I don’t think it’s important at all. Assume that it’s all real. Assign weigh inversely proportional to variability of dataset.
I’m not sure what variability measure you want to use though. I want to say Coefficient of Variation, but I don’t remember that much from university statistics. (Have I even passed it? 
)
Though according to this theory you should heavily discount my very random thoughts! 
Sounds like you are talking about riding in ERG mode, where it takes time for power levels and resistance to change? If you ride in slope mode you should be able ramp power more quickly up and down, just like outside… the Neo can capture power surges, like most modern / direct drive trainers (if it were wheel-on you may miss something due to tyre slippage, but not the case here). For sure in slope mode you could alternate between 30s at 2 or 3 x TP and 30s easy quite a few times, or 10s all out and 50s easy (depending on your signature of course) if you do want to test the power meters. Can even push for a BT if you are keen 
Anyway, sounds like your most recent ride outside gave better results which is good. Could be due to the nature of the ride as you say… wonder if it could also be due to @ManofSteele’s signature tweak where HIE was increased to more realistic levels (from the other thread)? But that’s one for the Xert guys to comment on.
That explains it.
I remember those days. Now I am more inclined to be “uh-oh, better save something for later”.
You are probably right about ERG, but still motor power response is order of magnitude slower then strain gauge. I will give it a try anyway (when I find enough motivation for this 
)
Update. I did another ride today. It was actually much easier then yesterday. The group was quite mixed, and for the most part I was noodling along, enjoying a nice, cold sunny day, but I did a few hard efforts, which was easy since overall level was low. Got a gold BT and both PP and TP bumped up by 20W 
Two differences I could think of from yesterday - I had autopause on today (but Armando says it makes no difference), and yesterday’s overall level was higher - not as much difference from baseline to max effort.
What I’m going to try is manually reduce signature changes from activities like this by 3/4. This is very arbitrary, but probably less arbitrary and more consistent then randomly mocking around with HIE.
I wish there was less arbitrary way to deal with it though… 
1 Like
Yeah something really strange about your data. Your PP is very low for someone with your weight, TP and HIE. This can cause TP to be a lot higher, especially when BTs are at high wattages where PP is contributing a very high proportion of the effort (e.g. PP accounts for roughly 1/3 of the power at your 1 minute power). Your efforts are also very spiky and errors in the data will compound the errors in the signature.
Which power meters are you using? P2Ms are like clockwork (never see issues like this). Some older Quarqs can be more problematic. Left side only PMs can be superb or troublesome, dependng on a variety of factors. Powertaps are great usually - hubs and pedals. Older Garmins … umm… Assiomas are great but abrupt stoppages might let an extra second or two of data creep in which can throw off things during a BT.
2 Likes