@Jem_Arnold - thanks for the tag, and apologies for the delayed reply. I took a brief post-TDF holiday to, incidentally, visit my good friend Uli Schoberer (the inventor of the [modern] powermeter).
Before I start, @Robert_Chung has covered a lot of good stuff already, so I’ll try not to overlap too much.
From professional cycling, I think there’s an acceptance that power metres do have a little variation between devices, even of the same model. And we (reluctantly) accept this variation.
Why? Because when a team has around 80 riders (Men’s, Women’s, WT, development, etc), and I’d expect 6-800 power meters (maybe more) – it’s almost impossible to be super fussy about the last few watts. They’re either “close enough” or clearly “wrong”. When one looks wrong, basic checks are performed, including batteries and zero-offset. If it’s still wrong, it’s replaced.
My personal view is that for general training, “close enough” is good enough. I appreciate the day-to-day variation in both measurement and human ability. Indeed, I triangulate a few things to guide training – typically, power, heart rate and perceived effort.
For power and heart rate, I have “zones” set up on my Wahoo head unit, and it displays both the number and the zone as a colour. I use the colours as my guide for training. E.g., for “endurance training” I want to see blue or green (from HR), for “threshold intervals” it’s orange (from power). I also look at both power and HR, and how they’re tracking relative to each other (and perceived effort) and adjust based on that feedback too. I don’t want to get into the weeds here, but this helps to manage effort based on fatigue, temperature, fuelling etc.
I have another page set up with a histogram for time in zone by power and heart rate. I get a “quick look” into whether I have been doing what I intended to do in this session.
So while riding, I’m not micromanaging things too much. I’m using the numbers and colours to guide my efforts, and I’m letting my brain and legs figure the rest out. I’m the same with “time” too. While I may plan to do 10-min intervals, they’re never to the exact second. +/- 20% is fine.
Where I do want accuracy and precision is in the saved data. As discussed earlier, I want to be able to answer those deep questions and if the data is clearly wrong, that’s bad. For this reason, I’m lucky (thanks Uli) to have 4x SRM powermeters which all get checked for offset and slope calibrations*.
I currently have a Garmin in my jersey pocket as a data logger collecting GPS, power, HR, ventilation, core temperature and (sometimes) muscle oxygen data. I do not look at this daily, but I might want to look at it over time and so I ensure that everything records well.
So in short: Day-to-day: Accuracy matters less to me as a training tool. Long-term: Accuracy is super important, as it enables us to answer questions from our data.
*the biggest source of inaccuracy, in my opinion, is people calibrating their power meters. It should be performed very carefully, and only when something is clearly wrong. NOT before each ride, or interval. I think over-calibration adds more uncertainty than an inaccuracy in a power meter itself. If it’s working fine – leave it alone.
