mandag 27. mai 2013

Course analysis Norwegian Spring

Almost one month has passed since Norwegian Spring. Here I bring some thoughts on the elite courses form my point of view as the course setter. I hope you have not forgotten all about the map, course and competition, and can bear the thought of reading this course summary even if it turned out way too long even after leaving out quite a lot and is written in my school english (Hopefully writing the article in english will make it slightly easier to understand for foregins than what Google translate offers.). If you have forgotten all of parts of the events you may refresh your memory e.g. by reading the two following blog articles:

Europacupåpning i Norwegian Spring
Norwegian Spring 2013

Maps, solutions and results may as ususal be found in the map archive.

Day 1, Varteig:
The final part of the day 1 course

The terrain for the first day was not chosen because it was an optimal terrain for trailo, since it is not, but in order to arrange trailo from the same arena as the main event for the runners. The area where you were looking over the fields in the end was the only part of the terrain I initially thought was challenging enough for an elite course, but for every hour spent in the terrain, other parts also become interesting. In the end I actually think the terrain suited quite well for trailo, even though the elite course more or less passed straight through the only real forest in the beginning of the course. In general I thought that the course this day was not very difficult and my guess was that 12 people would have all controls correct. 5 people made it and I am happy also with that number.

The event started with timed controls. The two controls should be fairly easy as the overview was very good. The contours were also drawn at natural places in the terrain - no trickery, just plain contour reading. What could confuse some was that the hill to the far right in the green area was not visible. A fairly good tactic here was probably also to use the parking area to locate the hill and the spur.

Egil Sønsterudbråten in action at control 16 to 18.
Starting on the course control 5 was the first zero control and the first really demanding control from my point of view. The first impression of the A flag could most certainly be that it was correctly placed, but carefully reading how the contour is drawn, and checking the distance from the cliff and from the path should show that the A flag was too close to the cliff. For control 7 and 8 I had some real problems with fitting the map to the terrain. At some point I considered to leave out these control as I could not figure out how to fix the map. Finally it turned out that the whole hill was too close to the houses, and I think I managed to fix it quite well (see the difference on the map in the bottom of the article). Control 9-11 demanded some precise map reading on distance, but it seems like most people managed that. The area of control 12 (as with control 4) was mapped in order to give some variety. I originally wanted a between-problem on no. 12 as well, but chose an international styled zero problem instead. Those which tried to solve the control from the decision point would probably miss this one, but those who moved along the road should quite easily see that the flag was too close to the well. Control 13 and 14 was supposed to be quite difficulty zero problems. 14 created problems, but as many as 87% solved 13. 14 could be solved either by reading that map and observing that flag B had to be on the spur and that flag C was on the same distance, or by walking far up the road, almost to the finish, to locate the correct knoll. 15 was an international styled control that was included in the course just a week before the competition. It was not intended as a very hard problem, but with only 43% correct answers, it most certainly was to the competitors. Sloppy control description reading is probably the reason for some of the mistakes (it was the middle of the spur, not the south-eastern). The control was intended to be solved by measuring the distance from the forest or the hill to the center of the circle. Then two quite easy controls followed - the 17th having 99% correct answers. On that distance I did not expect that only one person would miss that one.

The percentage that had control 17 correct was impressive, but the percentage that managed control 18 was not. Only 44% chose the correct flag here. Originally I did not have the D flag even close to the correct position, but it was included to make the problem more challenging. The D-flag was set almost one full contour higher, and I thought (and still think) that following the height from the B-flag and maybe also finding the boulder south of the control would show that the D flag was too high and the E flag was perfectly placed. However, I have no problem seeing that those trying to read from the right could possibly put the contour a bit higher and thus excluding the E-flag. It is however very hard to determine exactly were the contour goes in that area, so I would say it is an unwise decision from the competitor to rely on reading from the right instead of reading from the left were there was more distinct features. Looking at the problem now, I see that moving the D-flag up 1m to the next contour and maybe moving the correct E flag up half a meter could have given a raised the number of correct answers. To me it is however nice to see that even if the map probably was not 100% perfect with the height of the contour through the green area north of the circle, almost all of the "big names" answered E. The rest of the course was fairly easy.


Day 2, Kjeøya
While I initially was not perfectly happy with the terrain the first day, I was perfectly sure right from the start of the planning process that the second day was a very nice area for trailo. There were plenty of complex "black and brown features" in a quite small area, and the view was fairly good in most areas as well, making it possible to set a few distance controls as well. My guess was that this day was more difficult than the first one and guessed that 6 people would get 20 points this day. I felt that there were several quite tricky controls and also the maximum time was only 96 minutes. The course was once again slightly more challenging than I thought, leaving only 3 competitors with full score.

Overall winner William Rex on the 8th control at Kjeøya
The course started out with a quite challenging control. If you were not focused enough here, it would be easy to go for the C flag right away and not spotting the correct knoll behind. Control 5 was in an area with very complex ground details. It was one of the first areas I spotted in the planning process and still is one of the controls I am most happy with. It is hard area to map correctly, but I feel that the end result of the map was good. The 6th control was the hardest control, as expected. I did however not expect it to be that hard. Only 29% of the competitors answered Z, as was the correct answer. That could maybe have been OK if it was mostly the upper half that had this one correct (as was the case with control 1), but the Z-answers were spread out over almost the entire results list. The A-flag was placed 7m from the correct position, but the flag was placed such that most bearings would pass the flag anyway since the angle of view was limited. Further, the flag was placed on the top of a steep section with the contour below passing in the bottom of this steeper part. That is, the A-flag was not more than say 1m too low in terms of height. The third reason why I think this control was so hard was that you could not see very much behind the flag. My idea to solving the control was to determine the top of the hill behind the control (you should easily see it when going back towards the start). Then you should be able to see that the A-flag was too far from the upper part of the hill. Secondly, distance estimation from the clearly marked spur by the tree 30m south of the control should indicate that the flag was too far down. Finally one should be able to spot that there was no area of bad runability with spread out boulders below the A-flag as there should have been according to the map. When checking the controls after the competition I could agree that bearings from east were a few degrees wrong - pointing slightly more towards the flag than they ought to. By adjusting this bearing in some way, it is likely that some more people would have concluded that the flag was too far down. That being said I still think the control was solvable, but I have no problem seeing that the combination of a few degrees skewness and long distance unfortunately made the control a bit too hard.

Control 6 day 2 was a though one.
The 7th control was an international inspired control with all flags in a row down a single longer re-entrant and the problem was to determine the center of the circle from the surrounding features. The three next controls required accurate map reading on distance. The 9th was a zero control were the control placement was used to give the illusion of there being one flag at each of the three cliffs while all of them actually was on the same one passing through several curves. The area around control 15 and 16 was quite nice, but the bad visibility from the path made it hard to make the controls here challenging. For a timed control or tempo station the area would however have been well suited. Even though it should be fairly easy, quite a few missed the clear zeroed 16 control. On the way back there were two controls in a very complex cliff area with limited visibility, which I am quite happy with. I could probably have enlarged the map to 1:2000 scale here to improve the readability, but honestly did not think it was necessary here. The course finished off with a control with complex contour details in an area with limited visibility. I originally intended to put out more flags on this control, but the changed my mind when seeing the visibility some week before the event.

After the course there were three timed controls at the arena at some distance. I knew that reading the cliffs on the second task would be difficult, but still almost all the top competitors answered correct, just at different speed - just the consequence I would like to see on a time control.
The intention of leaving the last 6 competitors for a "grand finale" on the timed controls was that everybody could see them in action at the timed controls. Unfortunately, that intention was not communicated to the competitors - that I will hopefully remember in a similar situation.

The mapping process
Since I live in Oslo, an did not have much time to travel to Halden/Sarpsborg to revise the map in addition to the additional precise trailo revision around the control sites, I needed help with preparation of the map. Therefor Morten Dalby (which has mapped most of Halden SKs maps) was hired to make a new trailo map at Kjeøya (day 2) with 2.5m contours (the original map had 5m contours) and to revise the existing sprint map in Varteig (day 1). Although the mapping work started a bit late due to some misunderstandings, and the cold and snowy winter complicated the mapping work, Morten manage to do an excellent job at Kjeøya, making my job with the final adjustments easier. Unfortunately he got some back problems during the winter, making him unable to revise the map in Varteig. Therefore the job was transferred to Bjørn Paulsen just some weeks before Easter. Bjørn's work helped me, but as he was not used to the accuracy demanded in elite level trailo maps, I had to do quite a lot of both the general revision in addition to the control specific trailo revision. To show some mapping work that has been done three versions of each of the two maps are included below: Version 1: Original map, version 2: Final map from hired mapper, 3: Competition map with my adjustments. Click the images to enlarge the maps.




Summary
All in all I think the event turned out fairly well and I hope most competitors left the competitions happy, even if some of you got quite many mistakes. I want to make courses for the top level trail orienteers, and did that by demanding precise map reading both on shorter and long distance. That being said, I don't like correct decision rates under 40% especially when mistakes occur as often on the first page as the second page of the result list. I know for sure that I have reached the upper limit for difficulty and do not need to push the difficulty in the courses any further to "make it challenging enough," which for sure is relieving as it is much harder to make difficult courses that are perfectly solvable and not involve any chance games compared to making easier controls.

As those of you that were present at the event, I was "slightly (!)" stressed during the weekend. I had started my work early, spent a lot of hours both in the forest and in front of the computer to organize the event and from my point of view I did what I could regarding the preparation. I still don't understand how other organizers manages to arrange big events without working 20 hours a day the last week and running around on the arena like stressed monkey. If you got some advice on how to avoid such an over-stressing situation, I would really love to hear them.

Finally, if you have comments on the courses, maps or anything else regarding the event, please don't hesitate to write a comment below send me an email or talk to me at some event - in a language you assume I understand. Did you use a natural solution method that gave a different answer - please tell me! The only way for course setters to know what was good and not is to get constructive criticism and feedback on the courses.

- Martin Jullum (email: martin.jullum [at] gmail.com)

4 kommentarer:

Rusky sa...

Hi Martin, thanks to this article and your competitions. I have read this one very well and I agree with you. In the first day, I had two mistakes: n.5 ( the real countour seemed lower than the road on the left of point of view) and n.18 (I'm not "big name" and I saw only the right part); on Sunday, the number 6. I remembered n.9 (very interesting) and n.14 (I answered well using the distance from the bend of the path); I ask you how did you pick the flags at n.8...-))
Regards Marco

Martin sa...

Thanks for your comments:
D1C5: There is a well known problem with drawing contour lines over roads as the roads are often a bit higher than the rest of the terrain. In this place the contour was passing "under" the road, which I understand is a common way to solve these problem. To be 100% precise the map probably should have looked like this: https://dl.dropboxusercontent.com/u/32776513/div/mapNS.jpg
Some experienced mappers (Remo, Lauri etc.) could probably say more about this and possibly correct me if I am wrong.
D1C18: I wrote almost ;-) But I think that it is not a big coincident that most of the "big names" actually chose the correct one here. Often experience is more important than accurate map reading in that kind of situations. I have still not done done trailo on a map that I could not spot mistakes on while being at the course. (And yes, I will find mistakes in your maps as well, Lauri :D)
D2C6: Enough said I think, it was intended as the best control of the course, but it turned out too hard.
D2C9: I also liked this control quite much. But it could have been a disaster if I did not spot the contour mistake on the same contour as the correct cliff were just just in front of the decision point. (See the version 2 of the map). I also removed a 15m long path in the control circle. It would have been easier if that was there, but the reason I removed it was that it was very very small.
D2C14: Actually that was a solution method I did not think of and hence did not check directly - luckily it worked out well. May I ask why you chose that method?
D2C8: You mean how I set out the flags and took them in? Climbing with an axe in one hand and a handspike in the other one! ;-) Picking them in: Some Elite runners of Halden SK (Jurenikova, Eliasson, Coupat, etc) used the course as a training after the competition and therefore also picked them in. Best result form training: Fredrik Eliasson 17 points and 56 seconds when trying for the first time. Otherwise the results were not so good. Jurenikova managed all controls which were not zeros, thus she did not find the correct precision to use this time.

Anonym sa...

Hi, about D2C14: I was not sure between B and D. I saw the rock behind them but I didn't understand where it finished. So, I decided to use the distance from the bend of path and all became easy.
Regards Marco

Martin sa...

D2C14: To me the cliff was the secure thing if you were not able to follow the curve around to the right. The left side of the cliff was maybe not that easy to determine the end of, but the right side should be fairly clear.
Anyway, being a competitor and a course setter is very different. My impression is that you as a competitor think of and use many more solution methods than as a course setter.

Thank you for your comments, Marco, they are very helpful :)