add a parameter such as self.thin in the userParameters file that is used to thin the Markov Chain results so that only every nth model is included in output calculations & statistics. For example, with if the chain is run for 400,000 iterations with thin=4, the output statistics would be based on every 4th model (100,000 total). Accept/reject rules apply to all models in the chain as the algorithm runs.
This might help to deal with some 'noisy' outputs where the chain gets stuck occasionally, and will be a step towards future multi-chain runs
add a parameter such as self.thin in the userParameters file that is used to thin the Markov Chain results so that only every nth model is included in output calculations & statistics. For example, with if the chain is run for 400,000 iterations with thin=4, the output statistics would be based on every 4th model (100,000 total). Accept/reject rules apply to all models in the chain as the algorithm runs.
This might help to deal with some 'noisy' outputs where the chain gets stuck occasionally, and will be a step towards future multi-chain runs