(South Oxfordshire). Been tracking WW Solar f/c and Solcast f/c (via GivEnergy portal) against actual yield since 1 Jan this year.
Overall, on average (282 days) WW overestimates by 10% and Solcast by 20%.
But split it into days where yield is >= or < 10 kWh.
If actual yield >= 10 kWh, (198 days), then WW underestimates by 20% and Solcast is spot on (on average).
If actual yield <10 k (84 days) then WW overestimates by 90% and Solcast by 50%
I track this at about 0800 each morning. The final WW f/c figure in the dowloadable file will differ. Makes little difference when yield >= 10 kWh, but when less the final WW f/c is better but still 60% overestimated.
Yesterday (9 Oct) and 3 Oct were two recent days when f/c was very over optimistic, but when looking at very small yields, (<5kWh) I dont think it matters.
However, I'm on Agile, and can't make use of the f/c (except manually), so this is academic for me. Next step is to leran how to calculate and plot standard deviations etc.
