Use revised Pareto k threshold

[aloctavodia]

The newest revision of the PSIS paper now recommends:

• "good" ---------> k < min(1 - 1/log10(S), 0.7)
• "bad" -----------> min(1 - 1/log10(S), 0.7) <= k < 1
• "very bad" ---> k > 1

---

📚 Documentation preview 📚: https://arviz--2349.org.readthedocs.build/en/2349/

[aloctavodia]

"good_k" value is computed once and stored in the ELPDData. I am open to suggestions for a better name.

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 86.97%. Comparing base (eda7f38) to head (3d11015). Report is 15 commits behind head on main.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #2349 +/- ## ========================================== + Coverage 86.95% 86.97% +0.01% ========================================== Files 123 123 Lines 12722 12733 +11 ========================================== + Hits 11063 11074 +11 Misses 1659 1659 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.