When we add new factors after using the app for a while, the old days where we didn’t use that factor (even though we still maybe did it certain days, but obviously we couldn’t log it,) are still counted in the trends and correlations. This creates a skewed result, because it automatically assumes that factor was absent ALL previous days.