One of the signs that you should not trust a method or its evangelists is that they actively block you just for disagreeing with them (even in the most polite ways).

I will never trust Bayesian methods or Bayesians simply for the reason that they simply don’t have stomach for a good debate or a counter argument. The moment they feel they are losing the argument they block you.

I know Dr. Juan Orduz might be held in great regard in Bayesian circles, but it is not a sign of a good academic or statistician to block someone just for posing a counter argument.

I had written the below for which I was blocked and my comment deleted. You decide if it was inappropriate or not. As always I believe the audience should have arguments from both sides to make informed decision. Deleting comments robs the audience of knowledge (no matter if you disagree or you will it is wrong in some ways).

My comment below:

“In sample R Squared is not a good metric for decision making”

In sample R squared value is not a great metric but one always has Adjusted R Squared value. Both R squared value and Adjusted R squared value depict goodness of fit of a model. Also Adjusted R squared value varies only by few percentage points from R Squared value most of the time.

Any business decision should not be solely basis this. Good business decision should also come from external sources (something not available in the model) – domain knowledge.

I would also argue that Posterior Predictive Check (PPC) is also not a great metric along the same lines as R Squared Value is not. PPC is also an in-sample metric. How is it that PPC is better than R Squared Value?

“Good causal models are not necessarily those with the highest R Squared value”

High R Squared value is not a sufficient condition but it is one of the conditions. If we have fit a good causal model, then all the variables ‘explain’ the changes in the Dependent variable (KPI) to a greater degree. By virtue of this, the R Squared value will be high.

Good Causal models ideally have high R Squared values. But not all models with high R squared value are causal.

Regarding DAG: There must be an arrow from seasonality to Inquiries as well. Inquiries must be ideally placed at center before the sales. Generally, Inquiries cause sales (rarely the other way). The above depiction opens a lot of backdoor paths.

–End of comments”