Oz Report

Updated

FS|GAP|software

https://www.fai.org/news/fs-2022-r18-released

• a small miscalculation in GAP’s launch validity,
• improves result representation in CIVLCOMPS.ORG (EMS) for predefined categories by including the category name in the result title.
• stopped Races with multiple start gates or Elapsed Time races – we now consider the start that gives the pilot the biggest distance considering the permitted flight duration
• tasks where SSS is the first turnpoint (not allowed in FS, you must set a launch turnpoint)
• tasks with one or several turnpoints before SSS

Discuss "FS 2022 R1.8 released" at the Oz Report forum   link»

And a bit of cloud flying

CIVL|cloud flying|competition|FAI|FS|GAP|Joerg Ewald|Paragliding|politics|PWCA|scoring|software|Sté|Stéphane Malbos|Worlds 2022|Pre-Worlds 2022|European PG Championships 2022

From the European Paragliding Championships.

https://www.facebook.com/groups/britishteampara/posts/5550191491710292/?comment_id=5552368234825951

Report by Thomas Milko, from Brazil.

Today there was great confusion early on at the meeting of the heads of delegations. They announced that the scoring system developed by the CIVL, Airscore, had a totally unacceptable failure, providing leading points much higher than the correct values for the pilots who didn't reach goal. They pointed out that the French, who caught this mistake because they were running the SVL scoring system in parallel in the championship,which is not the official scoring program.

The meet organizers didn't want to use fs comp as the GAP 2021 formula used in addition contains known bugs. So we don't have final results on either of the first two tasks. It's been passed on to the jury to decide what will be done. Tomorrow we'll know what the decision is, which is sure to be tricky.

Also heard again a huge discussion about cloud flying as many pilots fly next to the clouds very high. Seriously effects event fair play. The 2022 CIVL rule cannot be operational, besides not having the support of a large part of the most competitive pilots.

It is pretty hard to believe the reports about the rules and scoring coming from the Europeans. A CAT 1 event is only supposed to use tried and tested systems and rules. Airscore, the CIVL's new scoring system, has been in development for 3 years now, but has been almost completely ignored by the paragliding community. So it was quite a surprise that CIVL elected to use it. My quote to Stan just before the event started was "Airscore is the new FAI software, based online, to replace the old FS software. Nobody has any experience with it yet. Expect it to have bugs.". Well, the prediction seems to have been correct.

What can they do about task 1 though. That was marked as Final. We await the jury decision.

As to the cloud flying rule! Well, CIVL did try to use a variation of it in the pre-worlds, but the problems were so great it prompted a meeting chaired by Stéphane Malbos half way through the competition to discuss the failures. It was agreed it would probably not be used in the Worlds this year, and that other methods should be explored in CAT2's first. So it was a surprise to hear that, without trial, they had tried for more cloud flying rule changes at this CAT 1 event.

Lets hope that the event can move past the politics soon, and get back to competition flying at the elite level a CAT1 represents.

I asked Joerg Ewald <joerg.ewald@volirium.com> what he thought about this leading points issue:

Confusion over a mistake in the PWCA scoring specifications that I unfortunately copy-pasted when CIVL adopted the same leading points calculation for paragliding.

Once the dust settled, it was clear that all known programs calculate those weighted leading points correctly. Well, more or less, there are still discrepancies because the PWCA rules leave a lot of room for interpretation.

Discuss "Leading Points" at the Oz Report forum   link»

Flare-timing scored comps on the web

Looking at the data

GAP|Phil de Joux|scoring

Phil de Joux <<philderbeast>> writes:

Flare-timing, the GAP scoring project, is a suite of command line apps. In the past month I've started to show some of the inputs, workings and result in a web app. I had intended to do something like this eventually if it could give pilots some visualization of how a task was scored. In the meantime I found that it was easier for me to check the results seeing them presented on a web page than picking through text files, line by line. That is good for the fine detailed checking but I needed at this stage to see the obvious errors.

Today I've deployed a new version, the 7^th so far, and it is hosted at https://flaretiming.com. There are release notes:

https://github.com/BlockScope/flare-timing/releases/tag/app-view-0.7

One of the big items for this release was sorting out how to parse the various ABS, DNF and DF pilots from the *.fsdb file. I have written up notes on this:

https://flare-timing.readthedocs.io/en/latest/extract-input.html#pilot-group

Each competition is getting its own sub domain and is hosted as a static website. I develop this locally as a dynamic website and switch to picking up the data from relative *.json files at deployment.

Another thing I've added with the latest release is a **Geo** tab. Have a look task 4 for the following comp, the longest ever task set:

http://2018-forbes.flaretiming.com

GAP 2015 Vs GAP 2014

Take the first start time, it helps if you fly slow.

CIVL|GAP|Jon "Jonny" Durand jnr|Jonathan Dietch|Zac Majors

Chippy writes:

Had the 2015 Sylmar Spring Air this weekend. Nice turn out, with Jonny, Kraig, Konrad Heilman (both Konrad and Kraig did not compete but did fly the course), Zippy and most of our regulars. Discussion after the scoring and awards was about leading points and why they should be included. I decided to enable them the best way possible using the scoring script you developed. However setting the parameter in your scoring program to use GAP 2000 wasn't working.

http://www.soaringspot.com/en_gb/sylmar-spring-air-2015/results/open/task-1-on-2015-06-06/daily

So I downloaded and scored the top three in FS 2015. Using OzGap 2005 in FS 2015 pretty much mirrors OzGAP 2005 in SeeYou, using GAP 2000 awarding departure and arrival position points, or the new Official CIVL GAP 2015 award squared leading points and arrival position points, changes the 2^nd and 3^rd place competitors, because of those parameters, as expected.

Live tracking here.

My response:

FS 2015/GAP 2015 uses the squared version of leading points plus arrival position points. What happens if you use GAP 2014 in FS 2015 instead?

Chippy writes:

Re-scored the comp again in FS after seeing that I needed the legacy scoring pack.

After getting all of the users in, I used the following formulas:

2014 Official CIVL Hang Gliding with squared leading coefficient value (GAP 2015)

2014 Official CIVL Hang Gliding without squared leading coefficient value (GAP 2014)

OzGap 2005

From what I can see, using the squared leading coefficient value rewards the pilots that lead (start) early, and also rewards pilots that lead but do not make goal. This scenario was being discussed in the LZ most vigorously by Konrad and to some extent Jonny. I think they are in favor of this.

When not using the squared leading coefficient value, pilots that lead and make goal get the same value or close to it, but those that lead (or took the first start), but did not make goal are the most penalized. The further away from goal, the larger the penalty.

In the examples, you can see Ken Andrews who was leading early with Rob Burgis but missed one of the turnpoints is rewarded with 47.1 lead points when using the squared value. Those leading points are really stripped away when not using the squared value and distributed to the pilots that took the second start. Maybe this is the way it should be, because at the point Ken missed the waypoint, he then failed to be leading.

OzGAP, obviously only rewards pilots that lead and make goal.

The reversal is completely related to the squared value. Since it cannot be removed using GAP 2015, I had to use GAP 2014 to really see it.

You can also see 3^rd and 4^th places swap (Ron Kienan and Jonathan Dietch). Of course, Jonathon likes this because it puts him in 4^th.

Since the comp was small, maybe the points are compressed and have a greater effect.

Mostly, our comps are small and we generally use one start and therefore, leading points are not required/needed. This time we couldn't get everyone off in time and multiple starts were needed. At this point, most pilots here "expect" leading points to be enabled. And currently there is not a way to do that in SeeYou. Pilots I'm hearing from want leading points for pilots that lead out but also for pilots that lead, but do not make goal.

Some might say, what's the point, you didn't make goal.

My response:

Why exactly should Rob be rewarded so handsomely for taking the early clock? Zippy flew much faster than he. Was Rob somehow disadvantaged and Zippy and Jonny advantaged because Rob took an earlier clock?

This comparison tells us nothing about whether anyone was pulling anyone else by getting out just in front of them. It is a matter of personal preference whether one wishes to give more or less points for speed or for taking the earlier start clock.

Discuss "GAP 2015 Vs GAP 2014" at the Oz Report forum   link»

Only the latest versions

CIVL|Joerg Ewald|FS|PWCA|GAP|PWC

Joerg Ewald «Joerg Ewald» sends:

http://fs.fai.org/2013/06/17/fs-2013-released/

Unlike previous editions, FS 2013 comes only with the most recent versions of both CIVL’s and the PWCA’s official scoring formulas: GAP2013 and PWC2013.

If older versions of either GAP or PWC formulas are required, these are now available with the optional FS Legacy Formula Pack, which can be downloaded and installed from the same link above.

Discuss "Latest version of the CIVL scoring program, FS 2013" at the Oz Report forum   link»  

GAP explained

Lots of text, number and charts, how much is understood?

GAP

1. Leading points are not leading points at all. The name is misleading and has mislead pilots for a decade. If used, they need to be called something else. Departure points are no better.

2. The nominal distance isn't what you think it is and to determine a nominal distance that divides days where the average pilot goes less than the normal task distance from days where the average is equal to or greater than the expected normal task distance requires greater nominal distances than have normally been used. It also requires a small amount of algebraic manipulation.

3. The GAP 2002 document states that: "GAP is designed to reward the pilots who make the decisions." GAP does not do this.

4. The three different versions of GAP produce about the same results. Arrival time points have the advantage of being understandable and actually correctly named.

5. I doubt that anything that I have said in these articles will lead to any changes in how things are done, except changes I make in competitions where I have the power to make such changes.

6. It doesn't really matter as the best pilots will come out near the top no matter which scoring system we use. If we used a completely different scoring system maybe that would change who the very best pilots are (as the previous statement is a tautology).

We dig a bit into the muck

GAP

(Switzerland)

When trying to explain the GAP scoring system to pilots, the developers of the scoring system and others have tried to stick to pretty pictures, because frankly the equations which our computers use to draw these pictures can get pretty ugly and without the pictures they don't make much sense. There is no real reason why one equation would be chosen over another or why one set of coefficients is the right ones, other than the fact that the pictures that they draw appear to be reasonable.

You'll find an excellent discussion of the GAP scoring system here along with its implementation in the scoring program FS here. But there is still this idea that the GAP system is somehow encouraging you to fly out in front. This document states:

The GAP scoring system rewards the pilot who makes his own decisions and stays in front of the others. Waiting on takeoff for the other pilots to fly, then following to be safer and faster, is a less valuable tactic in terms of this scoring. With previous scoring systems your best tactic to recover points over your opponent was to start a few minutes after him, then catch him. His best tactic was to wait for you and fly with you. Everybody was always waiting on takeoff because everyone was following the same tactic! With this scoring system a good tactic to catch up points on your opponents is to start early and fly fast. This way you force your opponents to take the risk of being an early bird with you, or wait for more pilots to be in the air for safer (scoring wise) flying.

As I have pointed out in the last article, I would take that statement with a large grain of salt. Flying fast using pilots out in front who took an earlier start time and getting to goal is almost always better than going early and trying to grab the points for being out in front and getting to goal early.

So where do leading points come front? Let's look back at a chart from the last article:

Case 1a.

If you formed a triangle with the blue line as a hypotenuse, the x-axis as one leg and the y-axis on the right side as another you would enclose the area under the first pilot's flight curve (actually a straight line). As here:

You can in a similar manner find the area under the second pilot's curve (which includes the little rectangular area bounded by the Y-axis from 0 to 10 minutes and and the x-axis from 0 to 100 km). See here:

Because the first pilot was faster than the second pilot you can see that his area under his curve is less than the area under the curve of the second pilot. It is the relationship between these areas that it used to define the "leading factor" which is multiplied by the available "near the front" points to see how many "leading" points the pilot gets. Now the pilot with the smallest area (whether they make goal or not) gets 100% of these points. The pilots with bigger areas get a smaller percentage of them.

The leading coefficient for a given pilot is defined as the area under his flight curve in seconds and meters divided by 1800 and by the task length in meters. The equation that uses these leading coefficients to find the "appropriate" multiplier is found in the document linked to above on page fifteen (out of eighteen). Also:

Of course, not all flights are as idealized as these flights so the curves can get pretty messy and we have to come up with a bunch of additional rules to handle them. In fact the pictures that display the curves can lead us astray when it comes time to actually figuring out how this works. For example:

And this one:

I hope that I've given you a taste of the idea that perhaps "leading" points are maybe not what you thought they were. They certainly aren't easy to understand and they certainly aren't leading.

The leading coefficient calculated for the first pilot, a pilot who flies 100 km to goal in two hours (a pretty darn fast pilot) is equal to two (notice how in this idealized case it is just the time it took the pilot to get to goal). You can see the relationship between slower or later pilots with their larger leading coefficients by looking at this chart:

You should be able to find three of the previous idealized cases on this blue line.

As the other pilots' leading coefficients increase relative to the first pilot's, their multiplier times the available leading points decreases. You will find additional charts that map out this relationship for different minimum leading coefficients here. Go here if that direct link doesn't work. Check out Appendix C, page C-8. This Appendix is an excellent resource for understanding GAP and the equations that are used for scoring.

I'll discuss leading points and compare them with departure points more later, but I'll first take a break and look at some GAP parameters.

Discuss "So what are these near the front ("leading") points?" at the Oz Report forum   link»

The GAP scoring system

Sketch out how the points should be allocated between distance, speed, etc.

(Lausanne,+Switzerland)

GAP|Gerolf Heinrichs|PG

What does the GAP 98 document say?

GAP scoring was developed for FAI by Gerolf Heinrichs, Angelo Crapanzano and Paul Mollison. The idea was to get a fair scoring system that was easily adaptable to any competition everywhere in the world, both for hang gliding and paragliding, with a philosophy easy to understand for the pilot, regardless of the mathematical complexity. To compare different tasks within the competition and to adapt the scoring to hang gliders or paragliders, different flying sites, pilot’s level and task philosophy, before the competition the meet director sets some parameters.

Is GAP easy to understand? We'll see if this is true as we proceed through this series.

One of the major scoring decisions that these three made was to apportion the points available to all pilots depending on whether they got to goal or not, depending on how fast they flew, depending on how early they started and in what order they arrived at goal. This is how they decided at first to apportion the 1000 available points, points that they thought each day should be potentially worth:

These essentially hand drawn curves express the idea that as more pilots make it into goal more points should be available for speed to give more reward to the pilots who went faster and less to distance, as getting to goal apparently wasn't that hard and getting distance was therefore not such a big deal. Arrival position and departure time points are a portion of the speed points (actually a portion of what's left over after the distance points are subtracted from the 1000 points), providing, it was hoped, some small incentive to go as early as possible, offsetting the risk of being out in front, perhaps alone.

Note the speed line at zero percentage of pilots at goal actually goes to zero points, although the graph doesn't show this. Arrival and Departure points also go to zero if no one gets to goal. This leaves us with 900 points for the pilot who goes the furthest on a day no one makes goal, which may or may not be a good thing. Is there a good reason to devalue such a day, other than it looks that way on the chart?

In order to use this pretty picture for actual scoring they came up with equations that matched the curves (that's right, the picture came first, then the scoring equations). I've had Excel chart a picture based on the equation that they came up with:

Now one might quibble that the pilots who landed out shouldn't get any points, but unfortunately that would encourage meet directors to call tasks that would be considered by most pilots to be too short.

Their work wasn't done and as they saw how the scoring system worked they decided to make some changes. Next came a revision to GAP 98 that eventually became GAP 2000:

In this iteration speed was now relatively less important as more pilots made it to goal and the pilots who didn't make it to goal didn't get hurt quite as badly. Also there were slight changes to Departure points with its multiplier increasing from 1.2 to 1.4, upping their value a bit. This is essentially the distribution of points that we've been living with since 2000.

GAP, explained

More formulas

GAP|scoring

Appendix C:

http://www.pwca.org/sites/default/files/Attachement-page/PWCA Competition Rules 2012 - Finalversion.pdf

More to come as I look more closely at scoring and scoring equations.

GAP, explained

A bit of the scoring formulas

http://fs.fai.org/trac/wiki/ScoringFormulas

The most current documentation on Gap as implemented in FS. GapGuide-2011-v1.pdf

The original GAP 2002 explanation from Angelo. GAP02_en.pdf

Another explanation on GAP written by Nigel Leigh (NZ). Thegapguide.pdf It's a good guide that explains the features of GAP quite well for pilots.

Discuss "GAP, explained" at the Oz Report forum   link»

GAP documentation

What's inside FS

Agust Gudmundsson|CIVL|GAP

Agust Gudmundsson <<ag>> writes:

CIVL has published new document explaining the GAP2008 scoring formula and options. The document is available at http://www.fai.org/hang_gliding/competitions/gap (see bottom of the page). The document is a great improvement over previous explanation documents.

This can be improved even further, but for that we need your assistance. Please contribute by sending CIVL or me directly any comments, suggestions and corrections. We plan to have improved version in few months.

You can download the document here. Be aware that GAP 2008 is not actually a specification of a formula for scoring (unlike GAP 2000, GAP 2002, and OzGAP 2005), but rather a generic name for the capability of creating a scoring system from parts of these scoring formulas. It has already been shown that you can get into big trouble if you are somewhat careless in how you construct your scoring system.

The document linked to above is quite nice and you'll also find the links to the documentation for GAP 2000 and GAP 2002 on the same web page. Anyone interested in scoring is encouraged to look at actually how the scoring is done.

My scoring program running in SeeYou uses basically the same algorithms and gets the same results (for the same settings) as does FS. So this documentation just gives you a better idea of what is going on behind the scenes in all the scoring programs that use the GAP formulas. 

SeeYou

I have implemented GAP 2000 and OzGAP 2005 in SeeYou

GAP|PG|SeeYou|US Nationals

These implementations of our hang gliding and paragliding scoring systems can be used with SeeYou to score meets. I haven't written a manual on how to do it, although there is some on-line help on SeeYou competition at www.naviter.si. I'll be using the OzGAP 2005 implementation to score the US nationals and the Santa Cruz Flats race (backed up with Race 2003).

This is where you can find GAP2000 (scroll down to my attachment). And OzGAP2005.

Discuss SeeYou at the Oz Report forum

GAP

Maybe a picture or two will help

GAP Parameters - the graphs

GAP

https://OzReport.com/data/GAPtest.xls

You can play with the spread sheet yourself to see what the story is. Here's what I get with a task length of 60 miles, nominal distance of 75 miles:

What the graph shows is that the pilots have to average only forty nine miles over this task to make the task worth 1000 points (we're ignoring the time validity factor at the moment).

So even though the task was shorter than the nominal distance (quite a bit shorter, I should add), the task was easily completely valid. You'll want to use the spread sheet linked to above to see how the nominal distance and the average distance interact.

Now, what about the time validity? Well the time validity falls off much slower from 1 than the distance validity.

If no one makes goal, the time validity factor is 1. So, yes, you sure can have a 1000 point day if no one makes goal, as long as the average distance flown is equal to or greater than the amount given by the distance validity factor equation graphed above.

In this case if the fastest pilot takes three hours or longer to get to the goal, then the time validity factor is 1. The nominal time is three hours. The nominal time should be set to the average distance that you expect for the tasks called during your competition divided by your expected speed for the winner on the average day (25 mph). Days where you blow down wind in a 15 mph wind for the expected task distance will be devalued, as they should be.

https://OzReport.com/forum/viewtopic.php?t=1930

http://www.metamorfosi.com/GAP2002_en.htm

GAP

It looks like we didn't set them too high after all

(Quest Air)

GAP|Quest Air

We've learned a lot about GAP parameters over the last week of discussion and debate. You can find out more here: http://ozreport.com/forum/viewtopic.php?t=1930. It turns out that maybe we set the value for nominal distance too low at sixty miles and 2.5 hours.

The values for the rigid wing pilots were 75 miles and three hours. Seems to have worked out fine.

http://events.dowsett.ca/flytechistory.

Discuss GAP at the Oz Report forum

Nominal Distance

Does anyone know what this means?

GAP

Distance Validity := 2 * (Number of pilots flying * average distance flown - Number of pilots flying * Minimum distance)/(Number of pilots flying * ((1 + Nominal Percentage at Goal/100) * (Nominal Distance -Minimum distance) + (Nominal Percentage at Goal/100 * (Maximum distance flown - Nominal Distance))));

Discuss "GAP nominal distance - the formula" at the Oz Report forum   link»  

GAP and more

Who really understands the GAP parameters?

David Glover|Mark Dowsett

David Glover «davidhglover» writes:

I must have missed seeing the "half that value" recommendation, perhaps because it isn't written any where that I'm aware of. For example, the Race Programs makes a one sentence statement which does makes it seem like nominal distance is minimum task distance for 1000pt day.

Quote:

"The nominal distance is the minimum task distance that should be worth 1000 points." Then there is a box to input a number which is auto filled by the Race program with 43.5 miles.

The Race default is 43.5 miles and 2.5 hours, 30% and 10km min distance.

According to Mark Dowsett's spreadsheet analysis of previous Flytec meet the average distance flown by competitors at Flytec for the last 5 years in all flex wing tasks is about 48 miles. This figure is recommend by one of the GAP authors (30 miles) is 37.5% lower than the average distance flown.

Discuss GAP and more at the Oz Report forum

GAP SS

You can play with GAP parameters yourself

GAP

Download it here. The nominal distance and time validity factors.

CIVL|CompeGPS|competition|cost|Dennis Pagen|finance|GAP|insurance|Ivan Twose|NAA|sailplane|scoring|SeeYou|site|transportation|USHGA|Worlds

The actual statement was: “The USHGA, in fact, inhibits competitions in the US (as well as helps).”

I don’t think that Steve and I have any argument here. He mentions a number of ways that the USHGA inhibits competitions and then goes on to come up with even more ideas than I came up with how the USHGA could help to promote competition. I think we are just arguing about words.

Let’s look at it this way. The USHGA has a competition budget of about $12,000/year. It takes all this money and gives it to the NAA or to fund our CIVL representatives. No money goes to support any of the ideas that Steve proposes above. All of the money goes to support a system that allows the very top US pilots to fly in the Worlds or set World Records. None goes to benefit the 99.9% of the membership that might actually take part in a local, regional, or national competition.

Now if there weren’t any good ways to spend that money that would encourage more competitions, then, okay, let Dennis Pagen have his playground. And let us select few get our opportunities to fly in the Worlds and set World Records.

But think about it. The USHGA could fund a seminar where Tim Meaney could train people interested in being score keepers. It could pay Ivan Twose for a license for CompeGPS for all US competitions and specify the changes that are needed to get CompeGPS and Race to work well together.

The USHGA could finance the development (to be paid back from competitions) of a version of SeeYou (the program now used for almost all sailplane competitions) that would incorporate Race, GAP, and other hang gliding scoring systems.

As Steve says above it could finance the mentoring of potential meet directors, scorekeepers, meet organizers. It could send around a meet organizer consultant that would help potential meet organizers get started.

It could eliminate the sanction fee, the deposit (under certain guidelines), and the site insurance fee for meets. All these would reduce the costs to pilots for entering meets. It could encourage meet organizers, by providing a subsidy, to allow new competitors to come to a competition at a reduced fee.

I’ll bet the major inhibitors keeping pilots from attending competitions are the difficulty in getting the time off work to attend and the costs: cost of getting to and from the meet, cost of lodging during the meet, cost of retrieval, and entry fee. More regional meets would cut down on transportation costs. If the USHGA provided support to meet organizers so that they could coordinate getting lodging and retrieval support to competitors, that might encourage more to come to meets. Using virtual goals is one way to cut down the costs that impact the entry fee.

The USHGA might even do a survey among its members to find out if they want more competitions, and what it would take to make them more successful.

All these things are possible if all our competition money wasn’t going into one pocket.

Discuss "USHGA – the competition budget" at the Oz Report forum   link»

Alessandro "Alex" Ploner|Alex Ploner|Angelo Crapanzano|cloud|collision|competition|Europe|FAI|gaggle|game|GAP|GAP 2002|Gerolf Heinrichs|GPS|Ivan Twose|midair collision|power|Richard Walbec|scoring|Thomas "Tom/Tomas" Weissenberger|weather

Angelo Crapanzano <angelo@metamorfosi.com> writes:

The pre Europeans is over and it was a good competition. The weather was varying but good overall and Richard Walbec was a very good meet director plus all the crew was friendly and willing to help (thanks a lot to everybody!) Unfortunately we lost what we discovered had been the best day ever in Millau: in this day a local pilot got up to 4700 m (15500 ft) while most of the competitors were blocked on takeoff by an unpredicted strong side wind.

This day Federico Bausone waited, ready for takeoff, for over one hour and fifteen minutes (with nobody pushing, of course), then the task was cancelled because there was not enough time for all the pilots to get the start gate. Fifteen minutes later the wind started changing and in half an hour it was perfectly straight but most pilots already packed the gliders :-(

As far as I know Millau was, both last year and this one, the only international competition using the GAP 2002 at "full power" (i.e. with the Leading bonus calculated from the actual tracklog of each pilot). Despite the scoring explanation was published besides the daily score, there were several questions on it; probably because pilots like to ask instead of reading and possibly because they know I like to answer.

I had the feeling pilots did like the idea of the Leading bonus when they exactly understood how it works (to know more have a look at my webpage www.metamorfosi.com and click on the GAP icon). On the first day Betiño did perfectly show how the leading bonus works: he started 15 minutes before the first gaggle, flew always on his own well in front of everybody but landed 4 km short of goal. Despite he was not in goal Betiño got the biggest leading bonus on that day to reward his early flying.

On another day Tom Weissenberger and Alex Ploner were for most of the task in front of the leading gaggle trying to escape but, just before goal, got stacked and the leading gaggle flew above them to goal. Alex and Tom got goal shortly afterwards with, of course, a worse time and less speed points but the leading bonus allowed them to reduce the loss of points as a reward for leading for most of the flight.

During the comp there was some discussion about start systems. Except on the last day, where it was set a pure race with single start for everybody, Richard always set a multiple start (mostly with 15 minutes interval but once 10 minutes and once 30 minutes). There was also a debate (mainly between Gerolf and me) about using a multiple start for everybody, except the first 20 in the general standing forced to get the central start.

This proposal come out because many top pilots prefers races but it was not felt safe to have 110 pilots taking the same start. Top pilots like races because it's more fun, they can fly in the same conditions, it's easier to know how one is doing compared to the others and it's also easy to control the opponents because everybody has to start at the same time.

I was strongly against this proposal because to set a different start system to some pilots compared the others is almost like to set a different task. All pilots in a competition are supposed to play the same game with the same rules but if we force the first 20 pilots in the total ranking to take a single start while the others can chose, we could end up with the same tracklog giving different points depending if the pilot was placed 20^th or 21^st.

To make it clear let's make an example with five starts every 15 minutes, from 14:00 to 15:00, but top 20 pilots have to start at 14:30. If a pilot placed in the first 20 is late and only manages to start at 14:45 he will be scored as he started at 14:30 while if he's not in the first 20 he will be scored as he started at 14:45. This means the same tracklog would bring to quite a different score and, in my opinion, this is not acceptable: all pilots are in the same competition and must play with the same rules.

Somebody says the priority on takeoff for the best scored pilots is already a different rule but I don't agree: it's just a "courtesy" - on the ground - to the pilots which are fighting for the top positions. Once in the air every pilot must follow the same rules!

Of course I do perfectly agree races are a lot of fun and easier to understand for everybody but, unfortunately, the risk of a midair collision gets quite high if there are too many pilots and there is an easy to get cloud base. The multiple start was invented after at the preworld in Ager '94 we ended up with 180 pilots waiting for the tarp in the same thermal for over half an hour: we were lucky enough not to have a midair collision but it was a nightmare!

Unfortunately it looks that pilots and organizers are too conservative and always want to do the same thing: in Brazil only races while in Europe only multiple starts but, in my opinion, in both cases it's not the best thing to do.

Let's analyze our possibilities: we can use three different start cylinders:

- start on Exit

- start on Enter without achieving the turn point inside

- start on Enter plus achieving the turn point inside

we can use four different in flight start time systems:

- Race start

- Free start

- Multiple start

- Open start. These, combined, give us 12 different options: some bad some better but, several, simply different.

Let's speak first about the start cylinder.

- Start on Exit Most pilots and organizers in hang gliding are used to the Start on Exit: normally it's a 5 km radius start cylinder centered on takeoff, where pilots have to be inside the radius just after start time. Pilot start time for scoring purpose is thus last exit from the cylinder (rounded to the previous start time interval which is usually 15 minutes). This system will somehow reduce the crowding only if the radius is quite big.

Unfortunately there is always a single optimum point where all pilots are likely going to met. With the Start on Exit pilots have to make a GoTo to takeoff then, after start, make a GoTo to the next turnpoint or activate the route which, depending on the GPS used, may require to press several buttons.

- Start on Enter without turnpoint inside this is mostly used in paragliding: Start is usually the first turnpoint and the pilot has to be outside of the start cylinder just after start time, then goes in and can immediately go for the next turnpoint without going to the FAI 400 m cylinder at the centre. This is not much different from the Start on Exit because there is still an optimum point where to make the start. However there are some problems calculating the correct task distance (Takeoff-StartCircumference-TP1) which shall be used instead of the distance calculated by the GPS (Takeoff-StartCentre-TP1).

- Start on Enter plus turnpoint inside this is the system we mostly use in Italy since this season and was often used at the pre Europeans in Millau: usually the start is the first turnpoint, with a big enough radius to have the start at a reasonable distance from takeoff. The pilot has to be outside of the cylinder after start time then has to go to the FAI 400 m cylinder at the center.

With this kind of start the pilot has to remember the start radius (which would be different on each day) but will only have to make a GoTo to the first turnpoint or just set the route. In this case pilots can, theoretically, spread along the whole circumference because any point of the start circumference is at the same distance from the turnpoint at the centre. Of course there would be better places than others depending on ridges, thermals and wind, but is undoubtedly more likely to get the pilots spread than with the other systems.

Giving this analysis, except possibly on some quite particular cases, the Start on Enter plus turnpoint inside should always be the preferred start cylinder: has no disadvantages, it's safer, gives more flight options to the pilot and it's the easier to handle with the GPS.

Now let's go to the start time systems.

- Race start this is for sure the simplest one: there is one single start for everybody. It's a lot of fun, everybody knows if he's doing good or bad and it's easy to understand both for pilots and spectators (are there any?). Pilots will fly in the same conditions and the fastest one wins. The pure race reduces the chance of a lucky start time but also eliminates the possibility to choose a better one.

This system gives to a pilot who wants to recover no options except to try to escape from the first gaggle by flying very fast and gives the pilot which is leading the competition the possibility to better control his opponents. The race will put all pilots at the start cylinder at the same time. This could be extremely dangerous if there are too many pilots and especially if cloudbase is too easy to achieve.

Of course there are systems to reduce the crowding which, in a proper day, could allow to make a safe race even with lots of pilots: should be a good day with good thermals and ceiling to spread pilots vertically, but little no clouds, start cylinder should be on enter plus turnpoint inside and wind should be at an angle to the Takeoff-StartTP direction to help the spreading along the start circumference. Start should also be reasonably far from takeoff (about 15 to 20 km) to reduce the crowding because some pilot would already bomb out and others would be late or low. Taking these expedients it would be possible to make a Race even with lots of pilots.

- Free start looks very simple and easy to handle but is totally unused: with this system the pilot start time is the last time he crosses the start circumference. Top pilots would need to wait forever (and bottom pilots would wait for them to go) because the best tactic is to wait for your opponent to start, then follow him after a few minutes and catch him. Being the tactic the same for everybody, all pilots will wait until it's almost too late to complete the task. This system is unsafe because we'll end up with lots of pilots waiting for a long time. Moreover we'll also waste the first part of the day and everything will end up later.

- Multiple start it's right now the most used system in hang gliding. Normally start interval is 15 minutes and there are 3 to 5 starts (half an hour to one hour). Pilots crossing the start circumference are scored as they started at the previous start interval. It reduces the crowding by spreading the pilots into several starts. Multiple starts also allow good pilots to play different tactics from his direct opponents to recover some places: one could start earlier and fly on his own to get more leading points or one could start later and fly fast by using the pilots in front. This system proved to be well suited on most conditions but this doesn't mean it should be the only one to be used.

- Open start this used by sailplanes and is a good option in some cases. Let's say start is open from 14:00 to 14:30. Every pilot starting while the start is open will be scored according to the last time he crossed the start circumference, while pilots starting after 14:30 will be scored as they started at 14:00.

With this kind of start is likely that several pilots will start as soon as they get cloudbase but the top pilots cannot afford to have their opponent starting just after so would likely wait for the start closing (14:30 in the example). If a good pilot has to recover he could risk to start on his own at start opening (14:00 in the example) and go for the leading bonus. In my opinion the start should, almost always, be 30 minutes long: shorter it would be useless and longer would oblige the top pilots to wait too much.

This kind of start is very useful in case of too much crowding (low and easy to get cloudbase for example) because most pilots would go away as soon as they are high and only the top ones have a reason to wait (but they are the ones which could handle the situation better). The Open start gets, practically, the same results of "forcing" the top 20 pilots to make a race (but using the same rule for everybody) but still gives a top pilot which wants to recover the option to play a different tactic from the others.

Conclusions: analyzing start cylinders format we ended up with the start on enter plus turnpoint in centre to be clearly better than the other possibilities but, speaking of start times, there is not a better one. We have to exclude the Free start but Race, Multiple and Open start have advantages one over the other depending on clouds, ceiling, wind and flight area. In my opinion it's simply wrong to use always the same system: all of them should be used depending on which one is better in the given day. They also test different pilots abilities and that's why, possibly, all these start systems should be used within the same competition: as a minimum it's more fun than playing always the same game :-)

Speaking about testing different pilots abilities, I've designed a system to be able to score an X-MAX task, with GAP scoring, within a normal competition. This way we'll test new pilot's abilities: to be able to find the best route and to correctly judge the day and himself (please don't tell me it's a matter of luck because I bet would be always the same in front…). Ivan Twose had not yet enough time implementing this but I'm sure it would be tested in Italy before the end of the season and would be ready next year.

Discuss competition formats at OzReport.com/forum/phpBB2

Discuss "Competition Formats / Start times" at the Oz Report forum   link»

GAP|GPS|landing|scoring|site

I asked Angelo Capanzano about this problem with the displayed speed in the Race scoring program output. He wrote:

If you use the last versions of RACE2003 (this is in beta and the Race web site says not to use it) and Compe-GPS the speed shown is the average speed over the Speed Section (from Start to Arrival). If you use old versions of Race you may get the average speed from takeoff or somehow wrong results (maybe different in Race and Compe).

In any case the speed score is calculated using the time from Start to Arrival and is always correct, so it's more a kind of "cosmetic" problem.

Start is last crossing of the Start circumference (when entering or when exiting, depending how the Start is defined) and is normally "rounded" to the previous imposed Start time (usually every 15 minutes). Arrival is when first crossing the goal line (virtual or physical) or the circumference of the Arrival cylinder (centred on a turnpoint or goal).

Older versions of RACE and Compe were using the correct Start and Arrival times but the wrong Speed Section distance (for example from takeoff to goal, forgetting the Start and eventual Arrival radius), thus leading to wrong average speed.

GAP scorings use the speed section times to calculate Speed Points thus results are always correct, regardless of the Race or Compe version used. GAP scorings use the total distance (from takeoff to landing or goal) to calculate distance points and results are always correct too.

Discuss Race at OzReport.com/forum/phpBB2

Discuss "The speed value in Race" at the Oz Report forum   link»

Angelo Crapanzano|Christian Ciech|CIVL|competition|Cristian Deacu|Davide Guiducci|Flavio Tebaldi|Florida|GAP|Giovanni Vitola|Manfred Ruhmer|Mino Bricoli|Olivier Burghelle|scoring|triangle|Tullio Gervasoni|weather

Angelo Crapanzano <angelo@metamorfosi.com> writes:

The X-MAX ended a few days ago. Despite quite a bad meteorological forecast we flew three days out of four.

First day:

1) Davide Guiducci (Litespeed 4) 80 km closed triangle

2) Mino Bricoli (Topless 3) 87 km 4 km short of goal out and return

3) Giovanni Martinuzzi (Talon) 63 km closed out and return

Second day:

1) Manfred Ruhmer (Laminar MR14) 113 km closed out and return

2) Angelo Crapanzano (Litespeed 4) 87 km closed out and return

3) Flavio Tebaldi (Laminar MR 4.2) 75 km closed out and return

Third day:

1) Christian Ciech (Laminar MR14) 180 km closed triangle

2) Manfred Ruhmer (Laminar MR14) 170 km closed triangle

3) Tullio Gervasoni (Litespeed 4) 164 km closed triangle

Manfred didn't fly on the first day because of the jet-lag coming back from Florida (and Christian flew quite badly probably for the same reason), but it was enough for him to win the X-MAX trophy (which only considers the best two flights).

With some of our problems with the Italian AeroClub resolved, this year the competition was also valid as second category event (all days valid) and the final results are:

1) Davide Guiducci (Litespeed 4) 2432

2) Angelo Crapanzano (Litespeed 4) 2067

3) Maurizio Bignami (Laminar MR 4.2) 1941

4) Mino Bricoli (Topless 4) 1874

5) Cristian Ciech (Laminar MR14) 1788

The CIVL president, Olivier Burguelle, was at the competition… and did help as driver too :-) He was very interested in the X-MAX competition format and quite satisfied with the results (practically without any need of organization we got long distances for the weather, no gaggles, no pressure on takeoff, pilots pushed to use their heads, almost no retrievals and everybody happy) and would like to push the idea in the Paragliding World Cup too.

Because of this interest, I've found a way to score the X-MAX format with GAP (or any other classic scoring program) and, very soon, it would be possible to score an X-MAX task within a standard competition using an X-MAX slightly improved formula, to consider flying areas different from Laveno. We'll live test everything at the end of May in an Italian League competition (again in Laveno).

Discuss competition and competition formats at OzReport.com/forum/phpBB2

Discuss "XMAX" at the Oz Report forum   link»

Alessandro "Alex" Ploner|Alex Ploner|Betinho Schmitz|Christian Ciech|cloud|Fantasy of Flight|flight park|Flytec Championships 2003|gaggle|GAP|GAP 2000|J.C. Brown|Jim Lee|Johann Posch|landing|Malcolm Jones|Mike Barber|Oliver Schmidt|Paris Williams|Peter Gray|Richard Walbec|SeeYou|Terry Presley|Wallaby Open 2003|weather

http://www.wallaby.com

First of all, the task:

It’s first twenty miles to the west to Rockridge and highway 98, then east to the Fantasy of Flight to keep us from going over downtown Winter Haven, then southeast to Chalet Suzanne on highway 27 near Lake Wales, then back to the Ranch. It’s a 70 mile task, with a substantial east wind of 10 mph predicted.

The new wrinkle is that the start circle is centered on the Rockridge and highway 98 intersection and is fifteen miles in diameter. The idea is that you begin the contest by entering the start circle instead of leaving the start circle. The reason for using such a start circle is to allow pilots to spread themselves out along the circumference of the start circle, all routes to the first turnpoint being equal from that circumference.

Now theory is one thing and practice is another. With a good wind component, only one point is optimal. With other pilots congregating in one location, they form an attractive nuisance and soon there is a party in one spot.

Still it’s quite manageable with twenty pilots looking to take the first clock at 1:30. Which reminds me of another innovation.

At the task committee meeting this morning Betinho mentioned that he wanted more time between start times to cut down on pilots waiting around for a later clock and then catching people in front of them (can you say Johnny Durand, Jr.?). Of course, we were using GAP 2000 at the Flytec Championship (and we are here), which has the minimal early departure/arrival bonus points so of course pilots are going to try to catch others from behind.

So I immediately suggest a half hour and, whoa, it is accepted. This will make it harder for pilots taking a later clock to catch the earlier guys and keep them from getting that extra added advantage without having to pay for it with reduced starting time points.

The task committee calls for three start times, although what they really want is a race, one start time. The compromise with three start times has to be made to allow the flight park time to get everyone in the air. Actually we could call a start time at 1:30 hours after the launch window open time, because that is enough time to get everyone in the air and at the circumference of the start circle, if people start launching at the launch window open time (they never do). So maybe we will have a race start soon.

The task committee consists of Richard Walbec, Betinho Schmidt, Mike Barber and I. But JC Brown runs the meeting and directs the discussion. Peter Gray and the goal crew are there also and Malcolm Jones hangs around the edges. This is quite a difference format (and dynamic) than Jim Lee , Terry Presley and I meeting in my trailer looking over the SeeYou task map as we did last week.

Mike Barber is a high anxiety person. He has extreme difficulty making a decision (especially one related to calling a task four hours in advance of the actual task). He wants to call off the decision until the last moment. He also is very distrustful of weather forecasting (even forecasts four hours out). The funny think is that he also is compulsive looking and listening to the weather forecasts. His anxiety infuses the task committee.

Mike is afraid of the possibility of over development, like the OD we had yesterday. I’m looking at the FSL MAP models and the BLIPMAP. I guarantee to him that there won’t be any OD during the task.

I compare the fear of OD with the actuality of higher winds than the models predict. It appears that the winds are stronger out of the east than the 10 to 12 mph that the models show will be the state of things later in the day. We later get a “sounding” from Paris Williams (taken, who knows how), with the east winds stated to be 20 mph. I’m concerned about the prospect for higher winds (as we are right on the cusp of where it makes a big difference), and I’m looking at the windcast showing a broad convergence and lighter winds to the west (as we had for many days at the Flytec Championship).

The task committee gets through the process and calls two tasks (we never called an alternate task at the Flytec Championship) just to handle all the anxiety (well, okay, that is a bit overstated). Then later, just before the pilot meeting, based on Paris’ “sounding” we call a third task to the west (I’m good with that) and make it the primary.

Now the pilot meeting is at 11 AM, which I consider to be about an hour too late, because it doesn’t give us time to get to the staging line and get ready to go before the launch window opens. I give my little weather spiel (know any good weatherman jokes?) and before we know it is too late to get to the launch line in time to get suited up for the 1 PM start window.

The three start times now become two start times (which is cool with me, as I go for the race concept) of 1:30 and 2 PM. I’m thinking that this is alright. Fifteen minutes are the launch window opens with no pilots launched (did I mention this?) and Bo on hold from JC Brown (talk about last minute anxiety) we call the secondary task (which earlier was the primary task) (see above).

They let the task committee stage in the front, but I get off early anyway, so I don’t clutter up the neighborhood. The lift is light to moderate (200 fpm) the southeast of the Ranch and we just hang out in the big smooth stuff to cloud base at 4,000’ slowly moving our way to the west staying just under the clouds.

Just before the start time at 1:30, there are twenty plus pilots at the edge of the start circle. There is light lift at cloud base so we are all just hanging near the top waiting for the clock to tick. I’m hanging back a few tenth of a mile in better lift getting up in the cat birds’ seat to watch the action in front of me. Position is a big deal for my psyche.

Kurt, Jim Lee, Johann Posch, Christian Ciech, Antoine, Manfred and a few other fast guys are around so I figure that this won’t necessarily be a loser move to go at 1:30, in spite of the weak lift we’ve been experiencing.

At 1:30 the whole group goes and there isn’t anyone else even nearby. It’s great having a few folks out in front and lower so that you can choose other lines to go to. I take a more southern route and hit the clouds early to find the next lift that we all get up in again just north of Polk City.

Christian finds the next one out front and puts himself in the lead with Johann and I right behind him and the flex wings trailing after us. It’s looking good to me. Christian goes on a long/fast glide. He’s doing well. But it is great to see that he is not over powering. He is flying faster than Johann and I, which most likely is a good idea. I’ll have to amp it up a bit more tomorrow.

Johann and I and a few flexies stop for a weak one as Christian heads to the turnpoint still not turning. We then head southwest to get under the clouds. Just before we get there, I see Christian starting to turn and head for him. I should have stayed with the gaggle as they get up quickly and it takes me a while to find the core under Christian. Now he’s really out in the lead. (Was that bonehead move #1?)

Johann and I get up next to the turnpoint and head toward Fantasy of Flight. Johann finds a good one to my right but in another bonehead move I continue on, having to slow down to find a good one seven miles out from this second turnpoint.

Many of the flex wings are doing well now above me and we are fighting up wind to get to Fantasy. I make a couple of more bone head moves and land 5 miles short. Johann gets high at Fantasy and continues to Chalet Suzanne.

Christian gets low by Fantasy also, out there out on his own down to 500 feet and takes 20 minutes to get back to cloud base. Still he is able to fly fast and get to goal first.

Manfred also took the 1:30 PM start time and makes goal soon after Christian. He wins the day for the flex wings (or at least I think he does as he was so fast).

Alex Ploner takes the 2 PM start time, and after a few 1:30 flex wings make it in, he makes it to goal to win the day for the rigid wings. He also got low but found a strong thermal right away. David Chamet comes in a few minutes later. The rest of the rigid wings trickle in quite a bit later. The flex wings are interspersed with the rigids (perhaps we will call separate tasks for them.)

Mike Barber cuts his knee on landing on his back wires. It is a pretty substantial cut with a bucket full of blood down his leg into his sock. He was wearing pants with holes in the knees (he lives out of his truck) and that’s where the wire got him. Can you say stripped wires?

Mike and Dorval came across the line very low and close together with Mike the lowest. Mike basically had to land without much time to think about it. He took out a down tube also without any other damage to the glider.

Discuss competitions at OzReport.com/forum/phpBB2

Discuss "The 2003 Wallaby Open" at the Oz Report forum   link»

Angelo Crapanzano|GAP

Angelo Crapanzano <angelo@metamorfosi.com> writes:

I see there is some confusion with NominalGoal and the other parameters used by GAP. I'll try to make it clear.

Why do we need a DayQuality? When adding different tasks we need to find out how much each task is worth compared to the other tasks, to find out who is the best pilot. Not all the tasks are worth the same because, just for example, we know some luck is often involved in our performances as almost everybody agrees:

- in a 30 minutes task luck is more involved than in a 3 hours task.

- luck is probably involved also in a task where 90% of the pilots bomb out and the others fly a long distance.

- some luck could also be involved if the task is too difficult, like flying strong headwind or in very stable conditions.

These are also the kinds of tasks where there is often a big spread in performances (read points) and, because we don't want the winner of a single task to get an unrecoverable advantage, we need to devalue this kind of task compared to more reliable ones.

As a too easy task is not a good one to find out who is the best pilot, such tasks are "automatically" devalued because, despite the fact that all pilots get a lot of points there is very little spread (and we are only interested in points difference, not in the absolute points achieved). For example, a task where everybody lands in the same field (not at goal) after 200 km and everybody gets 1000 points is practically worth nothing because there is no points difference (note that a task with everybody in goal cannot be considered "easy" because the difficulty is in going fast then).

GAP considers a task is fully valid if the time spent in the air is reasonable, the distance flown is reasonable and the pilot's distribution along the course is reasonable. Everybody would probably agree, but what does reasonable mean?

For example, a reasonable task distance is for sure different if we are flying in England, the Alps or Australia. A reasonable time is also different if it is a Club competition or a World Championship. That's why GAP requires the meet director to set some "NominalParameters" to let the scoring system know what kind of competition it's going to score. These NominalParameters are used to weight one task to the other and, of course, cannot be changed during the competition and are: NominalTime, MinimumDistance, NominalGoal, NominalDistance.

These parameters are used to find DayQuality, which varies between 0 and 1 and measures how suitable is the competition day to evaluate pilot's skill. It is obtained by multiplying the three validity coefficients: TimeValidity, DistanceValidity and LaunchValidity

TimeValidity.

NominalTime is the minimum fastest time to complete the speed section, the meet director considers reasonable for a task to be worth 1000 points. A reasonable task time is something which doesn't change much with the place but does change with the competition format and philosophy: If it's intended to be a race competition where most pilots are expected to goal everyday then NominalTime could be as short as 1 hour, while if it's intended like a very long distance competition then NominalTime could be over 3 hours. If the fastest pilot completes a task in less than NominalTime, the task will be devalued according to this graph:

DistanceValidity.

MinimumDistance is simply the distance below which is not worth to see the differences in performances between pilots: it is the distance normally achieved just by climbing over takeoff and gliding down. For several reasons, including safety, we don't want to make a difference between a pilot bombed out after 5 km and one bombed out after 7 km: it's worthless. This parameter is of course affected by the place we are going to fly: in the Alps we use 10 km while in England for a club competition could be as low as 3 km and in the Owens valley could be set up to 20 km. All the pilots flying less than minimum distance will score minimum distance.

NominalGoal is the percentage of pilots the meet director likes, on most days, to get at goal. If he's going to run a pure race competition where everybody is expected in goal everyday then NominalGoal should be set to 100%. If it is expected to be a pure free distance competition then it should be set to 0%. Nowadays, in international competitions, NominalGoal is often set between 20 and 30%.

NominalDistance is the task distance below which a good task, with the pilots evenly distributed along the course and NominalGoal pilots in goal, should be worth just less than 1000 points. This is of course heavily affected by the place and the pilots’ level. NominalDistance and NominalTime should be checked one to the other because NominalTime is also the time expected for the fastest pilot to fly NominalDistance on an average day.

Let's make an example to make it clearer: Once, for example, the meet director decides MinimumDistance=10 and NominalGoal=30 if he plans to use NominalTime=120 minutes and NominalDistance=70 km to check the numbers the questions the meet director should answer are:

Is, in my opinion, a good task where the winner flies for 120 minutes a task which is worth 990 points?

If yes he goes on:

Is, in my opinion, in this place and with these pilots, a task of 70 km with the pilots evenly distributed along the course and 30% of pilots in goal worth 990 points?

If the answer is yes then he should ask himself:

Is the winner likely going to complete this nominal task in 118 minutes?

If the answer is yes again he got the correct numbers otherwise he has to arrange the numbers until they are consistent one to the other.

These NominalParameters do affect how a task is weighted compared to another: to get correct results they must be carefully chosen by the meet director and cannot be modified during the competition.

MiniumumDistance, NominalGoal and NominalDistance do not affect "directly" the task value but are used to define the NominalDistribution as shown in the graph below:

In this graph the blue line connecting MinimumDistance at 100%, NominalDistance at NominalGoal% and MaximumDistance at 0% define the NominalDistribution expected by meet director in the competition, giving his preferences, place potential, weather forecasted and pilots skill (MaximumDistance is the longest flight on the task).

This NominalDistribution is compared to each task actual pilots distribution along the course: if the ActualDistribution, on average, falls below the NominalDistribution means that many pilot flew less than expected and the task value is reduced (of course a task value cannot be higher than 1).

LaunchValidity

Normally LaunchValidity is 1 because everybody on takeoff will launch but, on some days, takeoff conditions or weather could be questionable or could rapidly change. The final decision where to fly or not must be left to each pilot so LaunchValidity depends on the percentage of pilots launched compared to those present on takeoff (not the pilots registered but those who are actually present) according to this graph where the validity of a day is not affected by just a few pilots:

Now, I hope it's clear how GAP finds the DayQuality to compare, in a fair way, different tasks. (If you still have questions have a look at the GAP2002 explanation file under my web page www.metamorfosi.com)

Discuss "GAP – quality time, quality distance" at the Oz Report forum   link»

GAP|Lukas Etz|Tim Cummings

D Tim Cummings <timcu@mac.com> writes:

The GAP parameter "percentage of pilots in goal" is the minimum percentage of pilots in goal. If more arrive in goal, then there is no penalty on the scoring. If less arrive in goal then the distance validity may be reduced (depends on how far everyone flew). So as you can see, on a fully valid day, the "percentage of pilots in goal" GAP parameter has not affected the scoring at all

Lukas Etz wrote: Percentage of pilots in Goal? Mostly the comp organizers and task setters tried to get 30% in goal (so it was in the last worlds). In Deni and Hay the GAP-setting for “Nominal Goal” was 25%, and in the Bogong Cup it was 10%!!

Why is this? Is it really true, that we want (up to) 90% of pilots landing out of goal?? That’s really frustrating for the average pilot and extremely hard for beginners and new pilots.

Discuss "Pilots at goal" at the Oz Report forum   link»

competition|GAP|Lukas Etz

Lukas Etz <Lukas.Etz@gmx.de> writes:

For some years I am supporting more “open” types of task setting and a more simple scoring system.

There are many points in the present system which are not good in my eyes:

- Many times it’s not clear who is the day winner, until the computer lists are printed (sometimes the top two pilots are at goal, but the day winner is a third {average) pilot.)

- It’s completely impossible to transfer the pilots’ excitement to the spectators. (often the spectators only care about the landings and their noises)

- In the start cylinder it’s impossible for the pilots to know, if they should go for “early bird points” or not.

- The good pilots often have to play tactical games at the border of the start cylinder, instead of racing -flying the course in (sometimes perfect) thermals.

- Percentage of pilots in Goal? Mostly the comp organizers and task setters tried to get 30% in goal (so it was in the last worlds). In Deni and Hay the GAP-setting for “Nominal Goal” was 25%, and in the Bogong Cup it was 10%!!

Why is this? Is it really true, that we want (up to) 90% of pilots landing out of goal?? That’s really frustrating for the average pilot and extremely hard for beginners and new pilots.

- With our present type of task setting we can make big mistakes.

This is my main point. As a competition pilot, I have to spend so much for flying in a comp:

- lots of money (Costs for travel, transport, retrieval, towing, entry fees and material… look at the new glider after some hours before launch in Hay…),

- much time

- many personal difficulties (family, job…)

After having invested all this – I am at the place of my dreams: I am flying close to nice clouds, 10,000 ft above the ground, in a big comp, together with friends and in perfect weather conditions. I am getting some turnpoints, and then the goal is 30 or 60 km upwind instead of making 300 or 400 km downwind together with many others, I am landing some km short of goal.

The 4^th task in Deni and the first task in Bogong had seen this completely sad task settings. These days had amazing good weather conditions. These 2 days were the best of the comps, they were good for records or for many “personal best” flights, but the pilots had to fight against headwind for some hours. You can call such a task “progressively more difficult”, for me it was frustrating. I guess I am not the only one, who was disappointed. Have a look at the pilot lists:

Deni: 115 pilots

Bogong: 87 pilots

Hay: 64 pilots

(editor’s note: The Australia Open is scored for three classes, not just one, to encourage king posted and floater pilots. The Bogong Cup is limited in the number of pilots it will accept. Hay is a punishing place in the heat if you are not really totally motivated.)

So the number of pilots was decreasing strongly. Although Hay was called “Australian Nationals” which sounds like the only “real title”.

There are many advantages in “open” tasks, and there are so many new ideas to realize:

In an “open task” every pilot makes his own task. That means the tasks are not designed for 75-90% out landings! Everybody can reach goal.

Some proposals:

a) A day with strong wind and good conditions (f.e.in Hay):

The scoring could depend on the pilots best speed over a flown distance of 200km in a straight line. That means for the pilot: he can fly further then 200 km, trying to improve this measured speed, he is free to choose “his” direction. He can land close to a town or close to other pilots; or he can try to break his “personal best.”

b) A day with low wind and poor, average or good conditions.

An open triangle, like in the online contest. (Start cylinder and start time is given)

c) A day for the spectators

An open triangle with a latest time for crossing goal line.

d) An “assigned area task” like the gliding planes (so the directions of the task legs are given)

When I was proposing these tasks I normally got very skeptical or adverse comments by the best pilots and big interest by the inexperienced pilots. (We should more encourage and help them.)

This kind of “open” tasks is not the only way to go. But I am sure, that these tasks will make a lot of fun. It’s clear, that other pilot skills will be important, but the best (the same) pilots will win. These task settings and these scoring systems are worth it just to experiment.

In our German Hang gliding League we will make a first attempt (one single “open task”) in this coming season.

Discuss "More on competition formats" at the Oz Report forum   link»

GAP|Tim Cummings|USHGA

Maybe the USHGA should just use the built in validity factor in GAP.

D. Tim Cummings <tim@triptera.com.au> writes:

InAustralia we require the sum of day validities to be 4.000 for the competition to be fully valid. With GAP (and OzGap ), it is possible to have a fully valid day (validity 1.000) and no one getting 1000 points. On each day there are three winners, fastest to goal, first to goal, and departure points winner. It shouldn't be necessary for this to be the same person to maximise the validity of the competition.

GAP actually prints out the day validity on the last page of all the day score sheets.

Day validity = launch validity x distance validity x time validity

Launch validity decreases if you get several DNF's (people who turned up to launch, set up their glider, then decided not to fly e.g. wind too strong, wind gone over the back)

Distance validity is calculated from the nominal bombout distance, nominal task distance, nominal number in goal, and the day's task distance. If the non goal pilots are spread evenly between bombout and nominal task distance, and the actual number in goal is equal to the nominal number in goal, then distance validity is 1.000. If the sum of everyone's distance is even greater, then validity is also 1.000.

Time validity is 1.000 if the fastest pilot took longer than then nominal time (usually 90 minutes in Australian comps). Ensures the task committee doesn't set too short a task.

If no one makes goal, then the validity can still be 1.000, but the maximum score on the day is 900.

Discuss "GAP’s validity" at the Oz Report forum   link»

Thu, May 2 2002, 2:00:02 pm EDT

Angelo Crapanzano|GAP|Peter Gray

Angelo Crapanzano|GAP|John "Ole" Olson|Peter Gray

(?-i)John "Ole" Olson|Angelo Crapanzano|GAP|Peter Gray

Angelo Crapanzano|GAP|Peter Gray

Tim Meany, the scorekeeper at the Flytec Championship, and I have been checking out how using the OzGAP and GAP 2002 scoring formulas would change the results of the Flytec Championship. The changes are very small basically because the GAP system didn’t encourage anyone to leave early so the difference in departure/leading points is very small.

Tim has puttered around with using the more complex version of GAP2002 that requires the track logs from Compe-GPS. So far it looks like in this case there is very little difference between the straight line GAP 2002 method and the one that uses the tracklogs. Hopefully Tim will have a chance to work more on this.

We are hoping to get an updated version of Compe-GPS within the next week that gets rid of some problems and allows for us to more easily compare the results between these three methods. Of course, if the pilots are told that the meet uses the GAP system, they are not going to sprint out early and thereby get the departure points. Scoring the system after the fact with OzGAP or GAP2002 won’t change things much. The whole point of these scoring systems is to change pilot behavior and give pilots who leave early a fair return for their risk and difficulty.

Tim and Dave handled the scoring the Flytec Championship with ease, working well with the little quirks that they found in the Compe-GPS and Race programs. They didn’t take on the difficult task of trying to do the most complicated version of the GAP 2002 scoring system, which caused so many problems for Peter Gray at the Wallaby Open.

We expect to get reports from Angelo Crapanzano about future meets that are being scored with GAP 2002 and Compe-GPS.

A.I.R. ATOS|Aeros Combat|Aeros Combat 2|Aeros Ltd|Alex Ploner|Chris Arai|Christian Ciech|cloud|competition|Curt Warren|David "Dave" Glover|Flytec Championships 2002|Flytec Championships 2005|gaggle|GAP|Gary Osoba|Ghostbuster|job|Mike Barber|Moyes Delta Gliders|Moyes Litespeed|Quest Air|Ron Gleason|Steve Kroop|tail|tracker|tug|video|weather|Wills Wing Talon

David Glover was very smart and every day as the meet went on he would drag up folks to thank them for their help at the Flytec Championship. During the week he thanked the tug pilots, the volunteers, the ground crew, the Quest Air crew, the people who put the dinners together, Frank and Steve Kroop, the registration crew, etc. Because it happened every day everyone got more applause and more attention than if he had put it off until the last night, when everyone gets crammed together.

David and Steve did something also very clever, they had GW create a video taking footage and shots during the week. On Saturday night, the last night of the Flytec Championship, the video was ready to go and we got to see the whole video with the sound track. It was amazing that it had been done so quickly, all the while GW just looked like he was hanging out taking pictures and having fun.

But, not only did we get to see the video, all the pilots and tug pilots got a copy of the video last night. It really showed off what we do at a Floridaaerotow competition and we’ll be able to take it around and show it to our friends (if we have any outside of hang gliding).

Dave was a kick all week making announcements, telling jokes, getting pilots to come to the pilot meetings because they were so much fun. Belinda commented that we hadn’t seen Dave in his element in quite a while. While there were many many people who played keys roles in making the Flytec Championship such a great meet, I’ve got to feel that it was Dave Glover that really put it over the top and made it so much fun.

One of the key elements to its success (I feel), is that he was able to delegate responsibility to others, and in this case I’m referring to the task committee. I had written to him early on stating how giving the task committee the complete responsibility for calling the task was one of the keys to Tove’s great meets in Australia. David, like Tove, had the personality that allowed him to delegate responsibility and not get tied up into knots about it.

Chris Arai, Revo, and I had complete authority to choose the task each day, we took lots of pilot input and we did our best for the pilots to make the meet fair and fun. I can tell you that there was no barbeque task on the last day (although we did come back to Quest Air) as there has been in the past.

Having a task committee made up exclusively of pilots who without prodding from the meet director or organizer (well, we kept Dave away almost all of the time), is a key to having a meet (there are other ways to do it, but there are very very few individuals who can pull it off, and I can think of only one, Mad Dog, in Australia) that satisfies the pilots. This will continue to be a difficult issue and I hope that there will be a way to work this out at the upcoming worlds in Chelan.

Oh, yes, we did have a task on day seven. First, we heard from Gary Osoba:

Looks like a 7 day meet, thanks to the task planners, meet administrators, and reasonably good weather. Congratulations!

For a change, the entire soaring window today should provide for relatively consistent wind directions and strengths. Should make the planning a little bit easier. Here's how it stacks up:

11am Weak lift. Probably a bit early for clouds to be forming yet. When then do (likely closer tonoon), they should be around 2500'. Surface winds sse around 6. Winds aloft a little bit more southerly at 10-12.

2pm Good lift, strong in spots. Cb 5000' to 5300'. Surface winds sse 5-8, aloft sse 12-14.

5pm Moderate lift, good in spots. Cb could go to around 6000". Surface winds sse 8-10, aloft sse about 12.

"Only a fool would try to predict the weather"!

So we’ve got strong winds aloft out of the south, but good lift also. With the good lift we can come back against the strong winds, and that is exactly what we plan to do. We have really been counting on Gary’s forecasts, and he has done a great job for us. We really pick the tasks based on his weather predictions.

We call a 70 mile task (no barbeque task this one) that will first take us downwind 17 miles to the west, northwest to Bushnell, back against the head wind to Quest, south into the head wind to the intersection of highways 33 and 474, then north, downwind past Quest to Gator field, then 7 miles upwind for the final glide to goal.

The task keeps us near Quest, while at the same time making it difficult to complete. We come back over Quest to get everyone on the ground excited and it keeps us out of the swamp.

There are plenty of clouds when we start taking off, and it looks like maybe there are too many, maybe it will over develop after all. The wind seems awfully strong also, but it’s too late now to come up with an other task.

We hang around until the middle start time at 1:15although everyone is in the air in half an hour. We just stay at cloud base for an extra 15 minutes. GAP gives one very little reason to go out in front and leave the gaggle behind. Johann and I have already made up our minds that we will take the middle start time, and maybe everyone else felt that way also, or, when two pilots left, they all decided to go with us.

We can see all the darkness out to our northwest and it looks like we are jumping into a black hole. There are high cirrus clouds that dull the areas on the ground where the cu’s don’t block out the sun.

We spot the guys who took the 1 PMstart time and that makes it easier to make our way to the turnpoint through all this very gloomy looking area. Still we’re down below 1,400’ before we connect with the big lift that gets us to the turnpoint and back out again.

Well, that was a downwind task, but we averaged only 27 mph getting there, so the south wind didn’t help that much. Coming back will prove to be much more difficult (and most if not all of the pilots who don’t make it will drop out here), as the average speed will go down to 15 mph.

I’ll charge across some blue areas to get under what seems to be a cloud street, find nothing then push up wind to get under some pilots turning at 8 miles out from Quest, only to find myself at 450’ and working lift that averages 140 fpm, starting out at a much lower value. It turns out that every one will have difficulty making it back to Quest and will get low on this leg.

Christian Ciech and Alex Ploner are doing much better in this meet than the rest of us, and they have zoomed out ahead. They were half a mile ahead at the turnpoint, and I lost them coming back as I went more easterly, but they will also get quite low. They are flying together.

I’m flying with Johann, but I’ve lost him also. Given how weak the conditions are we are all struggling and it looks like a long day if we can stay up. I’ve got quite a few other gliders here with me, so we hang on and the lift improves, as we drift north away from Quest, but with stronger lift it is no problem. Twenty minutes after coming in low, I’m up to 4,000’ and on my way to Quest with Curt, Paris, Ron Gleason, and some other ATOSes.

We are styling now, hitting good lift and staying high as we come into Quest. We can see a gaggle forming to our south that includes the Swifts, so they must have struggled also. I can see Alex and Christian in the gaggle also.

Johann will lose it here and head off to the west to get under better looking clouds, while I’ll continue to the south to join up with the gaggle. Johann, who is in second place, will almost land at Quest.

I catch up with the lead gaggle, which is putzing along. I guess they don’t see any need to probe out into the blue to the south. All the clouds that we had by Bushnell have not come down here in the late afternoon, and there are mere wisps to our south.

With a bunch of rigid wings, Curt Warren, and above us all the two Italian rigid wing pilots, we start punching our way south only to find good lift, light sink in between, and long patches of buoyant air. We stay high and work light lift to get to 5,000’.

We are still running into the wind, so it takes a while, but we have no problems getting down to the south to get the turnpoint with Alex and Christian leading the way. The Swifts start to get ahead of us now, with Manfred taking the third turnpoint and coming back to greet us when we are 2 miles out from it.

As soon as we get the turnpoint, we can drift back north along 33 in strong tail winds and buoyant air. With the lift averaging less than 200 fpm in the cores, we are just taking a little bit here and there. After the long up wind grind it is a joy to drift toward the Gator turnpoint.

Now there are only rigid wings in the lead as we come into the Gator turnpoint and turn to get back to Quest. It’s been a long glide into Gator before our upwind final glide. My IQ/Comp has been acting up and not reporting any final glide info, so I’m just hanging with the four other rigid pilots. Heiner goes on glide and we all just speed up as it becomes clear that no matter that fact that we are going into a strong head wind, we will make it back to Quest without a problem.

Mike Barber who bombed out on the previous day (after passing up lots of lift trying to go faster) goes all out and will win in Class 1 as the flex wings will come in about 20 minutes behind Alex Ploner who takes first in the rigid wings. Then again he'll start fifteen minutes behind us, so you can see how much Alex and Christian were holding back, just tracking the rigids below them.

Class 5 on the last day:

Class 5 finals:

Class 1 last day:

Finals Class 1:

Preliminary results are up on the www.flytec.com web site.

Discuss "Flytec Championship – 70 mile fish bowl" at the Oz Report forum   link»

Mon, Sep 10 2001, 6:00:02 pm EDT

Rohan Taylor

Angelo Crapanzano «angelo» writes:

I do agree on Rohan analysis even if the problem, in Algodonales, was mainly due to the flight conditions and area. Pilots were forced to take-off early but good thermals developed later and it was not so easy to find good thermals on course. So it was much better (faster and safer) to fly in gaggles and having thermal indicators ahead.

In this situation it was too risky to fly alone and the early bird was simply not rewarding enough. I do not agree the "system helped create some of the largest gaggles we've seen" (it would have been even worse with any other scoring), but I do agree GAP scoring didn't fight enough the advantage of having thermal snoopers ahead. Juaki (the meet director) made, in my opinion, the best possible job but Algodonales should not had been chosen for a World Championship and the number of competitors should have been reduced (in a clever way).

(editor's note: The Oz Report repeatedly brought this issue to the fore.)

When I introduced the early bird bonus for the first time in 1988 most pilots where against it and they have been against it up to very few years ago. It's funny to see that now everybody wants to have it increased. Please note the early bird was kept low in the first versions of GAP for "political reasons" (i.e. to promote the acceptance between pilots).

From GAP96 to GAP98, the early bird points increased by 20% and then another 20% with GAP2000 (always at the expenses of speed points). It could (and should) be increased a little more but, as it is designed right now, it could not solve the "Algodonales" problem. These points come at the expenses of giving bigger problems somewhere else (remember that the scoring must work anywhere in the world, not only in a specific area with a specific situation).

Let's look at Rohan proposals:

(a) early bird departure points should be awarded independently of speed points. It was like that at the beginning, but we had to modify it. It was enough to takeoff early and float around to get points! The early bird bonus are intended to reward pilots who take the risk to fly in front, not just start in front.

(b) early bird departure points should be awarded even if the pilot does not get to goal, as there should be some reward for taking the risk of path finding. I already tried this many years ago, but did not work because we had several examples of slow pilots getting more points than fast ones landing in the same place one hour earlier. Anyway, right now, with GPS verification we have much more information on the flight so something like an "arrival bonus" at each turn point could be used. This, of course, would increase the importance of task setting, could bring the problem of pilots racing (and getting low) to the turn points and it would became very difficult for the pilots to weigh out advantages and disadvantages.

Rohan, I don't want to appear as one who says "NO" to any proposal. You see I agree 100% with your analysis (if not with your solutions), and for sure I do consider your feedback really important, but we must be very careful in making modifications because it's very easy to destroy the equilibrium in a scoring system.

Right now I'm thinking (not so much because I'm really busy with my work.) about the "turn point arrival bonus" and other solutions but I'm not fully happy with any of them. The idea of GAP is to have a scoring that, despite the mathematical complexity, has an easy to understand philosophy. It's mandatory to keep this characteristic alive. (If you are interested in GAP philosophy have a look at http://www.metamorfosi.com/GAP2000pdf_en.htm)

Fri, Sep 7 2001, 6:00:01 pm EDT

Rohan Holtkamp|Rohan Taylor|Worlds

Rohan Holtkamp «dynamic» writes in about the GAP scoring system. I wrote back to tell him that Angelo has said that they will work in this issue soon. Of course, the Australian meet directors are already modifying the scoring system for the upcoming Oz Meets:

In one of your recent Oz reports you issued an invitation to submit opinion on the current scoring system. The system used at all recent high level comps like the Worlds and various national championships, requires a little more tuning. To follow on from the discussions held nearly every morning at the team leaders briefings in Algodanales (World championships).

It was a popularly held opinion by most team leaders that the early bird departure points were not rewarding enough to encourage pilots to depart the start circle early, and many pilots held back to let the early pilots to find the thermals. Evidently the system helped create some of the largest gaggles we've seen, and made collision avoidance the highest priority at the start circle most days. (not too bad if you're on top:-)

It was agreed by most team leaders that;

(a) early bird departure points should be awarded independent of speed points.

(b) early bird departure points should be awarded even if the pilot does not get to goal, as there should be some reward for taking the risk of path finding.

Nearly all pilots I've spoken to detest this hanging back and 'leaching' style, but currently the scoring system rewards those that play it safe and don't have a go, so it becomes the state of play. All team leaders agreed that the scoring system could not be changed once the comp had been started so the status quo remained.

This is valuable feedback gleaned from a meeting of the worlds best and their coaches/leaders, perhaps should not be overlooked

Fri, Aug 31 2001, 5:00:05 pm EDT

Davis Straub|GAP|USHGA

Want to know how the GAP scoring system works? Check out: http://www.fai.org/hang_gliding/race/formulas.htm and http://www.hgfa.asn.au/gap/gap20.htm.

Other organizations have modified the GAP scoring system to emphasize other aspects of racing. I believe that we will be using a modified system this January in Australia.