---
title: "How does a faint signal become a neutrino discovery? | Sandbox Physics"
description: "A discovery rests on a chain: raw signals, calibration, event selection and tests of other explanations. One striking event cannot replace it."
canonical_url: "https://sandboxphysics.com/stories/neutrinos/neutrino-discovery-evidence-chain/"
language: "en"
page_type: "story"
alternate_languages:
  zh-CN: "https://sandboxphysics.com/zh/stories/neutrinos/neutrino-discovery-evidence-chain/"
last_modified: "2026-10-10"
---

[2026 Nobel Prize in Physics](https://sandboxphysics.com/nobel/2026/physics/)

# How does a faint signal become a neutrino discovery?

A discovery rests on a chain: raw signals, calibration, event selection and tests of other explanations. One striking event cannot replace it.

Sandbox Physics · 2026-10-10 · Instruments and methods

## Start with what the instrument measured

A sensor measures a response, such as an electrical pulse with a time and an amplitude; it never measures “a cosmic neutrino” directly. A scientific claim starts by keeping those measurements and understanding what they can and cannot tell us.

Calibration turns instrument signals into physical quantities. A clock offset can shift an inferred direction; a sensitivity error can change an energy estimate. In an optical detector, the material the light travels through is part of the measurement too.

So the first question is concrete: would a given physical event produce the predicted signals across the whole array? Known light sources, samples of ordinary particles and laboratory measurements check this before any unusual event is given a cosmic interpretation.

Sources: [1] · [2]

## Define the sample before interpreting it

Event selection is the rule that decides which events enter an analysis. It might keep only interactions that start inside the detector, reject activity near its edge, or require a well-reconstructed track. Different rules answer different questions and let in different backgrounds.

The rules also decide how much of the signal survives. Rejecting more background can throw away interesting events. This trade-off has to be measured or simulated and checked against suitable data, because a clean-looking sample can still give a biased view of nature.

The 2013 IceCube starting-event analysis used the outer layer of the detector as a veto against particles entering from outside. Without it, atmospheric muons could imitate energetic activity. An event’s geometry and timing decided whether it entered the sample.

Sources: [3] · [1]

1. 01 **Calibrate the detector**

   Clocks, sensor response and light travel

2. 02 **Name the alternatives**

   Selection and background uncertainties

3. 03 **Test the claim**

   Search scope and robustness checks

A checklist for weighing evidence. Significance cannot be computed from a single event display.

## Test the other explanations fairly

A background is any process that can get into the selected sample without the phenomenon you are looking for. Atmospheric neutrinos are real neutrinos, but they count as background when the question is about emission from distant astrophysical sources. “Background” names a role in a particular test.

Researchers need a background estimate with uncertainties, not one reassuring number. Control samples pin down some backgrounds; simulations show how physical processes pass the selection. Comparing energies, directions and event shapes can expose an explanation that the total count hides.

Counts of independent events with a fixed expected mean follow a Poisson distribution. The equation below is a starting model; the full IceCube analysis goes further. Uncertain background rates, correlated events and changes to the selection all need extra treatment.

$$
P(N=k\mid\lambda)=\frac{e^{-\lambda}\lambda^k}{k!}
$$

The observed count is a whole number, zero or more; for independent events the expected count is held fixed. A discovery test must also handle the uncertainty in that expectation and any other information it uses.

Sources: [4] · [3]

## How many chances did the search have?

Look at many sky positions, or keep changing a time window, and ordinary background gets more chances to make one spot look exceptional. A probability computed for one fixed test does not describe the whole search.

A local significance refers to one chosen location or hypothesis. A significance corrected for the search takes into account everything that was tried. The correction depends on the actual procedure, including trials that are correlated with each other, so there is no universal factor to multiply by.

A background probability says how unusual the chosen statistic would be if a specified background model were true. It is not the probability that a proposed source is false. Judging the source also needs the model assumptions, systematic checks and other observations.

Sources: [4] · [5]

## Repeat checks, and know which ones are independent

Changing the calibration model, using another class of events or examining a control region can all challenge a result. These are valuable robustness checks. But when they reuse the same data or assumptions, they do not count as new independent experiments.

A later observing period can test whether a signal persists, and another observatory brings different instrumental errors. A fair comparison depends on the claim: a transient need not keep repeating, and an instrument with a different energy threshold may simply be unable to see the counterpart.

The teaching Labs let you practise parts of this reasoning: estimate how often pulses pair up by chance, or see how a wrong light model shifts a fitted track. Their simulated results illustrate the method. The published discovery rests on the experiment’s own data, controls and analysis.

Sources: [4] · [6] · [7]

## What we learned

Good evidence lets others inspect the alternative explanations and the uncertainties. Calibration, control samples and a clearly stated statistical test are part of the discovery itself.

## Try it in the Lab

[Reactor coincidence](https://sandboxphysics.com/labs/reactor-coincidence/) [IceCube reconstruction](https://sandboxphysics.com/labs/icecube-reconstruction/)

## Primary sources & revision

1. [IceCube Collaboration · Instrumentation and Online Systems (2017; arXiv v3)](https://arxiv.org/abs/1612.05093)
2. [IceCube · The end of the tenth polar season (5 March 2014)](https://icecube.wisc.edu/news/collaboration/2014/03/end-of-10th-icecube-polar-season/)
3. [IceCube Collaboration · Evidence for High-Energy Extraterrestrial Neutrinos (2013)](https://arxiv.org/abs/1311.5238)
4. [Particle Data Group · Statistics (2025)](https://pdg.lbl.gov/2025/reviews/rpp2025-rev-statistics.pdf)
5. [IceCube Collaboration · Evidence for neutrino emission from NGC 1068 (2022; arXiv v2, 2024)](https://arxiv.org/abs/2211.09972v2)
6. [Reines et al. · Detection of the Free Antineutrino (1960)](https://doi.org/10.1103/PhysRev.117.159)
7. [IceCube · Public data releases](https://icecube.wisc.edu/science/data-releases/)

First published 2026-10-10; last revised 2026-10-10. Original explanatory text and diagrams by Sandbox Physics. Illustrations are schematic; no experimental event records are reproduced here. This is an independent educational publication, not an official Nobel or experiment collaboration publication.

[← Previous: A huge detector, a few events: reading neutrino scales](https://sandboxphysics.com/stories/neutrinos/neutrino-energy-distance-and-counts/) [Next: Read IceCube’s discovery paper, one figure at a time →](https://sandboxphysics.com/stories/neutrinos/icecube-original-paper-figure-guide/)

## Keep reading

Neutrinos · Reading path

### Start with the prize

1. [Article **2026 Nobel Prize in Physics: why look for neutrinos in Antarctic ice?**](https://sandboxphysics.com/stories/neutrinos/nobel-physics-2026/)
2. [Article **What can a neutrino tell us that light cannot?**](https://sandboxphysics.com/stories/neutrinos/high-energy-neutrinos-astronomy-window/)
3. [Article **A huge detector, a few events: reading neutrino scales**](https://sandboxphysics.com/stories/neutrinos/neutrino-energy-distance-and-counts/)
4. [You are here **How does a faint signal become a neutrino discovery?**](https://sandboxphysics.com/stories/neutrinos/neutrino-discovery-evidence-chain/)
5. [Article **Read IceCube’s discovery paper, one figure at a time**](https://sandboxphysics.com/stories/neutrinos/icecube-original-paper-figure-guide/)
6. [Article **Five ways a neutrino headline can lead us astray**](https://sandboxphysics.com/stories/neutrinos/neutrino-nobel-common-misconceptions/)
7. [Article **Francis Halzen and the team that turned Antarctic ice into a telescope**](https://sandboxphysics.com/stories/neutrinos/halzen-and-collaboration/)
8. [Article **Why do neutrinos keep appearing in the Nobel Prize?**](https://sandboxphysics.com/stories/neutrinos/neutrinos-and-nobel-prizes/)

Neutrinos · Topic

### How IceCube found neutrinos from space

Which events and checks turned a few flashes in the ice into a discovery?

1. [Article **2026 Nobel Prize in Physics: why look for neutrinos in Antarctic ice?**](https://sandboxphysics.com/stories/neutrinos/nobel-physics-2026/)
2. [Article **What can a neutrino tell us that light cannot?**](https://sandboxphysics.com/stories/neutrinos/high-energy-neutrinos-astronomy-window/)
3. [Article **A huge detector, a few events: reading neutrino scales**](https://sandboxphysics.com/stories/neutrinos/neutrino-energy-distance-and-counts/)
4. [You are here **How does a faint signal become a neutrino discovery?**](https://sandboxphysics.com/stories/neutrinos/neutrino-discovery-evidence-chain/)
5. [Article **Read IceCube’s discovery paper, one figure at a time**](https://sandboxphysics.com/stories/neutrinos/icecube-original-paper-figure-guide/)
6. [Article **How does the atmosphere make neutrinos all around us?**](https://sandboxphysics.com/stories/neutrinos/atmospheric-neutrino-production/)
7. [Article **IceCube’s 28 events: finding high-energy neutrinos from space**](https://sandboxphysics.com/stories/neutrinos/icecube-2013-evidence/)
8. [Article **TXS 0506+056: one alert, two separate lines of evidence**](https://sandboxphysics.com/stories/neutrinos/txs-0506-056-multimessenger/)
9. [Article **NGC 1068: what does a neutrino excess establish?**](https://sandboxphysics.com/stories/neutrinos/ngc-1068-neutrino-source-evidence/)
10. [Article **One neutrino sky, three different questions**](https://sandboxphysics.com/stories/neutrinos/neutrino-sky-diffuse-galactic-point-sources/)
11. [Article **When can a few neutrino events become evidence?**](https://sandboxphysics.com/stories/neutrinos/few-neutrino-events-statistical-evidence/)
12. [Article **Open a real neutrino event file and read it carefully**](https://sandboxphysics.com/stories/neutrinos/reading-public-neutrino-event-data/)

Neutrinos · Topic

### Unexpected results and sterile neutrinos

Gallium, reactor and timing experiments have all reported surprises. Which ones held up, and could a fourth kind of neutrino explain the rest?

1. [You are here **How does a faint signal become a neutrino discovery?**](https://sandboxphysics.com/stories/neutrinos/neutrino-discovery-evidence-chain/)
2. [Article **Gallium experiments: opening the Sun’s low-energy window**](https://sandboxphysics.com/stories/neutrinos/gallium-low-energy-solar-neutrinos/)
3. [Article **Why can a neutrino change flavor?**](https://sandboxphysics.com/stories/neutrinos/neutrino-flavor-oscillations/)
4. [Article **The reactor anomaly and the spectrum bump ask different questions**](https://sandboxphysics.com/stories/neutrinos/reactor-anomaly-and-spectrum-bump/)
5. [Article **Short reactor baselines: look for a change with distance**](https://sandboxphysics.com/stories/neutrinos/short-baseline-reactor-sterile-tests/)
6. [Article **OPERA: tau appearance and a corrected clock anomaly**](https://sandboxphysics.com/stories/neutrinos/opera-tau-appearance-and-timing-correction/)
7. [Article **Sterile neutrinos: an explanation must pass several tests**](https://sandboxphysics.com/stories/neutrinos/sterile-neutrinos-anomalies-and-tests/)

1 more group

Neutrinos · Topic

### Neutrinos and the Nobel Prize

Which neutrino results won the Nobel Prize, and what evidence stood behind each one?

1. [Article **Why do neutrinos keep appearing in the Nobel Prize?**](https://sandboxphysics.com/stories/neutrinos/neutrinos-and-nobel-prizes/)
2. [You are here **How does a faint signal become a neutrino discovery?**](https://sandboxphysics.com/stories/neutrinos/neutrino-discovery-evidence-chain/)
3. [Article **Read IceCube’s discovery paper, one figure at a time**](https://sandboxphysics.com/stories/neutrinos/icecube-original-paper-figure-guide/)
4. [Article **Pauli’s letter: proposing a particle no one had seen**](https://sandboxphysics.com/stories/neutrinos/pauli-neutrino-letter/)
5. [Article **Cowan and Reines: finding the neutrino in two flashes**](https://sandboxphysics.com/stories/neutrinos/reactor-first-detection/)
6. [Article **Why did physicists need a second kind of neutrino?**](https://sandboxphysics.com/stories/neutrinos/brookhaven-muon-neutrino/)
7. [Article **Super-Kamiokande: the neutrinos that took the long way**](https://sandboxphysics.com/stories/neutrinos/super-k-zenith-angle/)
8. [Article **2015: how did two very different experiments establish flavor change?**](https://sandboxphysics.com/stories/neutrinos/nobel-2015-two-neutrino-evidence-paths/)
9. [Article **SN 1987A: a supernova seen from underground**](https://sandboxphysics.com/stories/neutrinos/sn1987a-burst/)
10. [Article **1930–2026: a timeline of neutrino evidence**](https://sandboxphysics.com/stories/neutrinos/neutrino-evidence-timeline-1930-2026/)