Multi-Core Markov-Chain Monte Carlo (MC3)

Author:Patricio Cubillos and collaborators (see Collaborators)
Organizations:University of Central Florida (UCF), Space Research Institute (IWF)
Web Site:
Date:November 29, 2017


MC3 is a powerful Bayesian-statistics tool that offers:

  • Levenberg-Marquardt least-squares optimization.
  • Markov-chain Monte Carlo (MCMC) posterior-distribution sampling following the:
    • Metropolis-Hastings algorithm with Gaussian proposal distribution,
    • Differential-Evolution MCMC (DEMC), or
    • DEMCzs (Snooker).

The following features are available when running MC3:

  • Execution from the Shell prompt or interactively through the Python interpreter.
  • Single- or multiple-CPU parallel computing.
  • Uniform non-informative, Jeffreys non-informative, or Gaussian-informative priors.
  • Gelman-Rubin convergence test.
  • Share the same value among multiple parameters.
  • Fix the value of parameters to constant values.
  • Correlated-noise estimation with the Time-averaging or the Wavelet-based Likelihood estimation methods.


MC3 should work for both Python2.7 and Python3!


Attention!, there was a bug fix to the snooker random walk, where the posteriors were underestimating the parameter uncertainties as the number of free parameters increase. For those who have used version 2.2 with the snooker random walk, please update to the current version (v2.3).


All of these people have made a direct or indirect contribution to MCcubed, and in many instances have been fundamental in the development of this package.

Be Kind

Please cite this paper if you found MC3 useful for your research:
Cubillos et al. 2017: On the Correlated-noise Analyses Applied to Exoplanet Light Curves, AJ, 153, 3.

We welcome your feedback, but do not necessarily guarantee support. Please send feedback or inquiries to:

MC3 is open-source open-development software under the MIT License.

Thank you for using MC3!

Documentation for Previous Releases