Thread Tools
Mar 11, 2018, 11:22 AM
Good for what ALES you
awilmunder's Avatar
Thread OP
Discussion

Fundamental change in how scores are calculated


This thread is for discussion of the change in how SVSS scores are calculated.

The benefits of entering scores by phone are significant. Scorekeeping using printed cards has always been time consuming and problematic trying to read and enter flight times and landing points for dozens of individuals, so this is very much a step in the right direction.

What may have been overlooked is that along with adopting a new way to enter scores, this scoring system also enforces a way to calculate scores that is fundamentally different from the method the club has used for decades. This is not necessarily bad, but it is important that all club members understand and agree to this since it changed yesterday’s pilot rankings and YTD percentages all the way to the top.

The fundamental change is how landing points are factored into the equation. Previously at SVSS, landings have always been calculated to represent a uniform 9% of our overall score. Using the new software, yesterday the landing scores represented 14% of the round score on the 10-minute flights to as high as 29% on the 4-minute flights. This is as much as a 323% increase in emphasis on landing points.

Again, there is no ‘right’ or ‘wrong’, but this represents a significant shift in emphasis from a soaring contest to a landing contest.

In this case, fundamental scoring rules have changed, but the significance of the impact may not have even been clear to the organizers, and certainly not to the pilots. We are also letting a scoring system dictate our competition values, in this case, emphasizing landings, rather than deciding on the importance of landings and choosing a scoring system that reflects the club’s values.
Sign up now
to remove ads between posts
Mar 11, 2018, 11:34 AM
Good for what ALES you
awilmunder's Avatar
Thread OP
Mark Twain said ‘I never let my schooling interfere with my education’. Here’s the part where I try to explain why this is happening without having to copy and paste the calculations out of Excel.

The software used yesterday was built for F5J and on a 10-minute flight landings represent 7.6% of your score. That is actually pretty close to the 9% we have used at SVSS. At Fall Fest, using this same software, landings represented 4% of your score on 10-minute flights. So why are we seeing such a huge difference? The first part of the answer is quite simple. F5J uses a 50-point landing tape, Fall Fest uses a 25-point landing target, and SVSS uses a 100-point tape, so on a 10-minute flight, our landings are receiving 4-times the emphasis as Fall Fest. Now when you factor in shorter flight times, this percent increases even more.

SVSS has traditionally normalized the flight score to 1000 points, whether 4 minutes or 10 minutes, and then added up to 100 points for landing, so the landing points were 100 of the 1100 possible points, or 9.09%.

The scoring approach used for today’s contest adds the landing points and then normalizes. It was also developed for a contest format with 10-minute flights and 50-point landings, so 600 flight points and the landing would represent 50 of the 650 possible points, or 7.69%. This percentage is pretty comparable to what saw above for SVSS and it would be easy to assume that adopting this approach would not represent a significant change.

Here is why that is not the case. At SVSS, we use a 100, not a 50-point landing tape; so on a 10-minute flight, the points for landings are doubled and became 14.28% of the score. On today’s 4-minute flight, you could receive 240 points for the flight and 100 for the landing, so 29.41% of your possible score comes from landings. This is more than a 3-fold increase from the 9% that the club has used for many decades. This was significant enough to have changed the rankings for the contest’s top pilots.

Why wasn’t this raised as an issue when the system was used at Fall Fest last year? Primarily because they use a 25-point landing circle, so on a 10-minute flight, their landings represented 4% of the score and 7.6% on 5-minute flights. These are both significantly less than the14% to 29% for our contest flights
Last edited by awilmunder; Mar 11, 2018 at 11:45 AM. Reason: added some numbers
Mar 11, 2018, 12:17 PM
Registered User
I was one of a few poor souls who suffered from Aric's worst case description outlined above. I nailed the 4 minute time on the button but slid through the landing and hit myself. OK, my bad, so I accept the zero landing. But under the old system my lowest possible score would have been 1000/1100= 91% but the glider score assigned me 71%, a major hit to my final score!

But I do like the entry system. I agree with Aric that is a good step forward.
Mar 11, 2018, 12:33 PM
E sailplane thermal hack
I wasn't at yesterday's TD comp since my focus is only electric,, but in ALES and F5J the goal has always been to DE Emphasize the landings. My personal feeling is that this is a good goal as the emphasis on landings encourages harder dorking with resultant stress on the airframes.
Mar 11, 2018, 12:42 PM
NorCal Electric Soaring (NES)
SoaringDude's Avatar
One thing that I'm experimenting with now is re-entering the landings using a 50% de-emphasis template in GliderScore that will help compensate for the change in normalization. Will post the new results for review and comment soon.
Mar 11, 2018, 02:32 PM
NorCal Electric Soaring (NES)
SoaringDude's Avatar
Update: please ignore these printouts. Issue has been resolved. --Chris As explained above the attached PDFs are the result of rescoring using a 50% scaling on landings. This iteration has no normalization--am doing another sample with scaled landings + normalization. After a quick review this looks closer to what our old scoring system would do. Not 100% the same but closer. NOTE that in the Flight Scores printout you can see the effects of the landing scaling.

In the meantime I'm working with Aric who has agreed to reenter the current scores into the old system for a more direct comparison. Please keep in mind we cannot duplicate it exactly the same but it feels like this iteration is much closer (I hope!).

Chris
Last edited by SoaringDude; Mar 11, 2018 at 05:15 PM. Reason: Clarified that this iteration is using no normalization
Mar 11, 2018, 05:05 PM
NorCal Electric Soaring (NES)
SoaringDude's Avatar
A G.S. setup error on my part was responsible for the landings being normalized. Will post corrected printouts as soon as I get a thumbs up from several who are reviewing the corrected scores. My sincere apologies.
Mar 11, 2018, 10:17 PM
Good for what ALES you
awilmunder's Avatar
Thread OP
My reply to the email discussion seems to have gotten lost, so I'm posting my reply on this thread...

In short, I threw together an Excel spreadsheet so that Chris and I could compare results between the earlier SVSS scoring method and the new ones created by GliderScore. I'm including my spreadsheet and you will find that the differences are now less than two one-hundredths apart, possibly due to rounding differences. I think this issue is looking like it is resolved.

Thanks all.


(Spreadsheet wouldn't attach, so I included a photo)
Mar 11, 2018, 10:29 PM
NorCal Electric Soaring (NES)
SoaringDude's Avatar
Aric, thanks for putting in all this work, much appreciated and glad GliderScore is checking out now.

BTW I just checked the Google Groups list and there are no email messages hung up in the queue so it looks like your last email just didn't make it through. No idea why. If it doesn't get through after a few hours I'll forward a copy to the list.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Question: How are your local thermal duration contests organized and scored? Miami Mike Sailplane Talk 16 Jan 30, 2018 05:56 PM
Idea Fundamentals are Fun Cyberdactyl Life, The Universe, and Politics 0 Jul 04, 2016 10:14 PM
Discussion Fundamental Changes: Govt employees must be liable MtnGoat Life, The Universe, and Politics 27 Jan 02, 2016 07:02 PM
Question How does scoring work/calculate gtrbaby Hand Launch 13 May 30, 2013 03:11 PM