I would give these 5 stars if the pedal pods were more durable.
If you train with more than one bike, it makes a lot of sense to use the Vector. I primarily race track, but I train often enough on a road bike that it's handy to have one power meter that transitions easily between each bike. This process isn't quite as seamless as you might expect, and while it takes some getting used to, the convenience of being able to use one power meter across multiple bikes is priceless. Compare the size of other power meters, and note that the Vector is considerably smaller and in many ways, more elegant.
The installation isn't too complicated, and Garmin put together a series of excellent video instructions to ensure you're able to get the device up and working. This is a matter of putting the correct number of washers along with the pedal pod onto the spindle (this takes some trial & error to get right; you don't want the pedal pods flush with the crank, and some users advocate more washers than less, but too many can hamper connection between pod & pedal) and fastening it to the crank arm and connecting the pod to the pedal through a short cord. It takes only a bit longer than installing normal pedals. You'll likely read elsewhere that the torque needs to be exact, and that is indisputable--too little torque, and power readings are way off. In my experience, off by as much as 100W, and jumpy. Using a torque wrench with the Blackhawk crow's foot 15mm adapter simplifies the process. Basically, it's a matter of tightening the pedals firmly but not over-tightening, and it's worth being exact to ensure consistency from ride to ride. For me, the required torque is a bit tighter than what I'd normally do. Why so specific? It's a matter of getting them in tightly enough such that the installation angles don't change under the force you put it into them while riding. The pedal pod, which dangles between each pedal and crank, will function in any position, but you'll need to align it so it connects to the power measuring device (through the backside of the crank) without issue. This isn't complicated, but it can be tricky at times, especially if you're using too many washers or pulled the angle at which the pod hangs too far when torquing the pedal "just so." Any pairing issues I've had were due to my errors in installation. Generally, any snags with getting your bike computer to find it are usually solved by spinning the cranks some (ideally by actually riding, I've found that putting torque into these solves most issues), or if that doesn't work, swapping sides for the pedal pods. Garmin notes that it is possible for files in the Edge to be corrupted, which makes pairing across varying bike profiles difficult if your software isn't up to date.
Next up is calibration. You'll need to calibrate every ride, and fortunately, it's done quickly. After each installation, you'll have to set the install angles, which you'll be prompted to do--it's worth noting that you need to clip in and pedal, if you just spin the cranks you'll throw off the power readings. After setting the installation angles by pedaling between 80-90rpm for about ten seconds, you'll be given the opportunity to calibrate. Unclip, then let the cranks sit in the horizontal position, wait for the torque reading to hit 0.00 (it will vary minimally every couple seconds perhaps), and then proceed to the second part of calibration, which you begin after selecting "RIDE". Spin the cranks backwards eight times smoothly and slowly, around 30rpm, without torque. This is crucial: smoothly and slowly. If you go too erratically or rapidly, it'll throw power readings off wildly. I've started doing this by hand and it's been fine. It's also not necessary, because the first step of calibration will deliver the standard +/- 2% that's the norm for any power meter, and 99.9% of cyclists don't need to sweat such tiny disparities. You must calibrate the system every ride, a process that entails opening up the bike profile section, pairing the Vector with your computer, and selecting "calibrate." If you're doing a triathlon and leaving the bike overnight, the calibration should hold for perhaps up to 12 hours, so you're not fiddling with it during transition.
Quick note on the pedal pods. You've likely heard they're fragile, and that is absolutely true--one of mine broke on the first ride, which was an easy 60 minutes on smooth, open road. Frustrating, but then again, in my first race with them, I caught the back part of a large crash, rammed into another racer's floored bike, went over my handle bars, and probably put the bike vertical in the process (I mostly remember hitting the ground, not how I got there)...and the pods somehow survived that. In regular, controlled riding, the pedal itself will hit the ground before a pod will, and the pedal is the well-built Look Keo model. I've ridden these through wind, rain, and puddles without incident. When you're carrying your bike around, you'll likely be much more careful with it, but I've never broken a pod in non-riding situations.
Garmin's support with issues has been great. I've not tried updating any firmware yet, and they're accommodating with the pedal pod replacement. Candidly, for awhile I just fastened my broken pod to the crank with electrical tape and that worked without issue.
I can only compare power outputs on the Vector to my old wired PowerTap, which was quite reliable but probably near the end of its life and not very convenient. Whatever discrepancy exists is minor under normal riding. I have not compared Vector's readings to Quarq or SRM, though the consensus is that it tracks along very well with SRM. I've yet to find a credible source regarding the sampling rate, but I've seen enough evidence that readings are nearly identical to SRM's that it's likely not worth sweating. Effective cadence range begins at ~30rpm, so if you're Man 1 in the team sprint looking to measure every .25 seconds of your start, these are not a good option. Given that that's approximately .00002% of the cycling population, this isn't a concern. I do a fair amount of super-high RPM work on the indoor trainer, and going from 0 to 190rpm in a under 2 seconds will not register, but that's not a session in which you're really concerned about power anyway. The greater issue, especially for using multiple bikes, is not so much accuracy but consistency--if it's off by 2% (the upper end of the accuracy margin), as long as it's off by 2% whenever I use it, that's sort of acceptable. The installation process is relevant here again, because if you don't torque the pedals exactly right, readings will be off, and if you don't calibrate properly, readings will be off. There's a reason I repeat this so frequently, it's that important. If you switch bikes regularly, accept that it will take some time for you to get accustomed to this, but once you do, it's second-nature. The ability to track left/right power balance is interesting but I have no idea what to do with the information. In one ride, it also said I was on a 63%/37% left vs. right split which I somehow doubt. This may result from how the pods are installed, and the angle at which they hang--yet another quirk about the system.
The greatest drawback to the Vector is it's occasional finicky behavior and comparatively complex protocols. This is not a "set it and forget it" system if using multiple bikes, but it's also not mind-boggling. If you only install it once on the one bike you use for everything, then that's not a problem--get it done right the first time, however long that may take, and you're set (don't forget to calibrate each ride!). Until you break a pedal pod. Garmin will replace them under warranty, but it will take 10-14 days and you have to send in the broken one, and new ones cost $70. If they can fix this, you're looking at a completely amazing system.
Overall, the Vector is a phenomenal idea and mostly executed with expert precision, so just be patient with learning the in's and out's of it and you'll get plenty out of it.