# Question about Optimized Scalar Quantization

**URL:** <https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969>\
**Category:** Elasticsearch\
**Created:** [July 9, 2025, 9:32pm UTC](https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969 "2025-07-09T21:32:10Z")\
**Posts on this page:** 3\
**Page:** 2

<div class="post-metadata">

**Author:** ![john-wagster](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/john-wagster/32/139021_2.png) [@john-wagster](https://discuss.elastic.co/u/john-wagster)\
**Post date:** [July 24, 2025, 2:41pm UTC](https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969/21 "2025-07-24T14:41:17Z")

</div>

> [@yli](#):
>
> the correction terms more used during dot product calculation

They are used after the dot product calculation for distance to bring the outcome more in line with an approximation of the raw distance calculation.

> geometry example, to me seems more about describing the geometry properties of the vectors

agreed; I think it's meant to help ground the reader to understand why the math works to make this more efficient.

---

<div class="post-metadata">

**Author:** ![yli](https://avatars.discourse-cdn.com/v4/letter/y/bbe5ce/32.png) [@yli](https://discuss.elastic.co/u/yli)\
**Post date:** [July 25, 2025, 5:33am UTC](https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969/22 "2025-07-25T05:33:28Z")

</div>

@john-wagster as always, thanks for your quick reply.

> They are used after the dot product calculation for distance to bring the outcome more in line with an approximation of the raw distance calculation.

yes, that is why I thought that the correction is not directly related to the geometry properties of the vectors, and the geometry properties description seems more related to $\bar{x}$.

---

<div class="post-metadata">

**Author:** ![john-wagster](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/john-wagster/32/139021_2.png) [@john-wagster](https://discuss.elastic.co/u/john-wagster)\
**Post date:** [July 25, 2025, 1:45pm UTC](https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969/23 "2025-07-25T13:45:58Z")

</div>

> yes, that is why I thought that the correction is not directly related to the geometry properties of the vectors, and the geometry properties description seems more related to $\bar{x}$.

yes you got it!

[Previous page](https://discuss.elastic.co/t/question-about-optimized-scalar-quantization/379969.md?page=1)