I've tried to frame it in terms of standard Propagation of Uncertainty rules. This ignores any correlated errors between inputs, but it makes it easier to under

I've tried to frame it in terms of standard Propagation of Uncertainty rules. This ignores any correlated errors between inputs, but it makes it easier to understand. Things clicked fairly fast for me when I tried this. It's just a weighted average between a prediction (which you can do easily using the error-propagation rules) and a new measurement.