---
title: "Comment by Kartik N V J K on \"Should you trust an object detector’s confidence?\" | Hashnode"
description: "The 61.81% average confidence against 78.45% actually correct is a clean case of underconfidence, and temperature scaling fixing the scores without moving a sin"
url: "https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3/comment/6ac7ddf30a309763fa28d23e"
---

[Discussion](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3)

Comment [Should you trust an object detector’s confidence?](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3)

[Kartik N V J K](https://hashnode.com/@kartik-nvj)

[Oct 8](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3/comment/6ac7ddf30a309763fa28d23e)

The 61.81% average confidence against 78.45% actually correct is a clean case of underconfidence, and temperature scaling fixing the scores without moving a single box is the neat part. It is the lesson I keep relearning on evals: a healthy-looking number can still be miscalibrated, the way a green CI run hides a live regression (I wrote that up here: [kartiknvjk.hashnode.dev/my-ci-evals-were-green-a-…](https://kartiknvjk.hashnode.dev/my-ci-evals-were-green-a-regression-still-paged-me-at-3-am) ). Did 15 equal-width ECE bins ever mask something that adaptive-width bins caught?

[Oleg Serov](https://hashnode.com/@sirserow)

[Oct 18](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3/comment/6ac7ddf30a309763fa28d23e#reply-6ad422c08bdbbc8157d272f1)

Thanks, Kartik! Apologies for a long response. The CI analogy makes sense - one summary number can hide quite a lot.

In my experiment Platt scaling produced the large improvement, reducing ECE from 16.64 to 1.31 percentage points, however the temperature scaling only brought it down to 15.28. Both methods kept the boxes and rankings unchanged.

I haven’t tested adaptive-width bins in this experiment, so I can’t say they caught something the fixed bins missed. I did check 10, 15, and 20 equal-width bins: YOLO26n’s original ECE stayed at 16.64 points across all three. That confirms underconfidence finding, but still doesn’t really establish how sensitive the corrected result is to binning.

Thank you, comparing equal-count bins would be a useful follow-up, especially in the high-confidence region.
