Can someone who trains powermeters explain how important ACCURACY vs. REPEATABILITY are?
Sure. They're both important, but repeatability is much more important. Accuracy is important for a few reasons: quantifying the specific demands of an event (and tailoring your training accordingly), and comparing long-term changes over a period of time (I have almost 15 years of power data, and I've used a lot of different power meters over that time. It's nice to be able to compare the data and establish long-term trends).
If you're trying to peg a time trial based on a consistent pace you've hit in training, does it matter if you're 10W lower than what it actually is?
No, not really, as long as the difference is constant. And therein lies the problem...
Long-time average power wise it seems to follow pretty reasonably, but in individual sprints, the numbers are way off. I know that software uses various filters over the power data to interpret your workout, but if the STAGES meter incorrectly in the same manner for every workout, what effect does it have on your training decisions?
Again, to the bolded part, that's the problem. Pedaling asymmetry isn't going to be the same for every workout, or even WITHIN the workout, so even if the PM works perfectly, it's not going to present the data in a repeatable manner. As far as what effect it'll have on your training decisions, here's a real-world example. I typically do vo2 workouts to failure. If I'm shooting for 350W for 5 min, I'll pull the pin if I can't hit 325 or so. I've seen real-world examples of power files where the agreement between a single-leg measuring device varies by +/- 8%. I don't know about you, but to me there's a big difference between 322W and 378W. As someone who adheres to the philosophy of doing only as much as necessary, I like to know when my power is reduced to a certain level, so I can move on and start resting. That's why I use a PM in the first place. I device with that sort of variance doesn't have enough resolution for me to be of much utility. I'd prefer just to go back to a hill and a stopwatch (seriously).
The way I see it, you're still getting a direct measurement of power, so you don't get delayed measurements or transient noise that indirect powermeters (IE: powercal, iBike) provide. But I don't train using a powermeter, so I don't really know what kind of effect this would have.
Obviously if you want to claim you had a 2000W sprint for bragging rights, an inaccurate meter is a nogo.
Well, none of the power meters actually "measure power directly", i.e. they all make *some* assumptions. In my experience though, the assumptions made by the Powertap and SRM are small enough that the data are useful for training and competing. Devices which measure power (indirectly) and then double it put it into the category of "power guesstimator". Power guesstimators are neat devices, and I enjoyed playing around with the Ibike aero, but I certainly wouldn't use the information to make training decisions.
Here's some more reading on pedaling asymmetry, this going back even longer: http://www.ncbi.nlm.nih.gov/pubmed/979569?dopt=Abstract
" onclick="window.open(this.href);return false;" onclick="window.open(this.href);return false;
This should answer the question of "how likely is it that I'm just having a 'good left leg day'". Truth is, it's pretty likely.
And again, this is old, well-established research. It's why I didn't even bother reading this dude's review. Frankly, it's not relevant. Even if the device worked perfectly for him, all it would mean is that he happened to have perfect pedaling symmetry during the test conditions. It wouldn't tell me how the device would work for me or others, or even how it would work for him under different circumstances.