Rankings for the new 4 tier era

[Clive Coles]

The ranking tables currently suggest that about half of the ranked orienteers fail to attend the 6 events to allow them to be ranked on an equivalent basis to those more regular orienteers travel to find Ranking events outside their region.

With the introduction of 4-tiers, level C events will become ranking events. It should therefore be far easier for local orienteers to attend sufficient events in the future. This surely is a good move and by making the lists more inclusive rankings will become more interesting as a result.

But it would seem to me to be easier to earn ranking points at a Level C event where the quality of the field is drawn from regular Sunday morning orienteers than would be the case if points were earned when competing against a Championship or Level B field. Now we will have so many level C events being ranked this will surely distort the rankings more in favour of local orienteers. Is this what we want from a National ranking scheme ?

We used to have loading with the long forgotten Badge ranking system. is it now time to consider re-introducing the concept ?

For starter how about

Level A x1.2
Level B x1.1
Level C x1.0

I have forgotten the arguments that were raised against this idea in the past. (I am sure someone will soon re-educate me).

[EddieH]

I don't wish to repeat the arguments, save that Graeme convinced me that weighting is wrong.

There is no reason why a lesser field should increase your ranking points.

There are only 2 problems that I see with loads of localish events.

1 is that some poor navigators that are very familiar with their local patch may be able to get 6 big scores purely through running speed AND local knowledge.

2 is that if the majority of people stick to their local patch any current local anomolies (which inevitably are still there after the very poor seed data that set the whole thing up) will simply be maintained.

In the end it is all a bit of fun, and from that poinjt of view the more events the merrier.

[SeanC]

If the level of competition influences the scores*, then there could me more variations between events at the same level, than between events of different levels in the same region.

Which reminded me about our level B event just after Christmas, DFOK Westerham 9th Jan. Just checked and the closing date is tommorrow. Currently it looks like the queue for EOD will be massive, so I advise you to get your entry in soon or you could end up with the choice of Black or Short Green on the day.

*I'm not saying it does, I don't understand the algorithm either

[Clive Coles]

Quite agree with you EddieH

In the end it is all a bit of fun, and from that point of view the more events the merrier.

Currently I seem to be on page 174 ~ I go up and down the rankings without even attending a ranking event since last February !

But I guess it's an interesting benchmark and some folk obviously care about their ranking if posts in the other ranking thread are to be considered.

Perhaps Graeme's weighting arguments are still valid but with the inclusion of level C events the balance of qualifying scores will change.

[andypat]

But it would seem to me to be easier to earn ranking points at a Level C event where the quality of the field is drawn from regular Sunday morning orienteers than would be the case if points were earned when competing against a Championship or Level B field. Now we will have so many level C events being ranked this will surely distort the rankings more in favour of local orienteers. Is this what we want from a National ranking scheme ?

Likewise I dont fully understand the algorithm, but isnt it possible that the opposite is also true?
If you win a local (C) event at which only lower ranked runners turn up, isnt it less likely that you will get mega ranking points?

[greywolf]

Clive, you are labouring under a number of misconceptions:

1) the 4 tier system shouldn't in theory have any impact on which events are eligible for ranking*... the problem to date is that clubs (or is it regional associations?) haven't been registering events as intended... perhaps they will get it right in future, which may lead to an increase in rankings events. There are some possible minor disadvantages - Eddie mentions a couple and no doubt someone else will very soon be banging on about how they only run hard in major events (which may be a factor for some very good orienteers) - but they are IMHO outweighed by the gains from increasing involvement etc, and certainly a massive improvement cf all the ridiculous anomalies in the old system

2) the points scored by any individual competitor are dependent on a) their performance relative to the field of competitors**, and b) the strength of that field - so you don't automatically score more points running against a weaker field ***
(and yes, there may be some problems with the seed data which causes a similar effect...but this is the case in Level A & B events as well)

3) introducing weighting has been debated at length here - basically it just skews the entire list in favour of those who attend weighted events - regardless of the quality of performance

* it was always the intention that what used to be "district" events, plus Nopesport league races, etc would be ranking.

** strictly speaking, this = 16yo+ BOF members who complete the course

*** to all intents and purposes this is calculated by the same algorithm as used in the old ranking system - the main significant change is that the current system does one calculation for everyone who runs the same course, whilst the old system used to pretend that individuals in different age classes / genders were running different courses and make separate calculations for each subset.

[Clive Coles]

Not actually intending to promote a move to weighting ranking points Greywolf. Rather raising the issue as I think the introduction of 4-tiers does "in practice" change the volume and spread of qualifying ranking events.

You are quite correct, clubs could have upgraded their old Sunday morning District events to level 2 under the old 3-tier system. A lot of clubs didn't for a number of reasons ( no need to go back and redebate the merits of that decision). I believe however many clubs will now upgade these full colour code events into level C. This is IMO not before time because they are fixtures that often attract as many orienteers as a level B event.

So... with the dawn of the new era I thought it interesting to see if folk were content with the potential consequences of this evolution.

[seabird]

Clive, I suspect you will see few, if any, significant changes.

Round here we have been using what will be level C events for ranking purposes over the past year.

The runners in our club nearest the top of the rankings tend to have accrued the majority of their points at what will be level A or B events.

Lower down that is less consistent. Some runners have accrued most of their points at Level C events; others at Level A & B. I have accrued all of my points at what will be either Level A/B cross country or Level C urban races.

I do not see any runners accruing more points as a general rule at Level C events than they do at the higher levels, if anythng the reverse is true. And no runners focussing primarily on events within the region have acquired an unexpectedly high place in the rankings relative to their fellow club members who travel further afield.

[GML]

What the ranking formula doesn't reflect of course is that some (higher ranking) competitors may not be competing as hard at a local event as they would at a higher level event - they may be using the local event as training. If this happens then it will tend to inflate the scores of other (lower ranking) competitors. May be a case for higher ranking competitors to declare themselves non-competitive (or deliberately mis-punch?) if they are only using an event as training rather than competing at full intensity?

[Gnitworp]

The greatest variable, that disrupts the theoretical purity of the rankings, is mistakes, especially uncharacteristic mistakes; far more significant than people allegedly not trying. I've never been aware of any of my rivals in M60 not trying, whatever the level of event. Personally, if I 'train', I train at going as fast as I can, and often get a better result when there's no 'major event' pressure. In my experience too much 'intensity' can be a precursor to mistakes.

[graeme]

Not much to add about stats to what greywolf wrote: if you want to weight the higher quality of races, the right way to do it is that one level A counts for two (i.e. something like the square root of the ratio of the number of runners) level C. One can be confident in the ranking of someone with high scores from, say, BOC and JK, but maybe you'd want to see 9 or 10 good scores at local events before you believe it.

gnitworp is right about mistakes, but one should not neglect the randomising effect of bad map/bad controlling in causing mistakes at lower events.

Question: If you take time out to bag the high point on the map, or spend significant time helping an injured competitor or lostl child, would you DQ yourself to avoid distorting the rankings?

[70plus]

Exactly, Graeme. Over the decades we have had a succession of ranking systems and a similar succession of attempts to introduce fudge factors to get 'truer' results. There are so many interfering variables which generate finish times at any event, that we should do as EddieH suggests - treat it as a bit of fun. Especially if one's worst results are dropped to eliminate one's worst failures; only the best orienteers rarely produce poor results. As a POO (pretty ordinary orienteer) myself, I have had some amusement over the years, trying to get rankings higher than I deserve.

Maybe the worst side-effect of a ranking system was about 30 years ago, when Badge events were nominated as Ranking - sometimes in order to boost attendance, not because the terrain justified it (anyone go to Clocaenog then?)

[roadrunner]

Question: If you take time out to bag the high point on the map, or spend significant time helping an injured competitor or lostl child, would you DQ yourself to avoid distorting the rankings?

An interesting point - the original ranking scheme (well, the first one I remember anyway) used, if I remember correctly, the three highest-placed ranked runners on a course as the basis for calculations, the argument being that they were the people you could be most confident had had a "clean" run (and not spent time helping a child, bagging the high point on the map, etc ) I don't support the idea of weightings for the reasons already mentioned, but perhaps there's a case for excluding people whose performance is far worse than their ranking position would predict, on the basis that they were injured, not treating the event as a competition, stopped to help someone etc. But in general, I prefer the idea of keeping things simple.

[greywolf]

the effect of an individual running much slower than expected, for whatever reason, is dependent on the size of the field - but basically it's pretty small.

viewtopic.php?f=1&t=10392&st=0&sk=t&sd=a&start=115

[graeme]

roadrunner has a point that the slow outliers should be relative to expected performance, rather than absolute time. This would help solve the bimodal distribution problem you've mentioned in the past.