# More Like This Searches - an analogy?

**URL:** <https://discuss.elastic.co/t/more-like-this-searches-an-analogy/303929>\
**Category:** Elasticsearch\
**Created:** [May 4, 2022, 10:29am UTC](https://discuss.elastic.co/t/more-like-this-searches-an-analogy/303929 "2022-05-04T10:29:46Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![sandydesu](https://avatars.discourse-cdn.com/v4/letter/s/dbc845/32.png) [@sandydesu](https://discuss.elastic.co/u/sandydesu)\
**Post date:** [May 4, 2022, 10:29am UTC](https://discuss.elastic.co/t/more-like-this-searches-an-analogy/303929/1 "2022-05-04T10:29:46Z")

</div>

I'm using 'More Like This' (MLT) searches for a problem I have and I have found them to be really well suited for the types of searches I need. I need to explain how MLT searches work in a meeting and I'm after an analogy to help illustrate the idea, but this analogy is eluding me at the moment.

I've studied the [documentation](https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-mlt-query.html) on MLT searches, used the **\_explain** API to see how things are calculated and I've been through the inner workings of the [BM25 algorithm](https://en.wikipedia.org/wiki/Okapi_BM25).

My audience is technical, but they aren't specialists in data or search. I need something to illustrate the idea of what is going on with this type of search without having to go into all of the details. I am wondering if anyone here has any ideas? I would be grateful for any help provided!

---

<div class="post-metadata">

**Author:** ![Mark\_Harwood](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mark_harwood/32/10538_2.png) [@Mark\_Harwood](https://discuss.elastic.co/u/Mark_Harwood)\
**Post date:** [May 4, 2022, 12:33pm UTC](https://discuss.elastic.co/t/more-like-this-searches-an-analogy/303929/2 "2022-05-04T12:33:14Z")

</div>

Hi. It uses only the interesting words in a piece of text as a search.  
Technically you could use _all_ the words in a long piece of text in a search and achieve similar quality results but it would be slower. So instead we cherry-pick the rarer words because they are most highly-valued in searches anyway, being more informative. If a rare word is repeated in a text that is also a good indication that the text is about that topic.

---

<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:** [June 1, 2022, 12:33pm UTC](https://discuss.elastic.co/t/more-like-this-searches-an-analogy/303929/3 "2022-06-01T12:33:54Z")

</div>

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