Discussion: Fairness in TT start times

I know I have discussed this with ESC members before, but I can't agree with the fairness of the TT start list.
In the sprint all the reasonable top female finishers are starting within 5 minutes.
It seems to me that the last starter has a big advantage over the competitors starting sooner. This is cetainly true if you are a decent fast runner, your strategy could legitimately be , go out as fast as possible through the first few controls, catch sight of the earlier starters and just reel them in. This is a vialble startegy especially on a sprint in open area with 1 minute start intervals.
I predict they will travel around the course as a group, with the later starters having the fastest times.
Even in the mid distance, several contenders are starting within a few minutes, and as I undrstand, the classic race will again feature a "inverse chase" start.
Some have argued that this gives the advantage to the competitors who deserve it, those higher ranked.
I think the advantage should go to no one. Spread the TT ers among the general group. they are much less likely to gain advantage from them.
I think a start time spread per TT competitor of 4 minutes in the sprint, out ot 10 minutes in the classic would promote fairer competition.

Longer start intervals might be fairer, particularly in the sprint, but I think your prescription is excessive even if we agree that the mooted procedure is unfair (I might argue that all of the TT participants except the lowest ranked one have about the same opportunity to run hard early in order to catch sight of and then reel in people ahead of them.). Even conceding the problem, surely two minutes would be enough.

As to 10 minutes between TT participants in the classic, it appears from the start list that the "inverse chase" start for the classic will be applied to everyone who is eligible for the US Team, not merely to those who intend to go to Japan. The number of men falling into this category would require a start window of almost 5 hours, longer if we choose not to introduce an element of luck in who happens to have a fast non-US citizen starting shortly before or after them by sticking to starting all the non-eligible participants after the eligible ones. I can't see that flying. Not even in Cincinnati :)

There are reasons for and against doing it this way and, these have certainly been discussed by the ESC.

First, we must accept that any individual start race is inherently unfair. Some competitors will gain assistance by seeing other competitors in the woods, while others won't. Some competitors will start early (when it's cool) and others will start late (when elephant tracks have developed). A question is whether we want these interactions to be random or controlled in some way.

The main rationale for using the current system is to avoid the (e.g.,) 10th-best orienteer latching on to the (e.g.) 1st-best orienteer and getting a ride onto the WOC team. By seeding the start list, competitors will be racing amongst those who are roughly of equal ability and, thus, the likelyhood of being passed by someone who is MUCH better is reduced.

A down-side of doing it this way is that when competitors do interact, there is much less likelyhood of one competitor getting away from the other when they are of comparable ability. There's no doubt that if groups form, they will be much more long-lived if the start list is seeded.

Assuming the start-list is seeded, should the best start first or last. The general consensus is that starting later is a (small) advantage. This advantage should be given to the best orienteers (who have proven themselves and are actually vying for the WOC team slots).

Regarding the 1-minute interval in the sprint, I think 1 minute is a pretty big margin in a sprint, so I don't see it as a big deal. Sure, there will be some catching and some grouping, but I would expect it to be comparable to the middle-distance (which has double the winning time and double the start interval).

Finally, one significant rationale for doing it this way is that it is how it is done at WOC. The WOC finals are all based on qualifiers with best starting last. If 1 minute intevals are OK for a WOC sprint final, then I should think they should be just fine for the US trials.

I'm not saying the current system is perfect, but I do think there are some good reasons for doing it this way.

Yes, I agree that this method is fine. I don't understand the assertion that someone would be able to just go out fast to catch someone, because this is a sprint race, and people should already be moving as quickly as their navigation skills allow. If you do get close enough to have sight of the person in front of you and they show you the control position, then that's not ideal, but at least you've earned it by catching a real competitor. But if the TTers were spread among the general group then that would happen more often and more randomly, as people from other classes would be out on the course.

As for the earlier starters, I think they are helped, rather than hindered, by having faster people behind them. If someone catches you then you've already lost a minute in a 15-20 minute race so your chances of a good result are pretty low, but the presence of that person can help you salvage something by finding the control you've been searching for, or just making you run faster. In fact in this system it's only the highest seeded man and woman who don't get that luxury. If they mess up the navigation, the cavalry won't be coming to help.

It seemed clear enough to me - if a competent navigator running the same course is in sight ahead of you there's an opportunity to follow them at a faster speed than your navigation skills would allow if they weren't in sight. It would mean taking a chance on them making a rare error and there would be a risk of running faster than one's navigational ability in the attempt to get close enough to see them ahead of you but there's certainly a potential advantage to be had. However, as Brian observes, individual start races are inherently not perfectly fair. The best we can do is make them pretty close to fair.

I will note that a one minute start interval must be a great deal more generous in WOC terms - I haven't verified this but I expect the top ten qualifiers for the WOC Sprint Final all finished within a minute of the winning time for both genders. I'm not confident that this will or won't be the case for the male TT entrants and I would be astounded if it were the case for the females, i.e. I think it is much more likely that any nth seed will make up a minute on the n+1th seed at Flying Pig than at WOC.

The current system is as fare as any other system and it has been adapted by ESC well in advance before the team trials. I think this discussion can have place, but now it is out of time and place. If one has proposal for change it must be discussed before the rules are set - for next team trials.

Thank you, Greg. The ESC will be happy to discuss the start procedure in time for the 2006 Team Trials.

Real fairness will never be achieved. But I'll would say that current system is fair enough. How often will it happen that a runner that's not a top runner will have a top result based on unfair conditions? I would say never!

For example competitor A & B run a course of 10 controls with no interference, independent of each other. They both have the same time, each with one 1+ minute error. A's error occurs on the 3 rd control, B's is on the 8th
Now with 1 minute start intervals: A goes first and makes their error on 3, B catches them. At 8, B has A's help and doesn't have error. B wins
Then B goes first, A makes error on 3 falls 2+ minutes behind, B makes error on 8, A & B tie .
IE Results are different.
Not suggesting we change now, but excuses like all races are unfair, and this is the way WOC does it does not mean it is the best or fairest method.

It may be a fine semantic point but I think all interval start races being somewhat unfair and the desirability of emulating the WOC start procedures qualify as reasons rather than excuses. As Brian implied, we could make all races perfectly fair by using mass starts but we'd reduce the importance of navigational ability by doing so. By using any start interval at all, some potential unfairness is created. By lengthening the start interval to decrease the chance of unfairness due to following, the chance of unfairness due to heavy rain or other inclement weather beginning after some competitors have finished but others are still on course or not yet started will be increased. Good luck getting everyone to agree as to what is the ideal compromise (not quite an oxymoron but very close).

Some course designers started using butterfly loops to make it harder to follow.

I think the butterfly loop is a good concept, but difficult to execute on ridge and valley terrain. It's very difficult to find five good control locations that are all spaced about 100-150m apart. Making the loops much longer than that devotes too much of the course to the loops, particularly at the sprint and short distance.

Although others seem to feel differently, I'm of the opinion that interaction between competitors of similar ability is more significant than interaction between competitors of dissimilar ability. That is, when someone ranked 10-20 points higher than me passes me in the forest, there's no effect other than my being a speed bump for him and my getting a litttle wind burn. But if I encounter someone who's about my speed, we may both wind up traveling faster than we could individually. So I have not so much of a concern for the issue of a slower competitor following a faster competitor to a place on the podium. Rather, it's a question of whether the #6 ad #7 runners can hook up and go faster than the #4 and #5 runners on their own. And that argues for separating people of like abilities when creating a start list. An Eskilsson start (interval starts based on previous day's result, with the fastest going last) maximizes the chances of people of similar speed encountering each other.

I think butterfly loops are overrated as a panacea to this sort of thing. If two runners come into the butterfly together, what are the chances that they'll come out of it within sight of each other? Pretty high.

Of course the stated method is the one that will be used in this event--no changes to the agreed-upon procedure should be made now. But I don't see any problem with having a discussion about the topic while it's on people's minds.

Jeff's examples ignore the presence of additional competitors before and after these two, which is what gives A and B the same possibilites, to the extent that it can be controlled (unless they are the very first or very last runners, who I agree are disadvantaged). Of course the actual interaction that you have with others will depend on how those seeded near you actually perform, but that can't be controlled. The only way to completely eliminate any possibility of interaction would be to have a very long start interval, perhaps impractically long, depending on how sure you want to be. And it does make sense that a trial to determine the WOC team should take place under the closest we can simulate to WOC conditions, rather than under some alternate set of conditions in which interaction never occurs.

As for butterfly loops, I agree that they are OK but perhaps not worth the effort. Two people who reach the loops together may well have the same order. And what if three or more reach the loops together? Randomly, some will get to have partners, while others will have to go alone.

By all means it's a good topic for discussion, with the understanding that changing it for this weekend is off the table :)

I certainly do expect packs to form this weekend, although the people who are clean are most likely to catch up to a pack and the people who are not are most likely to get caught by a pack, so there's bit of 'fairness' in that...

And one perhaps good thing about this weekend is that we do have some variety in the start lists, with more or less random starts on Saturday.

I understand the concern about butterfly loops where competitors going into the butterfly together are quite likely emerge at about the same point as well. But what about a two loop course format similar to the billygoat sprint last year with two smaller loops passing through a common crossover control.

This way you could still maintain a 1 min start interval, but alternate which loop runners start on so that 'in the woods' you actually have a 2 minute spacing between runners. Or if you go with 2min start intervals you wind up with an effective 4 min gap. At least this doubles the amount of time one has to make up to close the gap on a rival assuming runners make no major errors - and if they do, then even if they hang on to the overtaking runner they are still two minutes back.

And I have to disagree with J-J about just being a speed bump when passed by faster runners - I can certainly run faster than I can navigate and in goat-style races when I manage to get ahead of faster runners (by their errors or well placed skips) then when they pass me I can certainly pick up my pace and hang with them for a while doing less navigation myself - so there is the potential for a good runner of less navigational ability being able to gain an advantage in this case - witness Dave Dunham latching on to Hammer at the Wachusett Goat a few years back - stayed with him until Mike figured out how to shake him towards the end.

But I also think that mimicing WOC practices is probably the best route to follow - they may be unfair in some regards - but that same element of unfairness will be present at the championship event - so why not consider it akin to training specificity and just accept it?

I really can't understand why someone would complain about fairness/unfairness, when it comes down to start time intervals, no matter when they start, 99% of the time the best competitors will prevail.One minute in sprint is a huge gap if you are caught and almost impossible to make up by a lesser competitor against one of the top orienteers.
Example: Back in 1992 IOF organized a demo park sprint-o in Central Park, Manhattan,NY,competitors all the national teams members from the 1992 Canada/US World Cup events.The majority of the top 50(male and female) orienteers were present and 40 more best from other national teams, for a total of 90 starters, all mixed randomly.Start intervals 30" and nobody even mentioned fairness/unfairness, and when you started If you were not first,or second, you could see the 2 previous starters.I was a good runner at the time but
coach's, presumed advantegeous strategy didn't work out for me and some others who try it.As Brian clearly pointed out they are other more "severe" unfair aspects of orienteering, which will give a lot more advantage/dissadvatage to some orienteers and none to others.Compared to those, start time intervals between competitors is minuscule, if not 0.

As somebody who used to do a lot of cross country skiing in the time that most races were 30 second interval starts, I can say that this discussion is most interesting. I know some people, including me, who would have fits over their starting positions before ski races. Clearly, starting immediately in front of or immediately behind your main competiton could have a substantial impact on who was going to win, particularly if you let that sort of thing bother you. However, a wise coach once told me that it was better to worry about the things that you can control (in orienteering, you physical and mental training) than the things that you can't control (your start position and whether someone was or wasn't going to pass you during the race). People running into other people in the forest is part of the game of orienteering, but a part of it that you can't control. So, its probably better, in my opinion, to concentrate on your own performance than that of others.

Assymetrical butterfly loops will fix the problem of competitors "seeing" each other. Since the difference at the top is not huge formation of "trains" is almost certain. Placing butterfly loops on the second half of the course will help braking "trains". It is not panacea but worth start using for national meets.

I can understand why someone would bring up potential unfairness in the start list. Maybe discussion about it will bring up a better way. I feel much the same way that Coach does but as a perennial also ran in the trials would hesitate to bring it up myself. I also have no solution to propose. I think it favors the higher seeded runners but assumed that was intentional. It is their reward for earlier good performances, the same as a home field advantage in the playoffs. We want our best to win and this gives them a better shot.

Some of the relevance of what happens in international competion escapes me. The original posting was about our women in the sprint who I do not expect to all finish with times a few seconds apart. If we really want to mimic WOC conditions perhaps we should have our best orienteers start first and followed by top Canadians or fast runners who already know the course.

Perhaps I missed it, or it is part of the sprint rules but I did not see anything in the TT rules which said what the start intervals would be.

I'm with JMM. My advice to the esteemed A and B is to run fast and not make any errors.

USOF does not yet recognize the Sprint discipline or have rules for it. In the absense of USOF Rules (not to say that if such existed, they would necessarily be followed), the event is governed by IOF Rules according to USOF 1.4. The IOF indicates a 1 minute start interval for the Sprint (Rule 12.14).

I think it's interesting that this discussion brings out more about peoples' attitudes.
The 2 viewpoints seem to be:
TOUGH!, Just DEAL WITH IT!
It is an issue, but can we, and should we, try to make it better (fairer)?

Before anyone answers coach's last question, I'd be interested to hear TT participants' and other observers' opinions as to how much group formation was a factor among the orienteers seeded high enough to have a realistic chance of making the team. Watching people run to the go control, my impression was that no large trains formed among the men who started after perhaps 9:12 (though some train formation may have been forestalled by Mikell Platt, Leif Anderson and John Frederickson missing their assigned start times). I saw what looked like one four person train among the elite women but one of the people in it told me that it had coalesced only at the very end of the race.

I'll add that it does appear from Erin's training entry for Sunday as though the formation of groups may have been more of a factor in the women's classic.

swampfox, eric, and I were within sight of each other for most of the second half of the men's classic.

In the classic, I slowly passed the people 3, 6 and 9 minutes ahead of me. 3 was in sight by didn't hang on long at all (200m?), 6 was fixing a contact, and 9 I passed on a different route choice. After that I was quite near Eddie for a few bags (I think he started 15-18 minutes earlier, due to illness...), and only at the 2nd to last control did -3, and 6 (again) catch and pass me.

There was a 5 person mass finish including Leif & Velichko as well as Swampfox, Ken & Eric, but I don't know when the last two converged. There are epunch splits on OCIN's site so somone could draw some nice pictures, if desired..

I plan to write up some comments about my classic race and put them on the extra ultra secret "Real, real, real Swampfox" training page, but doctors have ordered me to first take it easy for a few days and fully recover from a highly advanced case of multiflora rose thorn poisoning.

Much better to take a few days to recover from the multiflora rose than a few weeks to recover from, say, poison oak... I was happy to read on the BAOC site that they do indeed consider some areas just too thick with poison oak to offer them as A-event venues. I also noticed this event in their schedule...

ocin has added the results into "splitsbrowser", and if you select the view "race graph", it makes it easy to see where people grouped up. it works best if you only select a few people at a time (who started near each other), otherwise you can't really tell what's going on.

There are some very interesting things in those graphs (particularly for the Sunday race which is long enough for following to be an issue). Some people, although passed by multiple people starting soon after them, seem deliberately not to group up with them. And some people clearly speed up when caught by others behind them. Very, very interesting.

Another interesting thing is that it seems like the start clock drifted by about 6 seconds over three hours. Can someone reassure me this is the beeping clock drifting, not the SportIdent start control clock drifting?

are the non-groupies known to be slower runners? maybe they couldn't go any faster.

also, I overheard one of the OCIN guys at the saturday start mention that the beeping clock needed to be reset periodically due to drift.

The beeping start clock is known to slowly drift. They guy who made it is still working on that. We trust the SportIdent clocks.