSS scores, events' scores would waver up and down as more people logged in and had those calculated.
|
|
Dec 2017
11:24am, 7 Dec 2017
55 posts
|
Spideog
I've not paid a lot of attention to comparing the SSS scores on my runs recently as pushing the buggy adds too many other variables, and will also be screwing up the score for other people as my score won't be an accurate representation of the course or effort put in.
But I do recall seeing wildly different scores for myself for virtually identical times on the same course back in the day, the only difference being how many faster people happened to turn up that day for the event. On other weeks where there wasn't an unusual number of faster runners turn up the score did tend to adjust accurately in terms of the weather that day potentially effecting other peoples times and you could see that from week to week.
There are a few instances in the list where courses I've run are definitely out of order in terms of difficulty, the main difference being the standard of the local runners, and the number of times they have run there.
But I think we are in agreement that it's interesting, but not perfect.
|
|
Dec 2017
11:25am, 7 Dec 2017
6,782 posts
|
Ceratonia
It automatically allocates the performance to you if it can be sure it is you (for example, if you have a unique name and belong to a club.) For parkrun, it can always be sure it is you. Otherwise, you have to claim the performance.
|
|
Dec 2017
11:28am, 7 Dec 2017
6,783 posts
|
Ceratonia
Spideog- it's not the faster people turning up that changes the score, it's the faster people turning up and running fast. If they all have a big race the next day and jog round parkrun slowly, it will appear as if the course was particularly difficult that day (fast runners got slower times than normal) and your runbritain ranking would improve if you had done a good time on that day.
|
|
Dec 2017
11:30am, 7 Dec 2017
14,902 posts
|
Dvorak
As for Worcester, it looks an excellent course for a fast time. Flat, mainly tarmac (some gravel and people maybe running on grass beside this. Mostly straight or gently curving, with much of ir sheltered. From a map measurement on Fetch, compared to the course on the website, it is 5km. However ...
There are three 180 degree turns. Apart from slowing runners, it means that these have to be set in the correct place. There is a course download on Fetch from sheppardmk. This shows 5 km, but has a bit of wiggliness. Looking closer (and assuming his gps was correct at these points), the turns are both a bit inside where they should be, suggesting the course when he ran it was at least 100m short, up to 140m.
The indication therefore is that Worcester's rating may be due to being a potentially very fast course, which ALSO is being run short.
(I don't like to mention St*v*, but maybe someone who is on it could see what other runners have run at Worcester, specifically looking at the turns?)
|
|
Dec 2017
11:46am, 7 Dec 2017
56 posts
|
Spideog
Yep, the faster people turning up and running fast, and having done so more often over more years, will have an effect on the scores for the more established events in big cities.
|
|
Dec 2017
12:48pm, 7 Dec 2017
First-time poster!!
|
timgrose
Good to see you are talking about my little list!
The Worcester Pitchcroft one is puzzling. My hunch looking at strava and some of the leading times on it that the course is at least 100 metres short but is otherwise fast anyway. Have done enough of these investigations into possible short road races before to know that when nearly all GPS traces come up short it is almost certainly is. The course record holder, for instance, seemed to run faster on this course than he did in a track 5000 in the same timeframe.
As for the other points, look like got the general idea. For clarity, this for "average" conditions so if it is fast in the summer (e.g. firm grass) but slow in the winter (a quagmire! - think Mulbarton is one such) then it is going to get bumped down.
Yes there will be some effect if everybody does not try hard but the scores are derived from all runnings of each parkrun so such "one off" occurrences should have not much if any effect in a long term average.
For 5K at about 20 mins one "point" is more or less 30 secs. As such Worcester Pitchcroft aside you are likely to run at least 30 secs quicker on a good day on the track. That certainly was the case for me when was in good form at same time (a fading memory!). 16:24 on the track and 17:01 in a parkrun and I had several other track and parkruns just a tad slower.
|
