---
title: "Algorithms for Scoring Coreference Chains — Bagga 1998"
description: "The paper proposes the B-cubed (B³) metric for evaluating coreference resolution. Instead of counting links between mentions, as the earlier MUC score…"
authors: ["Amit Bagga", "Breck Baldwin"]
year: 1998
venue: "Proceedings of the LREC 1998 Workshop on Linguistic Coreference, Granada, 563–566"
source: "http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.47.5848"
licence: "No open copy found (LREC 1998 workshop proceedings, ELRA) — card only, no text"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/bagga-1998-algorithms-scoring-coreference-chains/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/bagga-1998-algorithms-scoring-coreference-chains/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [1998](https://md.research.elamur.ai/free/en/papers/1998/) / Algorithms for Scoring Coreference Chains

# Algorithms for Scoring Coreference Chains

**Authors:** Amit Bagga, [Breck Baldwin](https://md.research.elamur.ai/free/en/authors/breck-baldwin/) · **Year:** 1998 · **Venue:** Proceedings of the LREC 1998 Workshop on Linguistic Coreference **Link:** unknown **License of the paper:** unknown

[**0** citations](https://openalex.org/works?filter=cites:W7100926688) OpenAlex · 2026-10-14 · [**325** citations in our library](https://md.research.elamur.ai/free/en/papers/bagga-1998-algorithms-scoring-coreference-chains/cited-by/) of the 140673 papers whose bibliographies we read · Proceedings of the LREC 1998 Workshop on Linguistic Coreference, Granada, 563–566

- [Publisher's page](http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.47.5848)
- [OpenAlex](https://openalex.org/W7100926688)

**Authors (2)**

1. Amit Bagga
2. [Breck Baldwin](https://openalex.org/A5023998507) [ORCID](https://orcid.org/0000-0002-9974-4434) · h 10 · 1173 citations · 46 works

Citations and author metrics: OpenAlex, 2026-10-14; authors with the same name are not always told apart.

## Summary

The paper proposes the B-cubed (B³) metric for evaluating coreference resolution. Instead of counting links between mentions, as the earlier MUC score does, B³ computes precision and recall for every individual mention: how much of the predicted cluster containing that mention overlaps with its gold cluster, and vice versa. The per-mention scores are then averaged, optionally with weights. This mention-centric view fixes known blind spots of link-based scoring, for example that merging everything into one large cluster is not punished enough and that singleton entities are ignored. B³ became one of the three components of the standard CoNLL coreference score, together with MUC and CEAF.

## How Mova uses it

- `coref/src/score.rs` — the coreference scorer implements B³ alongside MUC, CEAF-e and CEAF-m; the CoNLL F1 reported for Mova's coreference resolver is the mean of MUC, B³ and CEAF-e.
- Formulas for predicted (non-gold) mentions follow the later reference treatment by Pradhan et al. 2014; mentions are matched by syntactic head, as in the CRAC shared tasks, with and without singletons.
- The scorer is self-checking: before every evaluation it reproduces the worked example from Pradhan et al. 2014 (B³ 35/84 and 0.50) and the "gold against itself = 100" check, failing fast otherwise; negative controls (one giant cluster, everything separate) must score low.

## Effectiveness in Mova

This is an evaluation metric, not a modelling idea, so it has no gain of its own. It is the measuring instrument for all coreference results, e.g. the sieve resolver on GUM test with gold trees: CoNLL F1 73.04 with singletons (MUC / B³ / CEAF-e = 69.18 / 75.26 / 74.68), 60.31 without singletons (B³ 58.79), per the development notes of `coref`.

---

👨‍🔬💥

Topics: [Coreference resolution](https://md.research.elamur.ai/free/en/topics/coreference-resolution/)

## Cite

The paper itself, not this card. BibTeX from the library's record of the work.

```bibtex
@inproceedings{bagga1998algorithms,
  title = {{Algorithms for Scoring Coreference Chains}},
  author = {Bagga, Amit and Baldwin, Breck},
  booktitle = {Proceedings of the LREC 1998 Workshop on Linguistic Coreference, Granada, 563–566},
  year = {1998},
  url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.47.5848}
}
```

- [← previous A surface-based approach to identifying discourse markers and elementary textual units in unrestricted texts](https://md.research.elamur.ai/free/en/papers/marcu-1998-surface-based-approach-identifying-discourse/)
- [next → An IR Approach for Translating New Words from Nonparallel, Comparable Texts](https://md.research.elamur.ai/free/en/papers/fung-1998-ir-approach-translating-new-words/)
- [↑ back to papers of 1998](https://md.research.elamur.ai/free/en/papers/1998/)