# Bug in fast sample posterior predictive?

**URL:** https://discourse.pymc.io/t/bug-in-fast-sample-posterior-predictive/6904
**Category:** Questions
**Created:** [March 4, 2021, 10:40pm UTC](https://discourse.pymc.io/t/bug-in-fast-sample-posterior-predictive/6904 "2021-03-04T22:40:35Z")
**Posts on this page:** 1
**Showing post:** 8

<div class="post-metadata">

### Author: ![bridgeland](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/bridgeland/32/2773_2.png) [@bridgeland](https://discourse.pymc.io/u/bridgeland)
#### Post date: [March 10, 2021, 3:45pm UTC](https://discourse.pymc.io/t/bug-in-fast-sample-posterior-predictive/6904/8 "2021-03-10T15:45:41Z")

</div>

That’s a nice ontology of PPC usages, Ricardo. You are correct. I want to do #5.

In [another post on a similar issue](https://discourse.pymc.io/t/sample-posterior-predictive-with-a-vengeance/5926), @OriolAbril suggested the technique of modifying values in an **arviz.InferenceData** trace, and then running **pm.sample\_posterior\_predictive()** to generate new distributions for unobserved variables.

That suggestion was made on a somewhat different problem, as I wanted to model intervention on RVs. In this case, I am trying to model intervention on deterministics.

So there are three possible interpretation of the behavior of the simple model in the original post, above:

1. It is a bug. **pm.fast\_sample\_posterior()** should be able to work on modified samples of a deterministic.

2. It’s not a bug. **pm.fast\_sample\_posterior()** can handle modified samples of RVs, but not deterministics.

3. It’s off-label usage of **pm.fast\_sample\_posterior()**, and should not be relied on. It happens to work for RVs today, but that is more or less an accident. I should find a different technique to model interventions.

I think you are claiming #3.

---

_[View the full topic](https://discourse.pymc.io/t/bug-in-fast-sample-posterior-predictive/6904)._