AxiomaticF2LOG and AxiomaticF2EXP scores drop by 1 ulp as term frequency crosses a power of two

Finding LFV-063Area: search / similaritiesPriority: lowImpact: silent wrong resultsNovelty: new

AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.

What it shows

You score with AxiomaticF2LOG or AxiomaticF2EXP at default parameters and no boost. You expect the score never to decrease as term frequency rises, which impacts-based pruning relies on. Instead it drops by 1 ulp when the frequency crosses a power of two, so a pruned search can return a different top hit than exhaustive search.

The test shows that with default AxiomaticF2EXP() and AxiomaticF2LOG(), boost 1 and the statistics of a real index, SimScorer#score drops by 1 float ulp when freq goes from 2^29-32 to 2^29 (the next float).

Expected vs actual

Expected score(2^29) >= score(2^29 - 32) (the SimScorer#score contract that impacts-based pruning relies on); actual F2EXP 1.0134975 -> 1.0134974 and F2LOG 0.10277987 -> 0.102779865, with -ea and -da.

Files

src/org/apache/lucene/search/similarities/TestAxiomaticF2ScoreMonotonicInFreq.java (JUnit test)

Run it

Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.

# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble   (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

javac -d classes -cp "$CP" $(find src -name '*.java')

# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
     -cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.search.similarities.TestAxiomaticF2ScoreMonotonicInFreq

Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).

Formal models and scripts this finding's part of the hunt used: study-similarity-impacts-smt

Feedback

Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.

Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.