Power comparison with Vector 3 Pedals

Of course Vector3 is not a gold standard, 17kg may not be enough for static calibration (btw. 11kg or higher is mentioned in Garmin’s instruction of static calibration procedure), drivetrain configuration has influence on the generated loss, and so on and so forth. But we agree that 2-2,5% is minimum loss which could be used already as a compensation factor which would eliminate probably majority of the complains, so why not doing it? I’m sure there are many clever heads in Garmin to develop respective algorithm.
300W measured with V3’s for someone weighing 75kg is 4W/Kg but in my example NEO would measure ca. 290 resulting in 3,86W/kg. I’m pretty sure climbing AdZ would result in much bigger time difference than just few seconds (closer to minute or more). Zwifters will agree that in racing scenario it would be quite significant disadvantage.
I may be too picky but still think this should be addressed somehow. I expect €1300 device with 1% claimed accuracy should perform better than low- or mid-range trainers.

i'm in the same page 

Btw, I had the same problem not long ago, my Assioma Duo has been doing up to 10watt difference no matter what. I noticed however that in my granniest gear, (big cog back, small front) the derailleur is very close to the body of Neo and while inspecting the chain line which should be straight, was actually quite crooked and not straight. I found a nice video by Tacx Faq, which shows how to insert 2 small ring spacers to Neo freehub body by removing cassette. I've added the spacers which pushed the cassette a bit more forward, making a bigger space from the derailleur and making the chain line far more straight. I've measured my results after and lord behold I am getting really a lot better numbers now. Seems there is very little difference now between the two readings. 

Maybe check if that helps you too here's the video: Tacx Neo Rear Derailleur Touching Lower Housing - YouTube

Also note that I have 11 speed shimano cassette which shouldn't require any spacers by default, but I guess it depends on the bike also, so check how much distance you measure from the body of Neo, could be your chain is not exactly perfect straight in gears and making your loss bigger than it could be.

"why not doing it? I’m sure there are many clever heads in Garmin to develop respective algorithm." Neither the Elite or Wahoo do this, if You want You can downscale your V3 pedals to ~98% and problem solved. Ahhh no, it is not good, because it is a LOWER number, the only LOWER what is acceptable is the number of the body weight on the bath scale :)  just a notice: in the real world your speed is coming from the power of the rear wheel, even if you put 300w on the pedal, but it is only 290w at the rear wheel your speed will be "calculated" from 290w ... so probably once the Zwift races will go with trainer power only and this problem will be solved.

Yes, the Garmin suggests 11kg or higher but 11kg (or 17kg) is not sufficient for perfect calibration , the error margin could be high, the head unit resolution is 0.03NM so apprx 20gramm (18gr) error in the weight can cause calculating error, usually the gym weight or dumbbells are not exactly has the weight as written on it, the bath scales also have error margin 100gr minimum or sometimes 300gr, 100gr difference (with 17kg) is 0.5% error adding to the pedal's accuracy (so You can easily adding extra 0.5% to the measuring, at 200w is is 1w, ok, not too much, the strain gauges in the bike powermeter can miscalculate more if the offset is not set right ) , if You do it with 11kg, it is almost 1%, if with 40kg, it is only 0.25%. And You can make mistake also when You try to finding the absolute highest value , this also can cause adding a few to the pedals... this is why the dynamic bench calibration is the best method to adjust (or validate ) it.

You have 5,6,7watts difference at 20,10,5min , i dont know what was your gearing during this test, i assume it was a sim mode ride , if it was on big ring front and mixed at the rear, that is totally acceptable , i have 5-6w difference at 250-300w range with more Neo1 and with more P2M NG , on small ring it is 3-4watts, everyday (yes, exactly the same difference day by day if i do everything same /powermeter calibration, cooling, etc/ ). You have 1-2w more, on lower watts, it could be your drivetrain , a faulty offset in the V3 (usually the bike powermeters measure wrongly due to the strain gauges, if the offset is not perfect, it over or undermeasure the power, the Neo always reports the same power, no offset difference day by day or activity by activity, lots of people has one day +5w offset for example with a bike power meter, another day minus 2watt, 7watts difference in offset , hmmm this kind of powermeters are toys and not real powermeters , that is my opinion)

i wrote the AdZ difference with your power, i didnt think (from your chart) that You can climb with 300w for 50mins . Yes, could be 40-50sec difference with such a power difference. But in Zwift races it is allowed, they ask You record with 2 powermeters (for validating) but You can use the higher one even if it is a damn single sided InPower which overmeasure the power with 5-10% and between the trainer and it has 10-15% difference ... But nobody will win a race with extra 5-10watts at 300w , that 100% sure, with 30-50watt, maybe yes, because he can "rest" more until the final sprint. (but needs good sprinting ability too)

here is an example, Drivo2 (0.5% certified accuracy, better than Neo's 1%) and Quarq, lets say it is enough accurate too, has 10-12w difference , problem? :) https://www.zwiftpower.com/analysis.php?set_id=72383  and he finished just at 20th place... 

and here is an another, Neo2 vs V3, and the Neo2 is HIGHER with 8w than V3 , now what shall we do? :) probably he used a 1000USD UFO chainlube/chain/joykeywheel stuff and saved 10watts and gone to minus :D www.zwiftpower.com/analysis.php

me too, Ride On! :) 

Thanks for the hint but I don’t think proper cassette installation is the problem in my case. I too have Shimano 11 speed (Ultegra Di2 with the same cassette on both NEO and the wheel) but it doesn’t touch the NEO housing even when chain is on biggest cog. Shifting works perfectly fine and I don’t even have to correct indexing when changing the bike between the trainer and the wheel. So no problem here.

I also didn't have exact touching nor shifting problems, I just observed the gap is very close and the chain is very much not straight,. Worth to check how straight is the chain lines in small - big, big - small, but yeah if it's fine like you say might not be a thing

Exactly the same here. Neo 2T vs Assioma pedals :(

Can You provide a comparison link?(ZwiftPower 4example) what kind of chainring do You use, round or oval?