# Anti-pattern mining

> The subtle idea in learning from strong boards: not every agreement between them is good. Some shared placements are real structure; some are a shared trap that caps every search just short of the top. Separate the two, and attack the trap.

- Canonical page (with interactive figures/demos): https://eternity2.dev/research/build/learning/anti-pattern-mining/
- Updated: 2026-07-16
- Topics: local-search, learning
- Source: Ropke & Pisinger 2006, adaptive large-neighbourhood search (the destroy-and-repair frame this steers) — https://doi.org/10.1287/trsc.1050.0135

---
Every technique so far treats agreement between strong boards as a good thing:
where the good boards concur, follow them. This one takes agreement apart, because
some of it is a trap. When many independent searches all reach a high but
imperfect board, they agree on a great many placements, and some of that agreement
is genuine structure while some of it is a shared bad habit: a local choice that
looks good, feels forced, and quietly caps every search just short of the top. The
whole idea is to tell the two apart, then attack the trap.

## Two frequencies, not one

A single count cannot separate real structure from a trap, because both show up
almost everywhere. The trick is to compute *two* frequencies for each candidate
pattern and compare them. Over a corpus of boards scoring across a wide band, for
every adjacent cell-pair and every piece-pair that ever sits there, measure:

$$
p_\text{high} = \frac{\#\{\text{boards with the pair, score} \ge 460\}}{\#\{\text{boards} \ge 460\}}, \qquad
p_\text{all} = \frac{\#\{\text{boards with the pair}\}}{\#\{\text{all boards}\}}
$$

together with a *ceiling*: the best score of any board that contains the pattern.
Two categories fall out. **Good consensus** is high $p_\text{all}$ with a high
ceiling, a pattern that strong boards share *and* the very best boards keep: real,
trustworthy structure. A **consensus trap** is high $p_\text{all}$ with a ceiling
stuck just below the top, a pattern almost every board adopts but no top board
does: the agreed-upon wrong choice that locks a whole family below the record. The
single-frequency view cannot distinguish them; splitting by basin is the entire
move.

> **[Figure]** Interactive: the trap map, corner-family by corner-family — interactive: PalimpsestDiagram. Rendered on the canonical page (link above); not shown in this markdown export.

## Steering, not banning

Knowing where the traps are turns the corpus into a map of which placements to
trust and which to pull apart. The [PALIMPSEST](/research/lab/experiments/raphael-anjou/learning/palimpsest)
experiment uses it to steer an [ALNS](/research/build/local-search/local-search-alns)
search: take a board caught in the trap basin, find the wedge whose removal breaks
the most trap patterns while sparing the good-consensus ones, deliberately break
those cells, and hand the board back to the repair loop, whose destroy operator
tears open exactly those broken regions and rebuilds them. Steered this way, it
reached 463, the best board this project has produced (the
[PALIMPSEST experiment](/research/lab/experiments/raphael-anjou/learning/palimpsest),
whose committed board reproduces via `just research-record-boards`).

The caveat is the lesson. Trying to use the trap list *directly*, forcing the
search to avoid every trapped placement, did not work and made boards worse. The
value was in reading the corpus to choose *where to aim* the search, not in
hard-coding its conclusions as a ban. A learned map is a good place to point a
search and a bad set of shackles to bolt onto it, which is the same boundary
between tiebreak and objective that runs through
[corpus priors](/research/build/learning/corpus-priors), seen from the other side:
here the danger is not over-weighting a good signal but over-trusting a signal
about what is *bad*.

Even steered perfectly, this does not clear the top. Rebuilding the trapped regions
tends to land back on the same known top board rather than a genuinely new one,
because the trap is not a mistake a smarter search avoids; it is the
[rigidity wall](/research/why/rigidity-wall) itself. That is the subject of the
[collapse page](/research/build/learning/when-learning-collapses).

## Related

- [Learned move-ordering](https://eternity2.dev/research/build/learning/learned-value-ordering) — When several pieces fit, which do you place next? Instead of one rule of thumb, carry several signals learned from strong boards and let them vote. Voting keeps the search from over-trusting any one hunch and marching into the same dead end.
- [When learning collapses](https://eternity2.dev/research/build/learning/when-learning-collapses) — The sobering half of learning from strong boards. A learned signal reaches the top of a search's own range reliably, and then stops. Over-trust it and the search collapses; even used perfectly it does not raise the ceiling, because the ceiling is not a thing the corpus knows.
- [PALIMPSEST](https://eternity2.dev/research/lab/experiments/raphael-anjou/learning/palimpsest) — Read every strong board to find the habits that quietly hold a board back, then break them. This experiment produced the project's best board: 463 of 480.
- [Local search and ALNS](https://eternity2.dev/research/build/local-search/local-search-alns) — Destroy part of a board, rebuild it better, and let the algorithm learn which demolitions pay. Adaptive large-neighborhood search is the most reliable polisher this project has, and the cleanest demonstration of the wall where polishing ends.
- [The rigidity wall](https://eternity2.dev/research/why/rigidity-wall) — Every record board we have is frozen in place. You cannot nudge your way from a great board to a perfect one, and we can prove it.