Scoring short tasks
- This topic has 5 replies, 4 voices, and was last updated 13 years, 2 months ago by
.
-
Posts
-
<p>As an exercise in learning, I have been trying to use FS to rescore the Corryong Cup. </p>
<p>When scoring the floater class competition, I have noticed some issues with short tasks.</p>
<p>The comp is scored with these parameters:<br />
Nominal Distance: 20km<br />
Nominal Time: 1 Hour<br />
Nominal Goal Percentage: 30km<br />
Minimal Distance: 5km<br />
Start Cylinder: 5000m<br />
TP Cylinders: 400m<br />
Goal Cylinder: 1000m</p>
<p>The only launch used in the comp is MtElliott (see tps below)</p>
<p>On the last day, the Task Committee called a very short task for Floater – ELLIOT to KANGCK. </p>
<p>In See You, the default measurement for task distance is to include the distance to the center of the cyliders, so this is a 9.4km task ( there is an option to shorten the distance to the cylinder boundaries ) – see http://soaringspot.com/corcup2011/results/mixed/task/day6.html</p>
<p>FS however only calculates the distance as cylinder to cylinder, so this task, is calculated at 3.4km. Now, 3.4km is less than the minimum 5km, so FS throws an error when trying to calculate the scores "ScoreTask: failed scoring task. ScoreTask: set_departure_points == double.NaN!!" Even when the Minimum is reduced to 3km, a pilot who landed in bombout (2km from Launch) is scored the same distance ( 3km )as a pilot who landed 1km short of goal because 1km short of goal is 2.4km along the course line outside the start cylinder, whereas bombount is inside the 5km start cylinder.</p>
<p>The following day – the last of the comp, the Task Committee called the same task for floater. It was a much better soaring day, but as we had the presentation dinner that evening, it is traditional to call short tasks to get everyone off the hill and back to town early enough for the scoring to be completed before dinner.</p>
<p>As it was a better soaring day, some pilots completed the task, however being a very very short task, one pilot was able to get very high inside the start cylinder, and flew between the exit cylinder and the goal cylinder in about 3 minutes. I havent’ yet figured out the exact cause, but if the Nominal Time is not reduced, then the same error message as I mentioned above is displayed when scoring the day. Reducing the Nominal time to .5 is the highest value which allows it to not error. Interestingly, See You also had trouble with scoring this task as the pilots performance devalued the day down to a zero point day.</p>
<p>Regards,</p>
<p>Dawson</p>
<p>$FormatGEO<br />
CLCCLC S 36 13 04.92 E 147 49 43.98 329 CLCCLC<br />
CORRY S 36 11 06.00 E 147 53 29.04 291 CORRY<br />
CUDG S 36 11 32.22 E 147 46 12.78 325 CUDG<br />
CUDGNO S 36 05 48.72 E 147 48 51.42 314 CUDGNO<br />
CUDGWE S 36 13 23.46 E 147 43 43.68 432 CUDGWE<br />
ELLIOT S 36 11 09.00 E 147 58 36.00 935 ELLIOT<br />
GEHI S 36 22 59.82 E 148 10 54.06 427 GEHI<br />
GRGGRG S 36 03 32.64 E 148 02 23.40 244 GRGGRG<br />
GUYFOR S 36 04 20.04 E 147 38 30.78 439 GUYFOR<br />
HOLBRK S 35 41 07.32 E 147 19 13.80 261 HOLBRK<br />
JINGEL S 35 56 01.02 E 147 42 02.04 230 JINGEL<br />
KANGCK S 36 15 50.76 E 147 56 18.48 376 KANGCK<br />
KHANCO S 36 12 58.38 E 148 06 35.22 290 KHANCO<br />
LIGHTH S 36 05 11.52 E 148 02 44.10 678 LIGHTH<br />
MTKOSC S 36 27 08.04 E 148 15 01.92 1945 MTKOSC<br />
MTMITA S 36 08 56.22 E 147 51 51.06 990 MTMITA<br />
MUNDER S 35 50 18.96 E 147 52 29.10 550 MUNDER<br />
THOWGL S 36 18 19.50 E 147 53 55.92 426 THOWGL<br />
TINTAL S 36 03 07.02 E 147 55 36.96 264 TINTAL<br />
TOOMA S 35 58 04.26 E 148 03 28.98 276 TOOMA<br />
TOWONG S 36 07 37.68 E 147 59 08.64 259 TOWONG<br />
WALWA S 35 57 58.38 E 147 44 03.06 224 WALWA</p>
<p>On the second last day of the comp, with dire forecasts, the task committee called a very short task. Mt Elliott ( Lauch
</p>oops, please ignore the last, broken sentence, I had reworded what I was writing and left that bit in accidentally.
It’s hard to debug with only Reflector to disassemble the code, but it appears as though the “set_departure_points == double.NaN!” error occurs when CalcTimeValidity() or CalcDistanceValidity() returns a negative number, causing CalcDayQuality to return a negative number.
I see that Davis’ Launch Validity calculation for SeeYou sets LaunchValidity to 0 where the script calculates it to < 0
It does the same for DistanceValidity and TimeValidity.
In FS, CalcDistanceValidity() does the same, but CalcTimeValidity and CalcLaunchValidity don’t do this, so could allow negative validity to be returned.Dawson
Hi,
I have never tried to score such a short task, and I do not think it is a good idea to use GAP to score something like this as you will have strange and unpredictable results for a 5 minute, 3 km task…
One of the advantages with FS is that it does calculate the shortest possible distance of the task, and you have just proven that it works 🙂
– Oyvind
Hi Oyvind,
You can’t say that this test proves that the calculation of the shortest possible distance works, because it simply didn’t work, it presented an error message, and failed to calculate anything.
It may be that such short tasks are unable to be scored by GAP, however, the scoring software should not report an error to the end user.
“ScoreTask: failed scoring task. ScoreTask: set_departure_points == double.NaN!!”
Does nothing in any way to assist the scorer who may be scoring their first competition ( as indeed I was at Corryong 2011 ) We used See You to score the competition which was able to score the day, it simply sored it with incredibly low validity. Had the decision been made before hand to use FS, then no scoring of the day at all would have been possible.
If FS is not to be able to score such tasks, then the documentation of the software needs to specify what the minimum requirements are for a scorable task, and when such an unscorable task is entered, instead of presenting a cryptic error message, it needs to inform the scorer of the reason why the task is not able to be scored, and which rules of the software have been broken causing the failure.
Alternatively, FS should score the task in the same way that See You does, and simply results in a ridiculously low validity for the day.
Regards,
Dawson
- You must be logged in to reply to this topic.