# Understanding Elasticsearch TF/IDF score in 6.8

**URL:** <https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009>\
**Category:** Elasticsearch\
**Created:** [March 18, 2020, 12:24am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009 "2020-03-18T00:24:56Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![AV7](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/av7/32/64706_2.png) [@AV7](https://discuss.elastic.co/u/AV7)\
**Post date:** [March 18, 2020, 12:24am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009/1 "2020-03-18T00:24:57Z")

</div>

Hi,  
I was looking at the lucene TF/IDF practical scoring function vs what is reported in the ES explain API. For ES 6.8, it used lucence 7.x for calculating the TF/IDF score which has a squared IDF term

 ![Screen Shot 2020-03-10 at 4.21.04 PM](https://us1.discourse-cdn.com/elastic/original/3X/c/9/c99d3dce983948886993709cfd38e8ff80857859.png)

However, when I look at the explain results for 6.8, I just see one idf term instead of a squared idf.

 ![Screen Shot 2020-03-17 at 5.22.30 PM](https://us1.discourse-cdn.com/elastic/original/3X/c/4/c4f8ceac93bab0f0fd488c3ef98fe23e07316694.png)

Am I missing something or is there a missing idf term in explain?

---

<div class="post-metadata">

**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\
**Post date:** [March 18, 2020, 4:14am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009/2 "2020-03-18T04:14:22Z")

</div>

By default I believe Elasticsearch 6.8 [uses BM25 instead of TF/IDF](https://www.elastic.co/blog/elasticsearch-6-1-0-released). What does your query look like?

---

<div class="post-metadata">

**Author:** ![AV7](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/av7/32/64706_2.png) [@AV7](https://discuss.elastic.co/u/AV7)\
**Post date:** [March 18, 2020, 7:04am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009/3 "2020-03-18T07:04:33Z")

</div>

I have set the similarity module to classic i.e. TF/IDF

---

<div class="post-metadata">

**Author:** ![AV7](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/av7/32/64706_2.png) [@AV7](https://discuss.elastic.co/u/AV7)\
**Post date:** [March 19, 2020, 3:59am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009/4 "2020-03-19T03:59:22Z")

</div>

The query has just a single term.

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [April 16, 2020, 3:59am UTC](https://discuss.elastic.co/t/understanding-elasticsearch-tf-idf-score-in-6-8/224009/5 "2020-04-16T03:59:23Z")

</div>

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.
