OPERA: the faster-than-light interpretation did not survive

Our assessment

Original interpretation overturned

Separate the versions

The September 22, 2011 preprint reported an arrival approximately 60.7 ns earlier than a light-speed calculation, with statistical and systematic uncertainties of 6.9 and 7.4 ns. Over the roughly 730 km baseline, the reported relative speed difference was about 2.48 × 10⁻⁵. This review grades the superluminal interpretation of that original result. The version-specific link matters: the later journal article under the same general title reports a corrected measurement, not an endorsement of the initial anomaly. [1, 2]

What went wrong

The correction concerned the time-measurement apparatus. CERN's February 2012 update identified a possible oscillator effect and a problem with the optical-fibre connection bringing an external timing signal to the detector clock; these effects acted in opposite directions. Its June update attributed the original anomaly to a faulty element of the fibre-optic timing system. This is a concrete instrumental explanation. It provides no basis for inferring dishonesty, and this review makes no such inference. A precise statistical fit cannot compensate for an incorrect clock calibration. [5]

The corrected measurements

The revised OPERA analysis replaced the original versions and reported an offset of 6.5 ns, with uncertainties large enough to include zero. The subsequent dedicated 2012 run used an upgraded setup and an additional independent timing system. For muon neutrinos it found an offset of 0.6 ns with statistical and systematic uncertainties of 0.4 and 3.0 ns. The later paper explicitly described this as confirmation, with improved accuracy, of the revised result. These findings remove support for the original early-arrival magnitude. [2, 3]

Why the cross-check mattered

ICARUS examined seven beam-associated events from a tightly bunched CERN beam and found travel times compatible with the speed of light. Seven events would ordinarily sound small, but event-by-event timing addressed the specific anomalous shift; sample size alone is not a measure of relevance. ICARUS shared the source beam while using a different detector and timing chain. Its result is therefore a valuable targeted check, with its own quoted precision, rather than a universal test of every possible neutrino-velocity model. [4]

How to read the low grade

The original claim earns no support or replication points because the measured anomaly did not survive correction and checking. Transparency earns substantial credit because the public record retains both the original result and the revised conclusion. Those are compatible judgments. The lesson is that systematic error can dominate an apparently impressive significance, and repeating a measurement through the same faulty timing component does not remove that error. The grade does not apply to OPERA's neutrino-oscillation programme, every publication by its members, or the quality of the corrective work. It is an editorial assessment of one historical interpretation against the selected primary record. An important distinction is between publishing a surprising observation for checking and treating that observation as settled new physics; the sources document the former process as well. [5]

Documented problems & their consequences

How we reached the grade

What this review does not establish

The original anomaly and corrected paper are different evidential objects. This review does not allege intent, misconduct or invalidity of OPERA's other research. Agreement within finite uncertainty is not an exact measurement of the speed of every neutrino.

The evidence over time

1. 2011-09-22

   OPERA posts the original early-arrival preprint.

2. 2012-02-23

   CERN reports potential oscillator and fibre-connection effects requiring checks.

3. 2012-03-15

   ICARUS submits its compatible-with-light-speed measurement.

4. 2012-07-12

   Corrected OPERA v4 replaces previous versions.

5. 2012-12-06

   OPERA submits dedicated 2012 results with improved timing.