[ste]

“- Arrival Time Points / 2 !!”

Actually they have, by design or error, used the OzGAP2005 style Arrival Time Points but with a twist:

m_fAvailableScoreArrival = 1000# * m_fDayQuality * (1 - m_fDistanceWeight) * 1 / 8

(as used in normal GAP Arrival Position Points) instead of

m_fAvailableScoreArrival = 1000# * m_fDayQuality * (1 - m_fDistanceWeight) * 1 / 4

which is used by OzGAP2005.

Stein-Tore

[ste]

It is not clear (from the rules or the comp reports) what GAP parameters are used.

It does look like they used
– Distance Points
– Time Points
– Leading Points
– Arrival Time Points / 2 !!
– Min. Dist = 5
in Korea.

[ste]

Hi Steve,

Have you got a clear definition for what should be considered a full scoring task?

Half the field having to thermal at least three times? Someone in goal? etc … What does a flown task have to look like before most pilots feel it has been a fair test of skills where the winner deserves maximum points?

Regardless of what mathematical formula one use to try to calculate some sort Day Quality factor one should have some common understanding of what sort of flying is required to make Day Quality = 1.0.

GAP attempts to provide tools for it by letting one define some values excepted for time and distance and no of pilots taking off.

I am not sure if this is (always/ever) the best way to do it. I do kind of like the way PWCA did it until 2009 with saying x pilots over y distance give DQ = 1.0. 1/x over y distance gives DQ = 1/x. Very simple to understand and everyone can quess what the DQ will be.

With the GAP DQ calcs it is very hard to give a good questimate of DQ after the day is over.

Anyhow here is the actual formulas for GAP DQ:

private static double CalcTimeValidity(double best_time, double nom_time, double best_dist, double nom_dist)
    {
      double tv_raw = 0;
      if (best_time > 0)
      {
        if (best_time < nom_time) tv_raw = best_time / nom_time;
        else tv_raw = 1;
      }
      else
      {
        if (best_dist < nom_dist) tv_raw = best_dist / nom_dist;
        else tv_raw = 1;
      }
      double res =
        -0.271
        + 2.912 * tv_raw
        - 2.098 * Math.Pow(tv_raw, 2)
        + 0.457 * Math.Pow(tv_raw, 3);
      if (res > 1) res = 1;
      return res;
    }

    private static double CalcLaunchValidity(int no_of_pilots_flying, int no_of_pilots_present)
    {
      double lv_raw = no_of_pilots_flying * 1.0 / no_of_pilots_present;
      double lv_raw2
        = 0.028 * lv_raw
        + 2.917 * Math.Pow(lv_raw, 2)
        - 1.944 * Math.Pow(lv_raw, 3);
      return lv_raw2 >= 1 ? 1 : lv_raw2;
    }

    private static double CalcDistanceValidity(double sum_dist_over_min_dist, double no_of_pilots_flying,
      double nom_goal, double nom_dist, double min_dist, double best_dist)
    {
      double dv_raw
        = sum_dist_over_min_dist
        /
        (
          no_of_pilots_flying
          / 2.0
          *
          (
            (nom_goal + 1.0) * (nom_dist - min_dist)
            + nom_goal * (best_dist - nom_dist)
          )
        );
      if (dv_raw > 1 || dv_raw < 0) dv_raw = 1;
      return dv_raw;
    }

    private static double CalcDayQuality(double time_validity, double launch_validity, double distance_validity)
    {
      return time_validity * launch_validity * distance_validity;
    }

Stein-Tore

[ste]

I have worked out how the PWC 2009 scoring formula is intended to work now (I think).

The base of it is GAP.

After scoring with GAP the following takes place:

– tasks with winner GAP score >=700 gives 1000 PWC points to the winner (the other scores adjusted accordingly).
– tasks with winner GAP score <=200 gives 0 PWC points to the winner (and the rest), ie the task does not count anywhere.
– tasks with winner GAP score >200 & <700 gives from 4 to 1000 PWC points to > the winner using the formula (x-B)(2*A-B-x)/(A-B)^2) where x is winner GAP score, A=700, B=200 (the other scores adjusted accordingly).

So, this is simple to achive, but why to do it I’m not sure…

In reality it is just a way to change the behaviour of the Day Quality calculation that GAP does (based on nom. dist., time, goal).

Stein-Tore

[ste]

One way to get the result in whatever format you like is to make a copy of “task_result.tkr.xsl”, the modify it as you like.

This requires some knowledge of xml and xslt. If you’re expert in xslt you can transform the results and do whatever calcultations you want here. Only your imagination (and xslt skills sets the limits).

Under the Reports menu there is the option to tick “Create source.xml file…”. You will then get an xml file which is the same as what is used as input for the xsl file. Use this to see what is available to play with.

Also Under the Reports menu there is “Reports using stylesheet (*.xsl)…”. Write your own xsl file from scratch. The input xml here is the whole fsdb file (Try to tick “Create source.xml…” and look at the .xml file).

As mentioned at the beginning, the more you want to do with this the more knowledge of xslt you will need.

Anyhow, to get something that can be opened in a spread sheet program one could create a task result xsl file which created a csv file with the relevant data.

Stein-Tore

[ste]

In FS Haversine is used with FAI’s radius.

In FsFlight under Tools > Calc. dist it can be tried.

For the above mentioned coordinates FsFlight get same dist as nxcl.

Note that the various instrument makers (GPS or flight instruments) does not all use the same methods so what a pilot see in the display will more often than not be different from what FS or GpsDump reports.

FS use a default errormargin of 0.5% (can be changed under Tools > Options) which should cover for any differences.

Stein-Tore

[ste]

See http://fs.fai.org/forum/topic/4-hour-cutoff-after-winch-launch regarding “Max time (hh: mm) en route from first crossing of turnpoint No”.

Stein-Tore

[ste]

In the test version of FS 1.2.13 ( http://fs.fai.org/forum/topic/fs-1213 ) there is an option in the task window to set “max time en-route from first crossing of a turnpoint”.

It should be used carefully and (probably) not with Race tasks. It is well suited for open distance tasks and possibly also with elapsed time tasks.

One good situation to use it would be as described over by ulf.
Set a circle covering the towing field, let the pilots be towed and once they’ve drifted out of the circle the first time their clock starts counting down.

Stein-Tore

[ste]

See http://fs.fai.org/ticket/174#comment:4 …

Stein-Tore

[ste]

Hi Gerogi,

It is a bug. I think it is fixed with FS 1.2.13 (or the last update of FS 1.2.12) but not sure. Could you please register a ticket (at http://fs.fai.org/newticket) and attach a zip file with the fsdb file and tracklogs for the task. Then I can check it.

Stein-Tore

[ste]

Going into a turnpoint circle or out of it does not make any difference.

All that matters is crossing it.

Regarding the two pilots mentioned in http://fs.fai.org/ticket/185:

703-Navarro Gonzalez reached the 3rd turnpoint and as such the last crossing of tp 2 (that was before the first crossing of tp 3) is registered as the valid crossing of that as this gives the fastest time btw tp 2 and tp 3.

705-Nuñez Alcon had tp 2 as the last turnpoint reached so the first crossing (inside the open periode of that turnpoint) is registered as the valid crossing.

Note, that in the task-result the time of the first start gate prior to their valid crossing will show as start-time.

Stein-Tore

[ste]

See http://fs.fai.org/ticket/173 …

Stein-Tore

[ste]

A “feature” 😉 of GAP:

From modGAP.bas (in old Race program):

m_fAvailableScoreDeparture = m_fAvailableScoreArrival * 1.4

.

Stein-Tore

[ste]

See http://fs.fai.org/ticket/139#comment:14 …

Stein-Tore

[ste]

See http://fs.fai.org/ticket/139#comment:13 …

Stein-Tore

[Author]

Posts