The biggest difference for a track bike is that the Garmin does not allow you to turn off auto calibration. Every now and then you notice that watts are a lot harder to get than they were a minute ago and realize that the head unit auto calibrated when things were not really at 0. On the road it works ok with powertap and quarq, but SRM seems to send calibration when still under light load - not sure if it is a low cadence thing.
As far as sampling rates with ANT+ my understanding is that the power meter is sending its latest values at some interval. The head unit is looking for the last packet at some interval. If the power meter is sending more often than the head unit some packets will be ignored. If the head unit is sampling faster than the power meter some will be duplicated. So it is somewhat of a lossy protocol with the expectation that the values are not that different and it all works out in the averages. As the intervals get smaller the samples in the average go down so the reliability of the numbers goes down too - kind of like looking at GPS speed data - get too short and the values jump all around but go over a longer time and it all works out.
The wired units can sync the send/receive to run a no-loss protocol if they wanted to easier than acking wireless signals which would require the sender to also listen.
If you really wanted fine grained data you would need multiple wheel magnets and cadence sensors to get a more frequent signal on the sensors. This is where the Stages and Garmin units might have an advantage with the motion sensors being able to determine intermediate crank locations. I think the track and high end SRMs used to have (still have?) multiple cadence switches to get half rotation values.
With a base SRM with once per crank rotation signal you are not gaining much power data if your cadence is low. From 60-120rpm you are dropping some rotations, over 120 you are missing half with 1/sec recording. So you might miss the first pedal stroke or 2 of a jump.
Powertap does a time based accumulation and sends once per second (wireless - 1.25 sec wired) so in theory you don't miss things with a 1/sec receive but clocks drift and signals are missed so a 2/sec receive would usually show 2 duplicate values but can avoid missing the stray value.
This is why if Stages is updating 64 times a second it is important to know what interval they are averaging over in the packets.