On Monday, December 21, 2015 at 1:12:42 PM UTC-8, Andy Blackburn wrote:

Typos fixed for clarity - sorry for the long post.

On Monday, December 21, 2015 at 6:31:33 AM UTC-8, Papa3 wrote:
Andy,

I'd argue that you simply don't have enough data points to make factual statements like the above. The majority of racing pilots in the US have between 1 and 2 years of experience using FLARM and are "hamstrung" by current user-interfaces that are optimized for the original purpose of FLARM - collision avoidance.


Perhaps not - but as far as I know I am the only one who bothered to generate several hundred datapoints from actual contests to test several of the theories that were presented for how Flarm leeching benefits competitors, specifically:

1) that competitors following by "Flarm non-stealth distances" (i.e. 2km) can follow the leader and core the thermal faster. Data - the opposite is true across multiple contests looked at including dozens of pilots and hundreds of thermals - later to the thermal generates poorer climbs and pilots who with a higher percentage of thermals they found themselves have statistically significant higher average climb rates over the entire task.

2) That pilots use Flarm to deviate to better course lines or saving thermals that help them speed up. Data - dozens of flight traces and hundreds of thermals over several contests show only a handful of cases where pilots changed course within 7 miles (maximum reliable non-Stealth range) to reach a thermal that they would not have come across with stealth on (that is, a course deviation that generates more than 2 km displacement from the closest approach to the thermal without the deviation). Pilots tend not to make big deviations for very far so the stealth range gets you visibility to most thermals anyway, just slightly later and most of the time for very little additional distance penalty. Additional Data - these deviations adhere to the data from 1), above - the below-average climbs make the deviation not worthwhile almost all the time - in one memorable case a pilot deviated 45 degrees for three miles and made a nice clover leaf search pattern exactly where the other gliders had been and found nothing.

Feel free to look at logs and provide your own analysis. Also look up Bayes Theorem. Based on the results from my (laborious - ugh!) analysis of flight logs from actual contests I would say we have sufficient data to conclude that there are not big advantages conferred by Flarm leeching.



It's absolutely foreseeable that one or more of the instrument manufacturers (or more likely the Open Source crew) will put some real thought into "Tactical Information Pages" (TIPS) for 2016. As discussed many times, the main purpose of those pages will be to apply "smart filters" for the raw data which is hard (though not impossible) to interpret on today's screens. Whether it's competitor location data, historical track data, location of the "lead gaggle", whatever, that information WILL be used, and the pilots will be able to configure that information to their liking. So, instead of a single "blip" or two from a target, the screen will show the average top-to-bottom climb rates, number of samples, etc. Another might show you only selected "targets" that you want to keep track of in the start cylinder. Will it drastically alter outcomes at the top? Not likely. Will it further compress the middle? Probably.


I dunno Erik - the information is so perishable when you are trying to make time typically by only taking top quartile to top decile thermals and I'm yet to find a race where thermals are continuous columns of lift with no cycling or variation. Not all processing of data yields useful information - and if it is bad more often than it is good information you are better off ignoring it.



If you want an interesting read, have a look at the post from Tom Arscott winner of the Junior Worlds Club Class. In a competition where the start was absolutely paramount, the ability to find (and hide from) competitors was make or break.


Read it - it was a classic gaggle day. I'm not sure it is a superior situation to have winners based on who is able to deploy team spies to find the German team (or hide from theirs). This all happened apparently with Flarm voluntarily in stealth mode - or so they say. The amount of "pickup" team flying that goes on out of the start is considerable - particularly where you don't allow starts out the top and finding a good initial climb is paramount. Pre-start jetting around the start cylinder to find the good pilots and stick with them has always existed if it's going to be a gaggle day - adding stealth makes your ability to do a systematic search for the good gaggle all the more an important determinant of the outcome. Is cylinder search technique the skill we want to test in contests? I'm not convinced - but clearly people went to some effort to do it at the JWGC.



Now back to sanding primer...


Yo - good luck with that Mr. White! (Erik is in full "Breaking Bad' attire for his refinish project)

9B