Emcee hammer

[ben18785]

Toy problems in notebook:

• two dimensional Gaussian: fine
• correlated Gaussian: fine
• banana: method struggles here
• annulus in 2d: slow but moving in the right direction
• multimodal: a few outliers here, which are removed in plotting
• cone: slow convergence but seems to be moving in the right direction

[ben18785]

Closing this one as it's all contained in the #11 PR.