Competition Review by Heiner Biesel

Date: Thu, 02 May 2002 22:36:57 -0600
From: Heiner Biesel
Subject: On competitions. Part 1


I can't claim to be an expert on hang-gliding competitions; many competitors are much more experienced than I am. Still, I've been to five world-class comps, and I've learned a lot from competing at that level, some of which I'd like to share here.

This report is in two parts: the first part deals with specific competitions, the tasks called, the scoring, the competitors, and the skills and experience required to do well there.

The second part will explore some alternatives, and speculate a bit about the future of our sport, and of the competition scene.

Please keep in mind that these are merely my personal observations and opinions, and that others will have very different ideas and opinions on these topics. I hope to generate a discussion, and perhaps some changes in competitions and the scoring methods used.

The Florida Competitions 2002

The two comps held at Wallaby Ranch and Quest Airpark from April 13th through the 27th were world-class events in every sense of the word. Not only were they attended by many of the world's best hang-glider pilots in classes 2 and 5, they were well organized, effectively staffed, and smoothly run.

It is difficult to appreciate how much effort and time goes into planning and running such a competition. The logistics are staggering: many tugs must be assembled, flown, and kept running. Carts have to staged, crews for their retrieval organized, line crews for staging the pilots must be trained and assembled, a goal-line crew must be trained and dispatched, food and water must be provided, toilets, showers, trashcans, water, cleanup, scoring, weather, lost-and-found, batteries, repairs, information, first aid, paramedics.a thousand details must be attended to.

Almost all of the staff and volunteers were friendly and supportive even when one of the inevitable glitches occurred. The pilots, with a very few exceptions, were supportive and fair. The competition remained in the air, not on the ground, except for the ritual dash for places in the staging lines.

At Wallaby, things ran even more smoothly than they did last year. The area for launching has been expanded, and grass is growing in formerly barren areas. A beer-and-popcorn tent was greatly appreciated by all, and the festivities ended with a wonderful fireworks display. JC Brown and Peter Gray initially struggled with GAP 2002, the new scoring system, and overcame weaknesses in the new software by putting in many extra hours of effort. The new version of the software is intended to reward more highly those pilots who leave early and lead along the course. However, the large majority of points still accrues from completing the task, and from flying fast.

Quest set a new record for the most launches in the shortest time. This year they also provided meals on-site for everyone who wanted to eat there, they expanded their facilities, put in a new and bigger office, and improved their camping area.

The tasks:

At both Wallaby and quest, the tasks consisted primarily of out-and-return flights and triangles. These are preferred tasks when winds are light and permit them. JC Brown experimented one day with a timed task, basically a speed task over a specified course, with the total being equal to the number of miles flown by 3 PM. This did not work very well, because no competitor managed to make it back to goal by 3PM. The GAP 2002 program also did not know what to make of this task, and the scores had to be calculated by hand.

On days with strong winds, downwind tasks to goal are likely to be called. Many competitors prefer tasks that return them to launch, or at least don't leave them very far from launch, because it simplifies and shortens retrieves. Flying over 4 hours in a contest for several days in a row can be very tiring, and everyone appreciates being able to return early.

The scoring:

JC Brown used GAP 2002 for the Wallaby Nationals while the folks at Quest stuck with GAP 2000. It is not possible to tell whether the standings would have been different if this had been reversed, unless one scores both contests with both systems and compares the results. Although GAP 2002 increases the rewards for early departures with additional points for leading at least some of the time, this bonus is not sufficient to offset the advantages of flying in a gaggle. Unfortunately, the precise algorithm for either scoring method is not easily accessible, nor are the tactical implications of these algorithms described in sufficient detail in order for most pilots to make optimal flying decisions on course. Competitors are therefore in a position of in effect being told "We won't tell you how you'll be scored, just be first to leave, first to arrive, fastest on the course, and you'll do fine." In fact, it is not possible for a competitor to know how well or poorly he has done, until all the flights for that day have been evaluated. The performance of other pilots not only affects speed and arrival points, but also the validity of the day, which in turn affects the total points available for that day.

It usually does not pay to leave early and leave the gaggles behind, because the usual fate of those who attempt this is that the gaggle as a whole flies a bit faster overall - since it is able to find and exploit lift more effectively than a single competitor - with the result that the lone pilot may arrive first at goal, but still come in 5th or 6th place. I experienced this personally on day 4 at Quest, which used GAP 2000, when I departed 15 minutes before the main group, flew most of the task by myself, arrived as the first ATOS-class wing, but came in 5th for the day because a gaggle of 4 other pilots had caught up to within 5 minutes of my time over the two hours it took to complete the task. That gave me 43 departure points vs. 26 for the fastest pilot in the gaggle, who also got 445 speed points for being ten minutes faster over the course vs. my 332. Overall that meant 85 more points for flying faster in a gaggle than flying first and alone.

The scoring systems are designed to exaggerate small differences in speed and arrival times for the fastest competitors, which creates a dash for goal among those pilots who are on final glide together. Ideally one arrives at the last thermal at the very top, gains just enough altitude to be able to glide to goal at high speed, and flies to goal as fast as this altitude permits. There is comparatively little difference in the raw performance of the top flexies and rigid wings, and so it comes down to ballast, reduced drag of pilot, glider, and harness, and tactical tricks, as well as the ability to climb well and to choose good lines of flight.

The competitors:

Since I fly an Atos I can really only comment on the rigid wing pilots that I've flown with. The best of the flexies are able to mostly keep up with stock rigid wings. When a rigid wing is flown by a top pilot with a clean harness, streamlined control bar and downtubes, and perhaps some ballast, no flexwing is able to keep up with it, with one exception. During the final glide flexies fly at very high speeds, faster than most rigid wings can sustain. This is due to the aeroelastic deformation of the flexies which reduces the pitch forces they experience when encountering gusts or vertical air currents at these speeds. Rigid wings become hard to control at high speeds because they do not deform, and the pilot experiences sharp changes pitch forces.

The pilots I've flown with fall into roughly four overlapping categories. The very best of them - Christian Ciech, Alex Ploner, Hans-Joerg Truttmann, Johann Posch, and perhaps one or two others - are of roughly equal ability, fly the best equipment available, and with ballast are able to pull away from the pilots in the next category. Some combination of ability and equipment account for this difference. The very best of them are also young, fly year-round, and have factory support.

The next group of pilots flies almost as well as the first, but will often be left behind on inter-thermal glides. Any one of them can win an individual day, but the probability of one of them winning a contest in which pilots of the first group are competing is fairly low. There are perhaps 5 or 6 American pilots in this group, and probably three times as many in Europe and the rest of the world. These pilots also fly with the best equipment they can obtain, including low-drag harnesses etc. but they do not enjoy factory support, probably do not fly as often, or in as many contests as group one does, and are as a group a bit older.

The next group of pilots consists of competent and seasoned pilots who have not taken full advantage of all the means available to speed up their gliders. Many of them compete relatively rarely. Some are new to the contest scene, others are old-timers who enjoy an occasional contest but can't or don't want to commit the large amount of time, energy, and money required to be competitive at the very top.

Finally there are a number of pilots who fly only an occasional contest, or who are primarily recreational pilots who fly in a comp for the fun and the learning experience.

The skill set:

Someone once wrote "The race is in the vertical." I used to believe this statement. Nowadays I'd say that perhaps half the race is in the vertical. The current scoring methods award points for finishing ahead of other competitors, even if the time difference is only a few seconds. This favors fast gliders and high speeds on the final glide.

Pilots flying in gaggles can usually cover distances faster than a single pilot, since the gaggle has a much better chance of locating lift. However, there is usually no advantage to being the first to leave a thermal and hunt for the next. Pilots staying behind in the very light lift commonly found at the top of thermals, and watching the progress of other gliders, can choose those lines that appear to produce the least sink. Being slightly behind and well above numerous other gliders is the ideal position to take advantage of a gaggle. This fact produces gaggles that continue to circle in weak lift at the top of a thermal, waiting for some glider to be the first to leave. According to speed-to-fly theory, one should leave light lift once it falls below a threshold value, in the expectation of finding stronger lift in the next thermal, but this is a sub-optimal strategy when flying with others in a contest gaggle. A consequence of this is that being able to climb well is an asset, but not as much of an asset as one might think. One must be able to climb at least as well as the average pilot in the gaggle, and a bit extra is a help as well, to make up for perhaps falling behind in inter-thermal glides. But being able to climb much better than the gaggle is not a great advantage, since one will just have to wait at the top for the gaggle to catch up, and for some gliders to head off for the next thermal.

To do well in the current competitions one needs to be able to keep up with the lead gaggle no matter what. It is a common experience to be slowly left behind by the faster pilots in a gaggle. After a couple of thermals one arrives at the bottom of the thermal as the faster pilots are leaving it at the top. At this point one has essentially lost contact with the gaggle. The lift is likely to have weakened, and one will be far behind on the flight to the next thermal. The only way to catch up now is to find stronger, better lift along a different course line, and to hope that the gaggle runs into areas of weak lift. Occasionally this happens, but not often enough to make it a viable strategy.

There are several different styles of thermaling. Most common appears to be that of following the other pilots, keeping essentially to the same size circle as everyone else. This has the advantage of minimizing disturbances in the gaggle, and of gaining altitude at the same rate as the average member of the gaggle. More adventurous pilots search for better cores and constantly adjust the position and diameter of their circles. This gets them a better climb-rate if they do it well, but it also annoys other pilots who often have to adjust their circling to avoid collisions. Of course, once a better core has been found, everyone congregates to it, and the game begins anew. Although the roll and yaw behavior of rigid wings and flexies are different, skillful pilots can fly these two types harmoniously in the same thermal.

The interthermal glides can make the difference between a good and an excellent performance. Typically, pilots fly considerably faster than speed-to-fly theory would suggest, and on these fast glides it is very important to minimize every form of drag. It is very frustrating to watch another glider pull away, and to be unable to do anything but watch. Top pilots pay great attention to every part of their glider, harness, and equipment, and seek the most streamlined form for everything. They maintain an optimal body position, tuck away everything loose, minimize strings and supports, put fairings around their hangstrap, and they fly with as much ballast as they can carry.

Without ballast even a very good pilot cannot keep up on long glides, because the additional weight affects the entire polar of a glider. As a first approximation, the best glide angle remains the same, but occurs at a higher speed. Of course, minimum sink is increased as well, but the small increase is usually more than offset by the faster inter-thermal speed one can achieve. Ballast presents problems in two situations: at landing, and when conditions get light and one wants to minimize sinkrate. In both situations one wants to get rid of the ballast quickly, and therefore containers of water or sand which can be quickly emptied are recommended. Unfortunately such containers are bulky, can leak, and modern racing harnesses make it difficult to place these inside the harness. Some pilots fly with lead weights. While these are compact, they present a danger during landings and especially during crashes, and they cannot be jettisoned.

I'm still looking for a good solution to the ballast problems.

Regards,

Heiner Biesel

From: Heiner Biesel
Subject: On competitions. Part 2


As promised (threatened?) here is part 2 of the report.

It will come as no surprise to regular readers of this list that I do not enjoy flying in gaggles at contests. I do it to the extent that it is necessary, but I am always seeking alternatives. However, current contest tasks, and the scoring systems used, almost demand that pilots fly in gaggles until the final race to goal, if they wish to place well in the contest.

I have already outlined the reasons for this in the first part of this report, and in this part I'd like to expand upon this theme, suggest alternative tasks and scoring methods, and comment on the contest scene in the United States. Again, these are my observations and opinions only, and others certainly view many aspects of all this very differently.

Since thermal lift is distributed semi-randomly, and produced sporadically, a cooperating team of pilots will always have a better chance of locating such lift, and of optimizing the use of it in performing some soaring task. This much is undisputed, and with current tasks and scoring systems results in gaggles of pilots who cooperate until the moment when they are able to go on final glide. Since the final glide is rarely longer than 10 miles, while most tasks are longer than 60 miles, this means that a serious competitive race exists typically only during the last 15% of a task, the remainder being a semi-cooperative search for lift and jockeying for position within the gaggle.

As a long-time cross-country pilot I have some deeply ingrained preferences and biases. I have learned to fly mostly alone, to evaluate routes and thermals in the absence of other gliders marking them for me, to make my decisions based upon the need to stay aloft and make good progress for many hours during changing conditions. Racing along a course defined by turnpoints toward a not-too-distant goal requires additional skills, as well as decisions and choices that would not necessarily be optimal on a pure XC flight. While flying a contest task one often flies faster than would be prudent if one were going for long distances, relying instead upon the ability of the gaggle to locate lift that one would have to hunt for if flying alone. There are various tactical games that can be played with the starting time, and along the course, and positioning within the gaggle becomes ever more critical until the last thermal is taken.

Many competitors enjoy this aspect of competitions, and the current GAP scoring systems reflect their preferences. I tend to prefer simpler formats that emphasize the strategic choices of a single pilot flying a course alone. I am not alone in this, but I suspect that pilots who view things as I do form only a minority among competition pilots. Nevertheless, I will outline some alternative tasks and scoring methods that appear to me to bring more of the spirit of cross-country flying into the competition scene.

If the intent is to place more emphasis on and to reward individual strategy, rather than collective flying in gaggles, some means of offsetting the advantages of gaggle flying must be found. Assessing penalty points for excessive gaggle flying, based upon an examination of track logs, has been suggested, but I think that this would not only be unworkable, but would also increase the complexity of an already overly complex scoring method.

More and more sailplane contests are flown in a format called distance-over-time, in which each competitor chooses his departure time and has a fixed amount of time - say three hours - during which he must cover the longest possible distance. Variations on this scheme include specifying an initial turnpoint for all competitors to get everyone to start in some specific direction, and a bonus for returning to the launch location after completing the timed portion of the task. I have written on this format previously, with relatively little response. Some pilots have pointed out that this sort of task does not prevent cooperating pilots from flying in a gaggle, and they are correct in this. However, unlike our current defined tasks, there is no specific location for goal, no maximum distance to be flown, no great advantage to remain with the gaggle until the final dash for goal. Instead, the rewards for remaining with the gaggle decline continuously over time, while the rewards of finding a better line or a stronger thermal increase, so that the dispersal of gaggles is more likely to occur earlier.

There is nothing sacred about our current scoring systems of the GAP variety. They have evolved over time to suit the preferences of some competition pilots, and have been tuned and tweaked to overcome what are perceived to be deficiencies. The coercive power of gaggles has been recognized for some time, and GAP 2002 and OzGAP are attempts to reward risky leading flying, in order to encourage more early departures and individual choices on the part of the competitors. Unfortunately, the scoring algorithms are now so complex that the optimal choice in any given situation is no longer clear. The rewards of leaving early and leading versus the penalties of flying slower and losing speed points are not knowable while flying on course, and so pilots are usually best served by flying conservatively with the gaggle. As others have pointed out, flying conservatively and finishing at goal each day will almost always result in placing quite high in the contest.

To break out of this rut, new tasks and new scoring systems should be created, rather than ever more modifications and additions to the existing ones. The fact that scoring via the GPS track log is now almost universal makes new approaches possible. Consider the following scheme: on a day with relatively light and constant winds, each competitor is assigned his own individual turnpoint, in some random direction, and at the same distance as every other competitor. Some pilots will have tasks that are mostly upwind followed by downwind returns, others have two crosswind legs, all have tasks of roughly comparable difficulty, if the assignment of turnpoints takes the local topography and abundance of lift into account. Certainly, some competitors will have tougher tasks than others, given the conditions on any given day, but judicious choices of directions and distances can minimize this inequality. Over several days of flying, such factors should average out.

Alternatively, a list of different tasks, including triangles, out-and-return flights, and straight distance flights could be generated each day, along with different difficulty ratings for each. Pilots could choose their own task, basing that choice upon their evaluation of the conditions for that day, their need to accumulate points, and their confidence in being able to complete the task. This approach contains the seeds for protests of unhappy pilots who feel their performance on a given day was not scored sufficiently highly, but since the maximum scores for each task could be pre-computed, such pilots would only have themselves to blame for making a poor choice. A more advanced form of this approach would permit each competitor to create his own task for each day by means of an interactive program. The software would compute the difficulty and maximum score for each such task, based upon established criteria as well as an accurate assessment of the expected weather conditions. It might also suggest alternative tasks with higher or lower scores, based upon the preferences of each pilot.

It could be argued that this approach turns hang-glider pilots into computer nerds, earnestly seeking to find the easiest task with the highest possible score, a situation where an understanding of the program's criteria for assigning difficulty to a task might outweigh one's ability to fly the task. This objection could be overcome by judicious programming, and by making all generated and selected tasks public and permitting pilots to change their selection. Rather than using the present convoluted scoring system, which awards points for departure, for arrival, for "leading", for distance, and for speed, one might provide additional flexibility by permitting each pilot to select tasks that are scored purely on distance flown, or pure speed tasks, or various combinations of each.

To minimize cooperation and collusion between pilots one could require that no more than 5 pilots fly substantially the same task on any day, and that pilots who have flown the same task on a previous day must select different tasks from each other on following days. In addition, pilots electing to fly similar tasks might be required to select start times that differ by at least ten minutes. This sort of open task scheme would require at least 30 terminals for 100 pilots, and sufficient time prior to launch to create and select tasks. At present, contest organizers are not prepared to provide this level of computer support, but it is certainly within the realm of the possible, and should be considered for national-level competitions, and for comps with significant international participation.

The United States has gone from a system of regional contests which qualified competitors for national competitions, to an open Nationals format, a few open competitions which are attended by pilots from other nations as well, and some regional contests that may or may not provide WPRS points. Other nations stage more numerous contests at local and regional levels, or operate in a league format, both of which provide more opportunities for pilots to enter contests and gain experience. This difference is reflected in the number of pilots who fly in contests, and in the quality of the competitors fielded by each country. In the US we have relatively few young pilots, and only a very few of our pilots are competitive at the world level. In part this reflects the declining numbers of active participants in our sport, and in part it is probably due to our system of contests which requires that a successful competitor must devote a large amount of time and significant finances to traveling to and flying in contests.

In particular, it almost necessitates travel to Australia, or South America during the winter months, or to Europe during the summer. A more extensive contest scene in the US might reduce or eliminate this need for travel, but the relatively low numbers of competitors and the expense and difficulty of organizing good contests make this an unattractive option for most clubs and organizations. This dilemma will likely persist unless we can significantly increase the number of new pilots entering our sport, or reduce the cost of staging and attending contests.

Regards,
Heiner Biesel