Just a quick one today. The last few days I've had a few questions about the accuracy of my Vector powermeter. I've got the generation 1 dual-sided model with the generation 2 pod upgrade. I've ran them with my Quarq Elsa powermeter and with a physiology lab Lode bike ergometer (http://www.lode.nl/en/product/n-a-127/3) during a step and ramp test. They've always been within a few watts of both of those, so I'd say I'm quite happy with the accuracy.

For those with Vector powermeters there are a number of points I've discovered over the past year that MUST be performed to maximise accuracy and repeatability. Firstly, ALWAYS ALWAYS ALWAYS torque your Vectors up. Garmin recommend 34-40Nm. Buy a proper torque wrench and the crow-foot spanner. If you don't do this you might as well just put your powermeter in the bin, the data will be all over the place. Secondly, set your install angles on a very smooth road at a constant velocity and cadence. Smoothness is the key here. Finally, always zero-offset prior to a ride. That's the Garmin 'calibration' in your Garmin bike computer.

From a data viewpoint, absolute accuracy isn't of maximum importance to me, repeatability is what truly matters. If it constantly under reads by 50%, so what? As long as it is constantly 50% low then that's great, because all of my training is based on those numbers. Repeatability, not accuracy, remember that.

There are ways to assess your powermeter's accuracy: analytically and empirically.

Analytically you need a mass of a precisely known value. Typically people use gym weights but the value of the weight is normally much different than stated. If you do use gym weights then ensure you have weighed them with a calibrated scale, such as at a post office or an engineering workshop (university engineering departments typically have great scales). The greater the mass the more accurate your calibration will be. At 1500w and 100rpm with 172.5mm crank arms you are exerting over 830N of force on the pedal, that's nearly 85kg of mass, quite a lot!

Empirical assessment requires the use of physical equations to determine what power output you need for a given velocity in a given condition. This is done by calculating all of the resistive forces experienced by a cyclist: aerodynamic, rolling resistance, gravitational and drivetrain inefficiencies. You can also account for accelerations and therefore inertial resistances, but reducing these makes the calculation more accurate. To minimise inaccuracies due to estimation of CdA and wind, it is best to perform this on a climb of known length and elevation gain with minimal wind or at least known wind conditions. Once you know the rider and bicycle mass, the course profile and weather conditions you can fire all of the number in to a calculator such as this: http://www.cyclingpowerlab.com/PowerComponents.aspx From which you can calculate the required power and compare it what your powermeter reads! 

It's an interesting tool to play with. I was playing around with variables for the Coll de Rates climb in Calpe , Spain that I rode a few days ago. A 1kg mass gain would cost me around 4w, whereas changing from holding the hoods to holding the tops would cost me 10w! And who said aerodynamics don't mean anything on climb?