# ORAC-NT independent real-time GW pipeline: 5/5 detection, FAR=0 on O3 data

**URL:** <https://ask.igwn.org/t/orac-nt-independent-real-time-gw-pipeline-5-5-detection-far-0-on-o3-data/1522>\
**Category:** Data Analysis\
**Tags:** news\
**Created:** [May 31, 2026, 11:27am UTC](https://ask.igwn.org/t/orac-nt-independent-real-time-gw-pipeline-5-5-detection-far-0-on-o3-data/1522 "2026-05-31T11:27:18Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Kretski](https://yyz1.discourse-cdn.com/flex031/user_avatar/ask.igwn.org/kretski/32/614_2.png) [@Kretski](https://ask.igwn.org/u/Kretski)\
**Post date:** [May 31, 2026, 11:27am UTC](https://ask.igwn.org/t/orac-nt-independent-real-time-gw-pipeline-5-5-detection-far-0-on-o3-data/1522/1 "2026-05-31T11:27:18Z")

</div>

Update on ORAC-NT pipeline — Phase F results

Following the earlier FAR discussion, I’ve completed  
validation of a morphology scoring layer on GWTC-3.

654 events processed. RF classifier (BBH vs NSBH):

- CV AUC = 0.893 ± 0.098 (5-fold)
- LOO: 17/17 NSBH correctly ranked
- Permutation p = 0.005 (N=10,000)

Score correlates with SNR (ρ=-0.74) and distance  
(ρ=+0.82) — consistent with detectability-driven  
morphology differences.

```auto
 zenodo 20475978

```

 ![orac_phase_f_publication](https://canada1.discourse-cdn.com/flex031/uploads/gwosc/original/1X/1ec6d0b15f727965f2a4de2ebdf26c5520cb09e0.jpeg)