Use tuning parameter as a factor to sigma (standard deviation parameter of Gaussian proposal). Increase tuning from 1.0 (default) if acceptance probability is too high. Decrease tuning if acceptance probability is too low (reduce step size). Optimum acceptance probability is 0.234 if using full-dimensional update. If component-wise (Gibbs) update, then it is closer to 0.44 (see Handbook of MCMC, chapter 4, page 102).
Use tuning parameter as a factor to sigma (standard deviation parameter of Gaussian proposal). Increase tuning from 1.0 (default) if acceptance probability is too high. Decrease tuning if acceptance probability is too low (reduce step size). Optimum acceptance probability is 0.234 if using full-dimensional update. If component-wise (Gibbs) update, then it is closer to 0.44 (see Handbook of MCMC, chapter 4, page 102).