FYI - I did a comparison a while back between my Sense Solar and Tesla / SolarEdge data, and helped another guy do the same. Both found something similar, a 3-4% “optimism” on the part of the inverter. I assume that the inverter looks at only power produced and doesn’t subtract losses or energy needed to run the inverter. The question I don’t know is how much of that gap is fixed, and how much of it varies with input power.