facebook js

Follow Planet-Soaring on Facebook

Thursday, November 1, 2018

What futur for F3F ?

Hello all,

No ... I will not talk about the drone law that is coming, this is another topic :) !

Following the F3F world championship in Rügen, I think it is the perfect time to make a step back, to learn the lessons and think about the futur of F3F:

Personnaly, the 3 main lessons learned in Rügen are:

  1. A round with 63 pilots can take between 2 hours and 3 hours depending of conditions. During this time window, flying conditions can change a lot. For exemple in the morning when the air is still wet and the drying when the wind is raising, leading to 10s time difference, if not more.
  2. Starting order draw must be carefully done, which was not especially the case in Rügen
  3. When groups scoring happens (round interruption or end of the day), scoring can become very unfair because of the third man effect.

F3F is a nice category because rules are simple, and because we are flying in very different and nice to amazing places. All this make me think that participation will increase in the futur and competition will be even more larger, especially the world championship.

We have 2 years before the next WC. Why not brainstrom in order to make the rules to evolve in order to organise fair competition with a large number of pilots, an ensure everybody enjoy and not consider that sometimes F3F is just like casino, depending too much of the stating order ... or, at the opposite, to conclude that the current rules are perfect as they are today, who knows !

On top of that, I think there are few point in the rules that should be clarified to avoid interpretation like penalties in case the round is cancelled, round interuption reset in case of flight attempt but not valid, etc ...

Personnaly, I don't want to see people who don't fly F3F to try to impose us rules that we don't share, as it has been in a recent time.

As a starting point for this brainstorming, why not first reconsider is the scoring method. France proposed a rule change in 2010. Timing with CIAM calendar was wrong so I think the proposal didn't reach the Plenary meeting. Also, at that time, community was reluctant to this sort of change. During the technical meeting at Donovaly in 2014, I re-explained this scoring method and some people thought further and I could feel some interrest in some nation. Maybe today time has past and with experience of large competition, people are now ready in their mind to look at this rule change with opened eyes.

So here is the rule proposed initially:


F3F rule change proposal

The objective of the present rule change proposal is to modify the pilot round result calculation to use the round average time instead of the best time. This make the round ranking fairer in case of group scoring. For details description, please refer to the presentation included in this document.

The result of the flight is stated as the time in seconds as hundredths of seconds obtained by each competitor. For the purpose of calculating the result of the round the competitor’s result is converted this way:

1000 * Pw/P

where Pw is the average result of all competitors and P is the competitor’s result.

Reasons behind are:

  • In case of group scoring, the final ranking is made much fairer 
  • By reducing the “third man effect” final ranking is made more consistent with real man on man ranking 


FAI current scoring formula

F3F scoring is based on a comparison between the time of the run and  the best time of the round:
FAI = 1000 * Best Time / Pilot Time (Best time is used as the reference)

The amount of points you get after your run depends on 3 parameters:

  1. Your skills (the parameter we want to measure).
  2. The flight conditions you get during your run (that induces the famous "lottery effect") 
  3. The flight conditions during the best run ( that induces what I call "the third man effect" and adds some additional "lottery effect" to your own run.

"Third Man Effect"

The number of points you get depends on your run but it also depends on what happened during the best flight:

- A "top pilot" got the "1000" with average conditions
- A top pilot got the "1000" with "the thermal of the day"
- An "average pilot" got the "1000" with a "standard" thermal

Between these 2 pilots: Who is the winner ?

You think this is Pilot 2 ?

The real answer is : we can't know !
The final ranking depends on the "third man" : the guy who wins the rounds

Because of "third man effect" today FAI scoring method may make the final ranking not coherent with the real man-on-man ranking. The "third man effect" enhances the lottery effect.

To reduce the "third man effect" it is proposed to replace the best time of the round by the average time of the round

New = 1000*Av.Time / Pilot Time

Using the average time instead of the best time makes the reference time much less sensitive to flying conditions. Lottery effect due to third man effect is quite cancelled. But lottery effect related to your own flight remains the same. It has been verified that both methods give quite the same ranking when flying conditions are stable. Differences may occur when flight conditions are very variable . They always make final ranking consistent with man on man ranking.

Group Scoring

It is well known that group scoring can be dramaticaly unfair. This is the case when a pilot get a huge thermal scaling down the scores of all other pilots in his group (but having no effect on the other group).

Exemple: first 2010 competition in France (only 3 rounds with group scoring in the third round)

Pilot “A” makes a 35s => only 2 pilots in his group are around 820pts and all the others have (much) less than 800. On the second day the air is much lighter but much more stable => in the second group more than half pilots have (much) more than 900pts. At the end of the comp all pilots in the top ten (except Pilot “A”) were from group 2.

In case of group scoring the new scoring method is much fairer because in both groups the "average pilot" will get the same number of points (around 1000) wathever the best time is. if some care has been taken to make both groups of egal value this result is very fair.


It is proposed to use the "average time" instead of the "best time"in the F3F scoring formula. This modification has two advantages :

  1.    global ranking is made consistent with man on man ranking
  2.    group scoring becomes perfectly fair

Don't be wrong ! This new method scoring will not affect the result of the pilot getting the thermal or flying very well, "killing" the group. The objective of this scoring method against the average time is to not impact the pilots compared to the other group where all pilots are flying around the same time.

Now that group scoring doesn't impact negatively the pilot scoring, why not consider group scoring to be used in every round for competition with more than xx pilots ?

We could imagine groups of pilots between 10 and 15 (or more) depending of the total number of pilot.

But to make the competition as fair as possible the composition of the group must be done carefully to avoid pilots to "meet" the same pilots several time. Such algorythm exists already in other categories like F3K or F3/5J, so could easily be adapted to F3F.

This approach would allow fair competition with large number of pilots, maximizing the number of round flown during the day as we can fly a group before the night, and resume the next day.

Of course all this must be studied carefully, with simulation, etc ...

But I really think that we should spend some time to evaluate the benefit of theses changes, and not stay on the current rules and complain.

So how to progress on all this topic ?

I was thinking that we could create a international working group with one representative of each (active in F3F) country, who could play the role of relay with the different national communities. This representative could be the official category representative of the country, or a delegate, an active and experienced F3F pilot. This working group could brainstorm, work on ideas, simulations and finally eventually propose some rule changes with a wide support, following the CIAM / FAI calendar. I know there is a wish of the FAI/CIAM to have such working group.

Ready to participate/contribute ? Stay tuned !


  1. I have implemented the average time scoring in f3xvault!

  2. Hi Pierre, I think reducing the effect of the super run is a great idea. Will be interesting to see what effect group scoring has on standard rounds with large numbers. Is it possible to simulate and older event such as the Welsh Open from 2018? I think the average time as a multiplier on its own, would be a great improvement if I understand it correctly. Is there a place where this discussion is active and civil?

  3. Hi Pierre.
    It's interesting idea but this new calculation has "Third Slow Man Effect" :)
    Please find example using your times:

    Pilot 1: 45, 45, 50, 50
    Pilot 2: 50, 50, 41, 41
    Pilot 3: 80, 80, 51, 51

    New Calulation:
    Pilot 1: 3539
    Pilot 2: 3488
    Pilot 3: 2585

    As you can see Pilot 1 has won :)

    Anyway new calculation needs additional analisis, possible that it has less distortion.

  4. Hi Pierre,
    Several years ago, I started to write a technical analysis of F3F scoring, which includes a discussion of some of the issues you raise. It is informed by formal statistical ideas rather than simulation, and is probably not easy for people without a statistical background to read. I haven't looked at it for a while. I was trying to improve its accessibility when other projects got in the way. I'd be happy to participate in the working group you propose.


Drone regulation: Everything you need to know and comply to when coming to France

The new drone law is in place in France since the 26th of December 2018. It is applicable also to aeromodellers from all over the world vi...