# NoneType error on custom distribution

**URL:** <https://discourse.pymc.io/t/nonetype-error-on-custom-distribution/12176>\
**Category:** General\
**Created:** [May 21, 2023, 3:05pm UTC](https://discourse.pymc.io/t/nonetype-error-on-custom-distribution/12176 "2023-05-21T15:05:40Z")\
**Posts on this page:** 1\
**Showing post:** 12

<div class="post-metadata">

**Author:** ![matsuo\_basho](https://avatars.discourse-cdn.com/v4/letter/m/3d9bf3/32.png) [@matsuo\_basho](https://discourse.pymc.io/u/matsuo_basho)\
**Post date:** [May 23, 2023, 4:23pm UTC](https://discourse.pymc.io/t/nonetype-error-on-custom-distribution/12176/12 "2023-05-23T16:23:14Z")

</div>

Not sure the priors are so wide.

```auto
from scipy import stats

smp = stats.norm.rvs(1.7, 1.7 * 0.01, 10000)
print(smp.min(), smp.max())

1.647 1.766

```

When you say too many parameters, what do you propose? I have a trimodal distribution that I’m modeling with 2 betas and one gamma. Each distr has 3-4 parameters. Open to ideas on how to simplify.

Point taken regarding normalizing the data. Regarding few datapoints in the observed data, I thought the implication of few points in the observed data is just that posterior distribution won’t be that different from the prior predictive.

---

_[View the full topic](https://discourse.pymc.io/t/nonetype-error-on-custom-distribution/12176)._