# stochastic calculate works with greta_mcmc_list objects

    `nsim` must be set to sample <greta array>s not in MCMC samples
    the greta arrays `y` have distributions and are not in the MCMC samples, so cannot be calculated from the samples alone.
    Set `nsim` if you want to sample them conditionally on the MCMC samples

---

    `nsim` was greater than the number of posterior samples in values, so posterior samples had to be drawn with replacement

# calculate errors if the mcmc samples unrelated to target

    the target <greta array>s do not appear to be connected to those in the <greta_mcmc_list> object

# stochastic calculate works with mcmc samples & new stochastics

    `nsim` must be set to sample <greta array>s not in MCMC samples
    the target <greta array>s are related to new variables that are not in the MCMC samples, so cannot be calculated from the samples alone.
    Set `nsim` if you want to sample them conditionally on the MCMC samples

# calculate errors nicely if non-greta arrays are passed

    `calculate()` arguments must be <greta_array>s
    The following object passed to `calculate()` is not a <greta array>:
    "x"
    Perhaps you forgot to explicitly name other arguments?

---

    `calculate()` arguments must be <greta_array>s
    The following object passed to `calculate()` is not a <greta array>:
    "list(x = c(2, 1))"
    Perhaps you forgot to explicitly name other arguments?

# calculate errors nicely if values for stochastics not passed

    Please provide values for the following 1 <greta_array>:
    `a`

# calculate errors nicely if values have incorrect dimensions

    a provided value has different number of elements than the <greta_array>

# calculate errors nicely with invalid batch sizes

    `trace_batch_size` must be a single numeric value greater than or equal to 1

---

    `trace_batch_size` must be a single numeric value greater than or equal to 1

---

    `trace_batch_size` must be a single numeric value greater than or equal to 1

# calculate errors if distribution-free variables are not fixed

    the target <greta_array>s are related to variables that do not have distributions so cannot be sampled

# calculate errors if a distribution cannot be sampled from

    sampling is not yet implemented for "hypergeometric" distributions

# calculate errors nicely if nsim is invalid

    nsim must be a positive integer
    However the value provided was: 0

---

    nsim must be a positive integer
    However the value provided was: -1

---

    nsim must be a positive integer
    However the value provided was: NA