---
title: "Discussion on \"Hybrid Search and Reranking: Fixing Retrieval After Vector Search Falls Short\" | Hashnode"
description: "Discussion on \"Hybrid Search and Reranking: Fixing Retrieval After Vector Search Falls Short\". Hybrid Search and Reranking"
url: "https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Hybrid Search and Reranking: Fixing Retrieval After Vector Search Falls Short](https://nagspidey.hashnode.dev/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short?utm_source=hashnode&utm_medium=feed)

Hybrid Search and Reranking

Nov 3 · 8 min read · nagspidey.hashnode.dev  
Read the article

-   [#rag](https://hashnode.com/tag/rag)
-   [#search](https://hashnode.com/tag/search)
-   [#machine-learning](https://hashnode.com/tag/machine-learning)
-   [#artificial-intelligence](https://hashnode.com/tag/artificial-intelligence)
-   [#software-engineering](https://hashnode.com/tag/software-engineering)

## Written by

[Nagappan S](https://hashnode.com/@imkarthik03)

@imkarthik03 · Chennai, India

  
[View profile](https://hashnode.com/@imkarthik03)

Building scalable products at the intersection of Full Stack Engineering & AI

I build scalable web apps, experiment with AI, and turn complex engineering problems into working products.

## Responses 5

Markdown supported · ⌘↵ to post

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Nov 6](https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9/comment/6aed94c5b2cd1143f171a2cd)

The two-researcher analogy assumes both ranked lists identify the same kind of item. In practice one search leg may return document IDs while the other returns passage IDs, and RRF cannot recognize agreement between them unless the fusion key is defined consistently.

I would record the ranking unit at each stage and test a document returned by the keyword leg whose relevant passage is returned by the dense leg. Decide explicitly whether fusion happens at passage or document level, then preserve the mapping for the reranker. That makes a high rank in both lists a well-defined claim instead of an artifact of incompatible identifiers.

[Nagappan S](https://hashnode.com/@imkarthik03)

[Nov 9](https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9/comment/6aed94c5b2cd1143f171a2cd#reply-6af1a5b7a83903c56a0414a7)

The post doesn't cover the fusion key, so thanks for raising it. I'd pick one unit for both legs, passage level, and carry the parent document ID as metadata so the reranker and citations can still use it. Your test case is the right one: a document the keyword leg returns whose relevant passage only the dense leg returns. Without a consistent key, a high rank in both lists isn't a well-defined claim.

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[1d ago](https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9/comment/6aed94c5b2cd1143f171a2cd#reply-6afb78694a85fc13205058d7)

Passage-level fusion with parent document IDs gives the reranker and citation layer a consistent item to work with. The remaining choice is how the keyword leg turns a document hit into passage candidates.

I would record that expansion explicitly and retain the matched text location, then test a document whose heading matches the question while its answer lies in a later passage. That shows which passage acquired the keyword evidence and whether the correct passage survives before reranking begins.

[AI Developer](https://hashnode.com/@lycore)

[Nov 4](https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9/comment/6aeaa0e2ead8145b951529fe)

The point that a reranker can't recover a document cut at position 51 is the one I'd test directly: log the pre-rerank rank of whatever the reranker ends up putting first, and if a meaningful share come from deep in the candidate list, widen the window before changing the model. It's also worth checking recall@50 of the fused list on its own, separate from final ranking quality, since that number caps everything after it. For identifier-style queries like calculate\_tax\_v2, BM25's default tokenizer may split on underscores or digits, so a code-aware analyzer (or an extra n-gram field) can matter more than the fusion weights. Since RRF ignores scores, have you considered weighting the keyword list more heavily for queries that look like identifiers or error codes, and measuring whether that beats plain k=60?

[Nagappan S](https://hashnode.com/@imkarthik03)

[Nov 9](https://hashnode.com/posts/hybrid-search-and-reranking-fixing-retrieval-after-vector-search-falls-short/6ae9d62e3b872791f3339ce9/comment/6aeaa0e2ead8145b951529fe#reply-6af1a5cb4b4391b4b0718317)

Logging the pre-rerank rank of whatever the reranker puts first is a cheap way to tell whether the window is the bottleneck, and recall@50 of the fused list is the number that caps everything after it. The tokenizer point on identifiers like calculate\_tax\_v2 isn't in the post and matters. On weighting the keyword list for identifier-style queries instead of plain k=60: I haven't tested it. The post only says to keep k=60 unless you have labelled relevance data, so I'd try it against a labelled query set before trusting it.
