Currently @observable puts observables into the GrandCanonical ensemble.
However, some observables can only be computed bootstrap-by-bootstrap. In particular, observables with disconnected pieces, such as correlations of density fluctuations,
< n_a n_b > – < n_a > < n_b >
The simple < averages > are called primary observables in Monte Carlo errors with less errors, while (general, nonlinear) functions of primary observables are called derived quantities.
It would be good to have a way to add derived quantities to bootstrap averages in the same way that we add observables to the GrandCanonical ensemble. Perhaps a @derived or @secondary?
Currently
@observable
puts observables into the GrandCanonical ensemble.However, some observables can only be computed bootstrap-by-bootstrap. In particular, observables with disconnected pieces, such as correlations of density fluctuations,
The simple
< averages >
are calledprimary observables
in Monte Carlo errors with less errors, while (general, nonlinear) functions of primary observables are calledderived quantities
.It would be good to have a way to add derived quantities to bootstrap averages in the same way that we add observables to the GrandCanonical ensemble. Perhaps a
@derived
or@secondary
?