CERN's LHC May Have Found Cracks in the Standard Model — Here's What That Means

First, What Is the Standard Model?

The Standard Model is physics' catalogue of what everything is made of and how it interacts. It lists seventeen fundamental particles — the quarks that build protons and neutrons, the electron and its heavier cousins, the neutrinos, the force carriers like the photon, and the Higgs boson — and it specifies the rules by which they push, pull, and transform into each other. Assembled piece by piece through the 1960s and 70s, it has survived fifty years of increasingly vicious attempts to break it.

"Survived" undersells it. The Standard Model is the most precisely tested theory in the history of science. Some of its predictions match experiment to twelve decimal places — the equivalent of predicting the distance from New York to Los Angeles and being off by the width of a human hair.

So why does anyone want it to be wrong?

Because it is unmistakably incomplete, and every physicist knows it. The Standard Model has nothing to say about gravity. It offers no candidate for dark matter, which appears to outweigh ordinary matter roughly five to one. It doesn't explain why neutrinos have mass, which we know they do. And it can't account for why the universe contains matter at all — the Big Bang should have produced equal amounts of matter and antimatter, which would have annihilated into nothing but light.

So the situation is peculiar. Physicists have a theory that is spectacularly accurate at everything it covers, and they are certain something bigger lies underneath it, and for decades they have been unable to find the seam. Every experiment keeps returning the same infuriating verdict: the Standard Model is right again.

Which is why a result that doesn't match gets attention.

The Decay With the Ridiculous Name

The measurement, published in Physical Review Letters by the LHCb collaboration, concerns a process called an electroweak penguin decay. Yes, penguin.

Why on earth is it called a penguin?

This is one of physics' better stories, and it's true.

What actually happens in a penguin decay

Strip away the name and the process is genuinely strange. It starts with a B meson, a short-lived particle containing a beauty quark (also called a bottom quark). The beauty quark wants to become a strange quark — a lighter, more stable relative.

In the Standard Model, it is not allowed to do this directly. There's no permitted one-step route from beauty to strange. The transformation can only happen through a convoluted detour: the quark briefly emits and reabsorbs other particles in a quantum loop, and comes out the other side transformed. The detour is so improbable that only about one B meson in a million decays this way.

In this particular measurement, the B meson breaks apart into four things at once: a kaon, a pion, and two muons. The physicists then measure the angles at which those four particles fly out. Those angles are the fingerprint. The Standard Model predicts a very specific pattern of them, and what LHCb measured isn't quite that pattern.

Why hunt in the rarest corners?

This seems backwards at first. If you're looking for undiscovered particles, why study something that almost never happens instead of building a bigger collider to smash them into existence directly?

Because of what lives inside that quantum loop. During the detour, the transformation borrows particles from the vacuum — briefly, virtually, without them ever becoming real. And critically, it can borrow particles too heavy to ever be produced directly. An undiscovered particle far beyond the LHC's reach can still leave a fingerprint here, by subtly altering the odds and the angles of a decay we can measure.

The analogy physicists like: you can't see what's in the next room, but if you know exactly what the room should sound like when empty, an unexplained echo tells you something is in there. The penguin decay is quiet enough that a faint echo would actually be audible. In a common, noisy process, any new physics would be drowned out completely.

What LHCb Actually Measured

The scale of this is worth sitting with. To collect enough of these decays, the team combed through data from roughly 650 billion B meson decays recorded between 2011 and 2018. One in a million of those is the process they wanted. Then they measured the angular pattern of what came out.

The result disagrees with the Standard Model prediction at a significance of about 4 sigma.

That number is the whole story, so it's worth understanding properly.

What Does "4 Sigma" Actually Mean?

Sigma is a measure of how surprised you should be. Specifically: assuming the current theory is correct and nothing new is going on, how unlikely is it that random noise alone would produce a discrepancy this large?

Every measurement wobbles. Count anything finite — decays, coin flips, votes — and you get a slightly different answer each time, purely by chance. Sigma tells you how far outside that normal wobble your result sits.

So 4 sigma means: if the Standard Model is completely correct and this is just bad luck in the data, that's a roughly 1-in-16,000 coincidence. Unlikely. Not impossible.

Worth noting how steep that ladder gets — each step up isn't twice as strict, it's dozens of times stricter. Going from 4 sigma to 5 sigma means the odds of a fluke drop by more than a factor of a hundred. If you want to see how sigma maps onto probability for any value, or work out what significance your own data reaches, you can run the conversion in our statistical significance calculator.

So why isn't 1-in-16,000 good enough?

Here's the part that makes intuitive sense once you see it: because physics runs an enormous number of measurements.

A 1-in-16,000 coincidence is very unlikely to happen to you, once. But it is not at all unlikely to happen somewhere, when hundreds of teams are testing hundreds of quantities across dozens of experiments, year after year. Rare things become routine at scale. That's why the bar is absurdly high — not because physicists are pessimists, but because they've done the arithmetic on how many chances they're giving randomness.

Drag the slider below to see the effect.

Run it a few times at a realistic scale and the lesson lands: 3-sigma results appear and vanish constantly. They are a normal feature of doing a lot of science, not a signal that something is wrong with the universe. Four sigma is genuinely uncommon. Five sigma, at any plausible scale of scientific output, essentially does not happen by accident — which is exactly why that's where the line was drawn.

The graveyard is full of exciting anomalies

Physicists are cautious here because they have been burned, repeatedly, and recently.

In 2015 both major LHC detectors saw a bump in their data at 750 GeV that looked like it might be a new particle. It generated hundreds of theory papers within months. More data arrived, and the bump dissolved into nothing — it had been a statistical fluctuation all along.

Closer to home: LHCb itself spent years chasing a different anomaly, in which B mesons appeared to decay into electrons and muons at unequal rates, violating a Standard Model principle called lepton universality. It sat around 3 sigma and was widely discussed as a likely crack in the theory. In 2022, with better data and an improved analysis, LHCb reported that the effect had gone away. The Standard Model was fine.

That's not a failure of science; it's science working exactly as intended. But it's the reason the people who found this result are the ones being most careful about how it's described.

The Boring Explanation That Might Still Be Right

Here's the honest complication, and it deserves more than a footnote.

Comparing measurement to theory requires knowing what the theory predicts — precisely. And for this particular decay, the Standard Model prediction is hard to calculate. There's a contribution from something physicists call "charming penguins": additional quantum processes involving charm quarks that feed into the same decay and shift the expected angles.

These contributions are entirely conventional — no new physics required. The trouble is that they involve the strong nuclear force in a regime where the usual calculation methods work poorly, so their size is genuinely uncertain. If charming penguins turn out to be larger than currently estimated, the "anomaly" is not an anomaly at all. It's the Standard Model, calculated imprecisely, and the discrepancy is with our arithmetic rather than with nature.

There is precedent for exactly this. The long-running "muon g-2" anomaly, another famous crack in the Standard Model, narrowed substantially not because the measurement changed but because theorists improved their calculation of what the Standard Model actually predicts. Sometimes the theory side moves.

Does Anything Else Back It Up?

Yes, and this is the strongest point in the result's favour. CMS — a completely separate detector, at a different point around the LHC ring, built and run by a different collaboration with different equipment and different analysis software — measured the same decay and found results consistent with LHCb's.

CMS is not optimized for this kind of measurement, so its numbers are less precise and can't confirm anything on their own. But independent agreement matters enormously, because the most common cause of a false anomaly isn't bad luck — it's a subtle flaw in one experiment's calibration or analysis. Two independent teams making the same mistake is far less likely than one team making it.

So the anomaly survives the first serious cross-check. It just hasn't survived enough of them yet.

What Happens Next

Two things, one soon and one slow.

More data already exists. The published result uses data collected between 2011 and 2018. Since then, LHCb has recorded roughly three times as many B mesons, sitting in storage and waiting to be analysed. That alone will sharpen the measurement considerably. If the discrepancy is real, it should grow toward 5 sigma as the data accumulates. If it's a fluctuation, it should fade — the way the 750 GeV bump did.

And a much bigger dataset is coming. Planned upgrades to the LHC and its detectors through the 2030s are expected to deliver a dataset around fifteen times larger than today's. That is the scale at which a real 4-sigma effect becomes an unambiguous discovery, and at which a statistical ghost has nowhere left to hide.

In parallel, theorists will keep working on the charming-penguin calculation. It's possible this gets resolved on the theory side before the data ever settles it — which would be an anticlimactic ending, and a perfectly legitimate one.

So How Excited Should You Be?

Moderately, and honestly.

Nobody has rewritten physics. The Standard Model has not been overturned, no new particle has been found, and the textbooks are not changing. What exists is a measurement that doesn't match prediction, at a level of statistical strength that would be considered decisive in most fields and merely promising in this one, backed by a second experiment, with a plausible mundane explanation that currently looks insufficient but hasn't been ruled out.

That is a real lead. It is also the sort of lead that has evaporated many times before.

What makes it worth following is the setup. Physicists have spent fifty years certain the Standard Model is incomplete and unable to find where it fails. Every anomaly is a candidate for the crack that finally opens the thing up. Most of them close. This one, for now, is still open — and unlike most, there's a clear path to settling it: three times the data already collected, fifteen times more coming, and a specific calculation that theorists are actively working to pin down.

Check back in a few years. That's not a hedge; it's the actual timeline.

Work out the significance yourself

Convert any sigma level into a p-value and into odds you can actually picture, one-tailed or two-tailed, with the 3σ and 5σ thresholds marked. Free, no sign-up, runs in your browser.