There is still a significant difference between the trackpoint calculation and the Doppler calculation for Alphas with the GT-11 (perhaps a little less with the GT-31)
The difference is that with the B1115 firmware, the Doppler seems to be even with far fewer strange spikes that were evident in the earlier firmware.
Earlier firmware for the GT-31 displayed a large percentage of invalid Alpha runs, with Trackpoints being worse than Doppler, but both were bad. The new firmware produces almost no strange speed curves/spikes and the Doppler and Trackpoint mirror each other very well. The trackpoint results still are higher most of the time but usually, in my experience, not as much as with the GT-11. Recent runs with both the GT-11 and GT-31 show pretty similar results from the Doppler on both, BUT, not the same. This is a source of frustration and the difference seems to consistently be more that between two GT-31's. I don't know if this will ever be solved. I think we just have to live with it and encourage all the faster guys to use the GT-31 so at least we are competing with the same thing.
As mentioned before, we need some more data from side by side GT-11 and GT-31, preferably worn on the head to eliminate the sharp movement on the arm during the Gybe. Anyone who can contribute, please PM me.
Mark, your speed graphs from RealSpeed look really good and clean. Just like good tracks should look with no mid gybe spikes. This is what one would expect from the newest GT-11 and GT-31 firmware (GT-31 is best in this regard)
It seems strange to me that RealSpeed would find a result with Dopper but not with Trackpoints because both methods use the trackpoints for the proximity. I think you should send that track to Mal Wright for him to examine. You may have found a bug in the software.
On the other hand, recent Alpha results from GPS-Results using Dopper are suspect. Manfred implemented a method of using Doppler positioning/direction data to calculate the proximity. The flaw with this is that if there is even a small error in the Doppler direction calculation (usually during the gybe) this error perpetuates itself resulting in the subsequent positions all deviating from the true position and the further you go, the larger the error. This does not happen with trackpoint position because each point is separately calculated. In this sense it is self correcting.
ps it helps if youve seen the movie zak and mirie make a porno to get that joke...
Hey Mark, thanks for passing the tracks onto me.
As I explained in my reply email, RealSpeed is reading the results correctly. Your best doppler result is performed where the incoming track and the outgoing track converge to the proximity distance at almost exactly 500m. This means there is only one or two combinations of points that meet the criteria for an Alpha in that run, with no result until they get close enough to fit inside the proximity, but also no result past the 500m total distance.
However, there are slight differences in speed between doppler and trackpoint calculations, which means the (slower) doppler *just* makes it back inside 500m whereas the trackpoint result is slightly more than 500m at the same point, and so doesn't count at all in that run.
Its certainly a bit of fun trying to work out why some of these results come about!
Cheers, Mal
hehehe
