“…One could therefore argue that the reduction in (hard to control) systematic error is paid for by a corresponding increase in (expensive to reduce) statistical error. • To have a further handle on the precise behaviour of the excited-state contributions, we have also considered two-state fits [4,7], i.e. we have explicitly fit G eff X (Q 2 ,t,t s ) to leading excitedstate contributions of the form (2.1) with ∆ = m π , ∆ = 2m π , both as a function of t, t s − t at each t s separately, and as a function of t s , t at all t s simultaneously.…”