GitLab

It appears that there are several mistakes in the case the boxcox calculation is used in .ErrorCrit.

First, an epsilon is added to Qobs and Qsim although it is not necessary for BoxCox:

  if ("epsilon" %in% names(InputsCrit) & !is.null(InputsCrit$epsilon)) {
    VarObs <- VarObs + InputsCrit$epsilon
    VarSim <- VarSim + InputsCrit$epsilon
  }

should become

  if ("epsilon" %in% names(InputsCrit) & !is.null(InputsCrit$epsilon) & !(InputsCrit$transfo == "boxcox")) {
    VarObs <- VarObs + InputsCrit$epsilon
    VarSim <- VarSim + InputsCrit$epsilon
  }

Then, the formula is wrong:

    VarSim <- (VarSim^0.25 - 0.01 * mean(VarSim, na.rm = TRUE)) / 0.25
    VarObs <- (VarObs^0.25 - 0.01 * mean(VarObs, na.rm = TRUE)) / 0.25

should be replaced with

    VarSim <- (VarSim^0.25 – (0.01 * mean(VarObs, na.rm = TRUE))^0.25 ) / 0.25
    VarObs <- (VarObs^0.25 – (0.01 * mean(VarObs, na.rm = TRUE))^0.25 ) / 0.25