O Log - Butterfly Loop


    Butterfly Loop (controls 11-18)

    Recently, I read an article (I believe in ONA) describing a "butterfly loop". More recently, I encountered one in a race (shown adjacent).

    The putative purpose of such, as I understand it, is to limit following, at least thru the loop, or to limit, in high visibility situations, the advantage of seeing what runners ahead are doing. It works like this: some runners get map X, and others in the same class get map Y. The courses on maps X and Y are identical except that the butterfly loop is taken in different orders on the two maps. On map X, using the adjacent example, runners must visit the features in the order 11-12-13-14-15-16-17-18, whereas runners who have map Y will visit these features in the order 11-15-16-14-12-13-17-18 (map Y runners, of course, will have them renumbered in sequence so they are presented on the map in the order they are supposed to visit them; this example would be clearer with control codes instead, but hopefully the idea comes across anyway). In any case, all runners will visit the feature at #11/14/17 3 times. In the end, everyone does the same course, just in different orders. I believe this was used in a recent classic distance WOC race.

    I found this thought provoking and interesting, tho my interests tend to be quite esoteric.

    First observation is that it looks confusing, especially with, in this case, the earlier part of the course in the same area. It looks easy to skip controls or take them out of order by mistake. In practice, however, I didn't have a problem with it. My only comment is that I think it would have been nice to see the printed 11 and 14 transposed, but I didn't lose more than a second sorting this out.

    That aside, two questions come to mind: "Is it fair?", and "Is it a good solution to the problem/does it meet its goal?".

    My opinion on the first question is that I don't think that it is fair, but that the unfairness varies by degree, and in many cases may fall into the acceptable tolerance of things unfair in an O race.

    First blush, I guess is, that since everyone does the same course, it must be fair. But the one time before when I saw this (tho it was not described as such), was the Hudson Highlander of 2001. In that race, everyone did a loop on one map back to a common point map exchange area, then half the field did Rockhouse, the other half did a loop on a much easier map, both back to the map exchange area, do the other map, then everyone did a loop on Sebago. It was the exact same principle.

    The consensus was that those who did Rockhouse second, rather than third, had an advantage because they were doing the harder map when they were fresher. The results, I think, bore this out, and tho no control was done for talent of the runners, some attempted seeding was done, as I understand it. I know I had Rockhouse second, and was quite happy about it, so I think there is something to this.

    So, if we accept that getting the harder of the two loops out of the way first is an advantage, and accept that one loop will necessarily be harder than the other (either technically, physically, or both), it is unfair. I don't find it too difficult to accept these premises. You could go on with 'my mind wavers late in a race but my body is strong, so I hope I get the more technical loop first, etc.'.

    But in this case, the actual magnitude of the difference in difficulty of the loops is small, so the unfairness probably falls under the radar. It is probably trumped by the unfairness of the luck of the start draw, or luck of changes in the weather, and all the other luck factors in O. But, I'd be wary of this construct as the loops become larger, unless the course setter is careful to keep the difficulty difference low, and be wary of other fairness issues that arise in the exploration of the second question.

    Secondly, is it a good solution to the problem of following? My opinion is that I don't think so, and, moreover, there are better solutions.

    First of all, in this case the butterfly loop is only 700m out of a 9450m course, or only about 7.5% of the race. It seems like a bit of extra work to potentially solve the problem for such a small portion of the race.

    Also, the course has been dumbed down to implement this construct by visiting the same point three times. While I realise (and have experienced) approaching the same control from a different direction can be the same challenge as the original approach, in this case it doesn't take much astuteness to make sure you memorize every detail of the circle and control area the first time, and my experience was that it did effectively trivialize two control points on the course (the second and third visits). So this is a cost, and IMHO a non-trivial one. (Moreover, it can be argued that if the approach from #13 is more technical than the approach from #16, for example, those that have the more technical approach third have an unfair advantage due to the familiarity).

    Elephant tracks will also develop into the nexus control faster, and more people will be streaming thru it at any one time. There was a regular queue there the three times I visited. Early starters will not experience either of these things, and thus it may potentially add to the unfairness, and certainly doesn't do much to eliminate the following factor for this control (or rather these three). Of course, in all fairness, my experience was with many classes using the same butterfly control; these factors might not be as bad if it is only for one class, especially for a field the size of a WOC.

    Finally, I was able to use a runner who I knew was on my course while I was in my second loop. I'm not sure if this is chance or not. I figure I could construct scenarios, based on time per K, start interval, and size of the loops, where the following effect isn't even mitigated unless the course setter is very careful. I can't claim the course setter wasn't careful in this case, of course, but it is something to think about when setting this construct to solve the problem of following.

    I'm not going out of my way just to dis this format, I just think the issues are interesting and worth thinking about. It is a bit presumptuous, I guess, to write so much about it having seen it only twice. My hunch is that the WOC people thought this all thru, and it is fine (I mean, what do I know?), but for "normal" races, I don't feel the solution you get is worth the costs. (Another cost, I guess worth mentioning, is that another runner I spoke with in this case thought visiting the same control 3 times was boring. Perhaps 5 times as many runners thought it was interesting; this is just anecdote).

    So, what would I do about following? First would be gigantic start intervals. I'm not sure how big, but big enough so that if you are following, you are either too far down already, or following someone much slower. This also has a big cost, perhaps too big, but I still like it the best if we really care about fairness.

    Secondly, I like forking much better. With the butterfly, you know where the 'do not follow' zone is. Read the map now. Not so with forking. Forking is also unfair, but I think with skilled course setting, can fall into the same realm of unfairness as the butterfly (and be acceptable), yet provide a better solution. The problem is that forking is perceived as inherently unfair, is controllable unfairness to boot, and seems easy to whine about/protest after the race. Not easy criticisms to shake, but I don't think these costs are as high as perceived, and the product you get for these costs is far superior. Perhaps a topic for another day. (Another topic for another day is to look at the Farsta format in this light; my hunch is that some of my thinking is colored by the fact that the butterfly looks like a butterfly, rather than the two loops overlapping like a Farsta, but all for another time ...).