Hi, I have a question about the power spike fixing. I did a sprint at the end of my ride today and it identified a spike as seen here. I understand how the spike fixing works, but my question is why is it correcting to only 160w? It seems to me like it should be correcting to somewhere in the 700w range based on my power curve model (or not correcting at all, as 30% over 700 is 910w). Was wondering if anyone could give some answers, thanks.
It slides a moving average over the data and when it finds a spike it uses the average of the power before the spike to fix it. Thats quite a long sprint so the moving average window will be quite large hence the low power.
You also have HR spikes on that ride so if your heart rate monitor was working properly you need to raise your max HR.
That makes sense, but I’m just curious how it decided to correct all the way down to 160w, and only for that little chunk. 160w is definitely below my ftp so it seemed more like a bug to me, especially because I have had power spikes on other rides that are corrected down to values that make sense based on the model curve.
It uses moving average data from your actual power trace to fix, not the model curve. Real power spikes might happen when you are toddling along in Z2 at 200w so rather replace with that than what your power curve says you could do.
Ah, so the 160w thats being replaced in this case is because I was sitting around there for the 5mins or so before the sprint and the increase in power is so drastic that it assumes its a spike and replaces with that moving average of the power before the spike? Is there any way to prevent this besides setting the power spike correction to a high enough % that it doesn’t fix it anymore?
The last time I did this sprint it also marked it as a spike, but for a much shorter window, and only corrected for 3 seconds to ~570w.
The best is to just set the threshold high. If the power is real you don’t want it clipped!