The full prediction

Everyone we track, how likely they look, and what would change it. Close scores are grouped, not ranked; a tenth of a point between names is noise. Every act carries the facts behind its placing: when they last headlined, and the size of rooms they play now.

Loading the latest run…

The full board, and where our two views disagree

Every act the betting market prices, and every act the model rates as a possible first-time headliner. Both figures for each, and where each figure actually comes from. The rest of the field is in the banded list below, where the gaps are too small to rank.

Unpriced
No two independent bookmakers price this act, so no market figure is used. The figure is the model's alone.
Priced and scored
Both views contribute. The prices behind the market half, each with its date, are listed further down this page.
Priced, not scored
The market's view of acts our model doesn't score. The figure is the market's alone. That is a decision about what the model covers, not a judgement that the act is unlikely.
Ruled out, figure withheld
The festival's organiser has publicly ruled this act out of this edition, with a date and a citation. Bookmakers may still be pricing them; we do not publish that price as a probability while the statement stands.
Excluded by rule
A standing rule of this edition's model scores this act at zero. These are editorial rules, the modeller's judgment rather than something new data forced; each states its basis and the public condition that retires it, on the act's own row and on the How It Works page.

Loading the latest run…

Returning headliners

Acts in the same band are equally plausible; inside a band the order is alphabetical because the gaps mean nothing.

Loading the latest run…

Where the betting market fits

Two figures per act: evidence alone, and evidence with the betting market folded in. How the two lines work

The market figure only moves when enough independent bookmakers price the same names. A handful of prices, or several outlets echoing one source, is a rumour with a number attached; when that is all there is, the figures stay evidence-only and this page says so.

How good the market has actually been

Checking historical odds boards against what actually got announced, the clear market favourite went on to headline about 6 times in 10. But when we can see everything a bookmaker priced that day, the eventual headliner is missing from it more often than not, in 6 of the 8 cases where we hold a complete board.

Those two readings do not deserve the same confidence. The first only counts markets a journalist wrote about at the time, which tends to be the ones the market got right. The second counts every named act on a board we know was complete, whether the market got it right or not, and it is the more trustworthy of the two.

About the betting odds in these figures

These figures combine bookmakers' own standing prices, read from their boards on the day shown, with older quotes carried by named publications on the dates they reported them. Every quote names its bookmaker and its source, and each one either carries its reported date or says plainly that it has none.

Prices we would no longer call fresh are shown with their age rather than dropped; removing them would change the figures without telling you why. The full source list, with each price's date, is on the How It Works page.

One limit on all of this. The historical comparison behind it only really tells us how reliable bookmaker odds are 3 to 6 months before a headliner is announced. Most of the prices behind today's figures were reported further out than that, so we can't yet say how well this pattern holds at that distance.

The prices behind the market figure

The acts the market prices, and how far the market and the evidence disagree. Every price behind these figures, with its date, is on the How It Works page.

Loading the latest run…

Notable absences

The biggest names missing from the board, and the specific reason for each.

Some acts carry a “with the market” percentage and no model figure. That means the market is pricing a name the evidence cannot support yet: no completed Pyramid headline set (a cancelled booking does not count), or no current activity to score. Those percentages are the betting market's alone.

Bad BunnyEvaluated, currently 0%: below the first-time-headliner threshold

His two Tottenham Hotspur Stadium shows (June 2026) are fully counted. The model rates his scale as unquestionable. What he lacks, on current data, is the UK festival-circuit footprint the model requires before rating a first-time Pyramid headline slot as bookable: the same threshold applied to every first-timer, on the same evidence. UK festival appearances or a deeper UK touring pattern would change this. BLACKPINK, Karol G and BTS sit in the same category for the same reason.

Justin BieberOn the watch-list, 0%: no announced upcoming tour dates

He has returned to performing since cancelling the Justice tour in 2022: 12 shows in the past two years, most recently in July 2026, including SoFi Stadium and MetLife Stadium. What the model has no record of is a single announced future date: nothing is currently on sale anywhere, so there is no forward touring signal to score. An announced UK/EU tour, or a new release paired with announced touring, moves him onto the scored board automatically.

RosalíaNot on the board: no Glastonbury pathway on current evidence

Her LUX world tour is real and large: 40 shows across 12 countries in the past two years, including two O2 Arena nights in London (May 2026), all counted in the model's data. She has no Glastonbury appearance history at any level, and arena-scale touring alone doesn't clear the model's evidence bar for a first-time Pyramid headline slot. A UK festival-circuit presence or UK stadium-scale headline shows would change her position.