2012 Urban Speculation

[swat]

Would it be feasible (and fairer) to score based on rankings points

It probably would be fairer, but there are a couple of problems: juniors don't score ranking points, and chasing sprint/mass start type races are excluded from the ranking scheme (also night events - so I guess night urban events probably are too). Otherwise it's a great idea - I had the same plan to use ranking points for our club league but got caught out by juniors not being included

[DJM]

Saturday 28th July Aberystwyth

Not quite sure of the rules but, as this race is structured as an age class competition, with L and S, A and B classes on multiple courses, I'm guessing it won't fit in to the nopesport series ... It will be a UK Masters Cup race however.

Mind you, having test run one of the courses yesterday on behalf of Croeso 2012, I can promise a brilliant day

[distracted]

Couple more provisional ones for the list:

Oct 20th - Oxford
Nov 24th - Sheffield

[Jayne]

but there are a couple of problems: juniors don't score ranking points, and chasing sprint/mass start type races are excluded from the ranking scheme (also night events - so I guess night urban events probably are too):(

I hadn't realised the juniors didn't get ranking points. There's not a lot I can think of doing to get over that. Although, i guess it would be possible to score juniors using the nopesport method and everyone else using the ranking points. Another way would be to apply a weighting to take in the number of people running but that might just over-complicate things.

Re. night/mass start. Although these don't count for running points, would it be possible to run the ranking point model on them to get the points - does anyone know who does the model - they wouldn't need to be uploaded to the website but we could possibly get "pseudo"-rankings to use for a league.

(or I could stop making things complicated - just thinking aloud!)

[DJM]

Appendix K: the ranking Scheme:-

1.2.3 All competitors [who are members of British Orienteering] in the top year of classes M/W16, and in all classes M/W18 or older, at Ranking events shall score Ranking points.

[awk]

(or I could stop making things complicated - just thinking aloud!)

It might be 'fairer' (whatever that means) to use ranking points, but I'm no fan of them for use in a head to head racing league. Using a simple points system as used by Nopesport, Yorkshire, and the UK Masters, means that people are encouraged to attend a wider range, and puts more emphasis on the fact that head to head competition means being there to ensure you get the points and your rivals don't. Using rankings points makes it more a remote mathematical exercise where you don't learn how you did until sometimes long after the event, and less about the racing that actually happens on the day.

(This view may be somewhat biased by the fact that what interest I had in the rankings has waned the more I see that it doesn't reflect my personal perception of the relative merits of my own results)

[graeme]

would it be possible to run the ranking point model on them to get the points - does anyone know who does the model

I've written twice to BOF to ask for the code, but to no avail. Let me know if you have any more luck.

the more I see that (the ranking list) doesn't reflect my personal perception of the relative merits of my own results

There's certainly some problem at the margins, (see above), but it affects scores only at the level of a few percent. By contrast my NUL and Masters Cup points bear no resemblance at all to "the relative merits of my own results" and everything to "whether there was a big field or not".

Take a look a Masters' Cup M35 and assume the scores are 7/16 ignoring the last three races. You'll see I was ahead of Geoff Ellis (hurrah!) despite the fact that he beat me on all but one occasion . My lead is due to:
1/ Getting big points for beating nobody much,
2/ Lots of people who aren't me beating Geoff at the big races.
Now of course, the system is what it is, but any idea I would have been the "right" winner is a bit weird.

In this case the best man won in the end, but there are similar examples based on full final results, I just can't be bothered digging for them.

[IanD]

I think it depends on whether you're trying to score the league to produce the 'right' winner, or to produce consistent scores for a given level of performance further down the field.

If you want to score for the 'right' winner of the league, my personal preference is for the present system, whereby the winner of an event always get the same score. A nice simple scoring system, and the points are known immediately the event finishes.

If you want people in mid-table to get roughly the same score for a given level of performance, then points based on placings don't work, since the size of the field matters so much. A ranking-list approach would avoid this problem.

A side effect of the present system may be to encourage people to travel to what they think will be the less popular events. Is this a bad thing?

My vote goes for keeping things as they are.

[King Penguin]

In the NOPE Urban league standings there are plenty of people below me who beat me every time (MVet). My over-elevated position is due entirely to them not having 6 results, and with just one event to go it's too late for them to overtake me.
Rather disappointing that only 4 runners in the class have 6 results, and no-one else can now get more than 5.

[andypat]

Speaking as a keen Scottish Urbanite, you do need to be quite keen (and prepared to travel far and expensively) to get scoring 6 events. I know there will be some who have maanged all but it is a bit hard with all other priorities. In some ways thats a good thing as it does give those of us who arent likely to trouble the podium an incentive to do at least 6 runs. I am going to manage 5 (assuming a clean run at Penrith), but then I was responsible for one of the Scottish ones so lost that opportunity!

I think you should jyst bask in your current glory KP - in years to come there may be more challenge

[awk]

Now of course, the system is what it is, but any idea I would have been the "right" winner is a bit weird.
In this case the best man won in the end, but there are similar examples based on full final results, I just can't be bothered digging for them.

Yes, I've been at the other end of that recently, where I've beaten people way more often than they've beaten me, but they've been ahead because they got to 'easier' scoring events. Doesn't bother me - that's the name of the game. I still prefer the simplicity and the emphasis on the head to head racing, which ranking points completely misses out. If I decide not to go to the less popular events (which, for various reasons I did), that's my problem - it's my choice.

On that note King Penguin, I agree with andypat: you've got to be there to win.

[andypat]

....then I'm willing to set up the scoring. This goes for as many regions (within reason!) as wish to do it, so we could have South and Scotland as well as Yorkshire.

Roger - I have pm'd you a query about scoring for a possible Scottish proposal.

[Roger]

From Caz's list - potential Southern League events...

Saturday 21st April Huntingdon
Sunday 24th June Canterbury
Saturday 25th August Guildford
Sunday 26th August Basingstoke
Monday 27th August Didcot
Saturday 22nd September London
Sunday 30th September Windsor & Eton

...

Questions:
1) Is it an issue that three of them are over the same weekend (it would probably be best 4 races out of 7).
2) Would it be feasible (and fairer) to score based on rankings points. Looking at my own scores in the Yorkshire/Nopesport leagues the races I did best in were the ones with fewest competitors (kind of obvious) but the respective ranking points I got seem to remove this bias.

From the races I removed I'd say there could be a Midlands league too!

London is unlikely to qualify because it's too big to use the standard Urban League categories. Having vastly more competitors than the other races also distorts the scoring, as you've realised.

Re scoring, I reckon that the idea is to encourage competition: to provide an extra reason to go to a particular event and give some head-to-head competition. I had an excellent battle with Spookster two years ago, with each of us knowing what was needed in the final race. (He won.) As a side-effect the current system may well increase participation, especially in events that may otherwise not attract too many competitors -- which I think is a laudable goal in its own right. As others have said, the current system gives more exciting (although less precise) results. My vote would be to keep it.

(This is not just because I instigated the present system. At inception, using the ranking points wasn't an option and spreadsheet calculations were the order of the day. The publishing of the ranking points opens up the alternative possibility and if I thought it was better then I'd advocate it as an improvement.)

And I'd say include the Midlands events too, to give a wider range of possibilities. Scoring 4/7 (or 3/6) is challenging if you miss three because you're a White Rose fan.

[Jayne]

Interesting discussion, and it looks like the conclusion is to use the Nopesport scoring, with thanks to Roger for offering his services. I'll post another suggested "Southern" list when more races are confirmed - and it looks like I'll take South as meaning pretty much anything below Derby.

Caz, thanks for the offer of help. Send me an email (jayne.sales (@) gmail.com) and we can discuss.

[graeme]

I've been wondering about a different (even less fair?) simple system.

Double points for the Final race.

This would have a number of effects. Primarily, it would be essentially impossible to win without going to the final, which would gave that race some kudos. It would therefore keep the competition open to the end - one weakness with the Masters Cup was that by the last few races many of the classes were decided, and there was no focus to the competition.

When you look at the maths, for two people who are quite close the final race doesn't have such a big effect on relative position: for competitiveness in the competition as a whole the last race is everything, but for determining who beats who it only has double the importance of the other 8 races.