# The Recall Trap: A Recall-Maximizing Retriever Configuration Reduces Issue Resolution in Fixed-Budget Code Context

arXiv: https://arxiv.org/abs/2608.14838 · HTML: https://mlsanity.com/paper/2608.14838/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 36 references that are in the map.
- First posted on arXiv 14 Aug 2026.

## Builds on (most cited first, 26 of 26)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering](https://mlsanity.com/paper/2007.01282/) (2020)
- [The Power of Noise: Redefining Retrieval for RAG Systems](https://mlsanity.com/paper/2401.14887/) (2024)
- [Context Length Alone Hurts LLM Performance Despite Perfect Retrieval](https://mlsanity.com/paper/2510.05381/) (2025)
- [Deeper Text Understanding for IR with Contextual Neural Language Modeling](https://mlsanity.com/paper/1905.09217/) (2019)
- [Evaluating Retrieval Quality in Retrieval-Augmented Generation](https://mlsanity.com/paper/2404.13781/) (2024)
- [Dense X Retrieval: What Retrieval Granularity Should We Use?](https://mlsanity.com/paper/2312.06648/) (2023)
- [SWE-PolyBench: A multi-language benchmark for repository level evaluation of coding agents](https://mlsanity.com/paper/2504.08703/) (2025)
- [Repoformer: Selective Retrieval for Repository-Level Code Completion](https://mlsanity.com/paper/2403.10059/) (2024)
- [PARADE: Passage Representation Aggregation for Document Reranking](https://mlsanity.com/paper/2008.09093/) (2020)
- [ContextBench: A Benchmark for Context Retrieval in Coding Agents](https://mlsanity.com/paper/2602.05892/) (2026)
- [More Documents, Same Length: Isolating the Challenge of Multiple Documents in RAG](https://mlsanity.com/paper/2503.04388/) (2025)
- [Compressing Code Context for LLM-based Issue Resolution](https://mlsanity.com/paper/2603.28119/) (2026)
- [RGFL: Reasoning Guided Fault Localization for Automated Program Repair Using Large Language Models](https://mlsanity.com/paper/2601.18044/) (2026)
- [FastContext: Training Efficient Repository Explorer for Coding Agents](https://mlsanity.com/paper/2606.14066/) (2026)
- [Byte-Exact Deduplication in Retrieval-Augmented Generation: A Three-Regime Empirical Analysis Across Public Benchmarks](https://mlsanity.com/paper/2605.09611/) (2026)
- [Practical Code RAG at Scale: Task-Aware Retrieval Design Choices under Compute Budgets](https://mlsanity.com/paper/2510.20609/) (2025)
- [On the Role of Fault Localization Context for LLM-Based Program Repair](https://mlsanity.com/paper/2604.05481/) (2026)
- [What Context Does a Coding Agent Actually Need to Act?](https://mlsanity.com/paper/2607.09691/) (2026)
- [When More Retrieval Hurts: Retrieval-Augmented Code Review Generation](https://mlsanity.com/paper/2511.05302/) (2025)
- [Scaling Retrieval Augmented Generation with RAG Fusion: Lessons from an Industry Deployment](https://mlsanity.com/paper/2603.02153/) (2026)
- [The Limits of Long-Context Reasoning in Automated Bug Fixing](https://mlsanity.com/paper/2602.16069/) (2026)
- [DeepDiscovery: A Location-Inference Framework for Task-Level Repository Understanding](https://mlsanity.com/paper/2606.22906/) (2026)
- [The Powerless Noise: How Experimental Settings Shape the Reported Power of Noise](https://mlsanity.com/paper/2607.03615/) (2026)

## Abstract (arXiv)

Retrieval components for code assistants are tuned against retrieval metrics: a configuration that raises recall@k is adopted, and downstream task success is assumed to follow. We report a controlled case study in code repair, not a new phenomenon but a deployed-flag, execution-graded instance of the known relevance-diversity and objective-mismatch tradeoff (Levy et al., 2025). On SWE-bench Verified we inject a retriever's hits as a fixed 12-slot context pack with no search tools and toggle one flag (one-chunk-per-file deduplication) on an otherwise identical stack. The flag is the higher-recall configuration (gold file present in 0.878 of served packs against 0.806 disabled), yet disabling it, trading file breadth for within-file depth, raises the single-shot resolve rate: gpt-5.6-sol +7.6pp (39.2% to 46.8%, n=500, McNemar exact p=0.0003), and a pre-registered open-weights replication any reviewer can re-run (Qwen3.6-27B, +3.6pp, n=499, p=0.0133); both survive repository-clustered inference. The gain tracks within-file anchor dose, and a random-chunk control refutes an argmax-selection artifact. We map where it holds: it reverses on a lexical BM25 retriever (-3.2pp, significant cross-paradigm interaction), is not detected under unrestricted-Read agents (a powered null), and across four languages (SWE-PolyBench, N=617) is positive but not significant (+2.6pp, p=0.056), a mapped boundary rather than a confirmed extension. Operationally, at a tight fixed budget: do not hard-deduplicate by file, and A/B packing policies against the task, not the metric the flag was tuned to.