When you are riding outdoors you usually want to preserve momentum and not keep power constant. You want to match momentum to match an average power target.
You also need to forgo power targets if traffic demands it. If I overtake something in an iffy situation I might go to Z4 for some brief moments.
Not really. Outdoors conditions can change really quickly and really rapidly. Me passing by a row of trees can easily change my power output by 50 W. Keeping power constant during accelerations and deaccelerations is not useful. Riding outdoors ≠ riding on the trainer. That’s why outdoor workouts are different from indoor workouts.
I didn’t read through the entire thread so forgive me if this was covered before.
I recently swapped trainer setups from a road bike to my XC. I did a 1 hour Zone 1 spin this morning and to my surprise, I had a lot of trouble getting my cadence up to where I expected it.
Old setup ~92 feels comfy and New setup ~90 was difficult.
So my conclusion is that bike fit has a lot to do with this as well. Or maybe bike familiarity. Or both!
What gearing were you using between both bikes? That can have a real impact on pedal feel.
And yes, bike fit may also be a factor.
Like others have said, you should always include 0 W for your power averages, but discard 0 rpm for cadence averages.
Freewheeling = 0 W, so you are dragging your average power down. But 0 W may be what is called for due to traffic, traffic lights, etc.
Was that on the trainer or outdoors?
If you were outdoors, I have a pretty obvious guess: wheel inertia. I reckon that on a road bike you were going much faster, so your rear wheel was spinning more quickly, which means it has a larger inertia. Most riders prefer a higher cadence when the wheel inertia is high. On your mountain bike, you tend to ride at lower speeds with lower wheel inertia, which means you would prefer lower cadences.
Does that soudn about right?
It does, and it impacts the analytics software. From an analytics software point-of-view, I’m pretty sure exclude power zeros behaves like auto-stop on your bike computer.
For example a ride on December 30, 2016 with auto-stop enabled, and I stopped for a couple minutes mid ride.
WKO peak 1 hour power is 223W average power / 243W normalized:
^^^ but it only has data for 56:06, because I stopped mid ride.
Here is the “223W for an hour” in Intervals to show the issue is not just in WKO…
basically the same, average power is 222 and normalized is 244 (versus WKO, Intervals is -1W for average, and +1W for normalized).
Ok, so both WKO and Intervals agree on average and normalized power. Because they don’t create data and will not insert zeros for power.
What does Power Duration Curve show?
When you include stops (no power data), I only did 207W for an hour.
So what timeframes did I actually average 222W? From power duration curve its 41 minutes. Not an hour. Just 41 minutes.
Its crazy making, you can’t get the power duration curve to match up with timeline selections. Forget about comparing different efforts - I could have two efforts that “averaged 222W” and one could have 4 minute stop and the other could have a 30 minute stop.
And an aside - how hard you are pushing on the pedals is called force or torque. Power is both torque and cadence. Changes in cadence, for example 60rpm vs 90rpm, means that I can be “pushing as hard” during:
as during
rough estimate on the cadence numbers, from my own recent data.
The latter may be true, but ignoring zero values is most definitely NOT the solution. As I said, all it provides is a distorted view of reality.
Real life example #1: back around the turn of the century, I knew an SRM user who never downloaded the data from his PowerControl IV (which didn’t allow you to include zero values for power…that only became possible in the head unit with the PC V, proving that, in fact, Uli Schroeber did actually listen to his customers/read the wattage list). As a result, he thought that he averaged a Lance-like 400+ W during criteriums.
Real life example #2: by your thinking, I too can legitimately claim to have averaged over 400 W for 1 h. How? By doing 15 s on/off microintervals, then ignoring all the rest periods that enabled me to generate 400 W during the “on” periods.
Once again I sense the pernicious influence of the misconception that there is some sort of magic to training in specific “zones”, such that you need to constrain your power output to a narrow range to induce adaptations.
I’ve said it before, but I will say it again: not only is that not true, it is counterproductive to being a good cyclist. You NEED to be able to increase (or decrease) your power rapidly, something that being a “trainer drone” does not help with and (if done excessively) can actually impair. To then apply the same thinking to cycling outdoors is the surest route to “triathletehood”, i.e., to being such a diesel that you are OTB at the drop of a hat.
To be clear, the issue here is your inappropriate use of the autostop feature, not how the software handles the data.
Somebody really, really, REALLY needs to write a book explaining best practices when using a power meter, and how the data - properly acquired and analyzed - can provide important insights into the demands of racing, how to train, etc.
here are the 3 good ones that are always within easy reach:
the first one on the left, it was so good I bought the updated 3rd edition in the middle as soon as it came out!
Took me about 9 months after buying a power meter to turn off auto-stop. Almost 6 years of non-stop data collection. A shame I didn’t turn it off sooner.
I went back to when I started riding 9 years ago. Easy, hard, inside or out. My cadence was 80/81
I checked a few of my rides in the last week, my cadence was 92/93. I had one TR workout that I did and still do, Petit. My 90k, 3 hour group ride on Sunday was 93. My solo 50k ride last week at the same speed but higher power was 92. I now ride generally with an over 90 cadence.
I dont know if it was self selected as I had to work to increase my cadence when I started. I found muscle fatigue was my early limiter so a higher cadence just resulted in an easier ride. It required me though to monitor cadence in my early riding to keep trying to find the gearing that allowed a higher cadence. At some point I just became used to riding at a higher cadence.
I do think there is a correlation for me with fitness and cadence.
Makes sense. I’ve seen many new riders start off with low self selected cadence and slowly work it higher (without consciously trying) as fitness and experience grows.
Different gearing. My normal trainer is an SB20 Stages bike. The XC bike is on a Kicker with SRAM 10-50.
Thing is, even at very low power for me, I couldn’t get my cadence up on the XC bike
Indoors
Mountain bike cadence is lower for the same reason road pros grind over the cobbles in Paris-Roubaix, it helps smooth out the bumps.
So what does the average power that you produce only when pedaling tell you?
It doesn’t tell you anything about how fast you traveled relative to the conditions (e g , wind, terrain).
It doesn’t tell you anything about the physiological demands.
It doesn’t tell you anything about your physiological capabilities.
It doesn’t let you validly compare different training sessions.
As I said before, it’s good for one thing, and one thing only: ego strokes.
I’m the opposite. When I started TR my FTP was 193s and my cadence was around 95 rpms. I regularly spin into the 100s and preferred it. 5 years later my FTP is 317 and my cadence is low 80s. Over 90 feels fast!
In the beginning I didn’t have muscle endurance so I relied on my cadence to avoid that annoying heavy pressure on the pedals. Now I actually look forward to pressure as it’s comforting and familiar.
I trust this was a deliberate continuation/acknowledgment of the tongue-in-cheek comment, @WindWarrior? 
Otherwise, 
And why does that matter?
That including zero values brings down the average is self-evident. The points you seem to be missing are that:
it increases the power you can generate while pedaling, and
it reflects neither the physiological demands (which are why drive adaptation) nor your true capabilities.
Most people generate very little net power during the upstroke - typically, only enough force is generated to get the trailing leg out of the way of the rising pedal. So, maybe we should also exclude those brief periods of “coasting”, and only include the downstroke in the calculation of “average” power? I think that most folks would realize that that is absurd, but it is the logical extension of your way of thinking.
TL,DR: Almost all exercise is actually intermittent in nature. Muscles are stimulated to briefly contract, but then relax (“coast”) until they are stimulated to contract again. Excluding any period of rest is therefore an arbitrary decision that doesn’t reflect reality.
I’m sure that it was.
However, I am a bit wounded that he threw Skiba’s book in there, rather than all three editions of ours.
