Who's playing Worthy Farm, 23 to 27 June 2027? We went through 25 years of Glasto lineups, every big summer festival we could scrape, setlist.fm gig diaries and the bookies' prices, and turned it all into odds.
The chance each act gets a big slot, meaning anywhere on the Pyramid or the top five on the Other Stage. The pale part of each bar is the range we'd believe. Tap a name to see what's pushing their number up or down, and their gig diary around Glasto weekend.
The current top eight. One point per update.
Every act the bookies have priced. Along the bottom is their chance of headlining, with their margin taken off. Up the side is our chance of any big slot. Dots below the line are acts we think the bookies are too keen on. History backs that up: since 2013, autumn favourites priced above about 35% headlined only 37% of the time, and a third of the real headliners weren't priced at all.
We checked 2000 to 2025: artist-years across 40 festivals, against everyone who played the Pyramid or the top five on the Other Stage.
Festivals that point to Glasto
Count the festivals
Does a gap in the tour mean Glasto?
Who's gigging when
We tested it the strict way: each year from 2007 is predicted using only the years before it, so it never gets to peek at the future. The festival effect also holds up in both halves of the data: three or more festivals is worth about × from 2000 to 2011 and × from 2013 to 2025, after allowing for fame.
Does 20% mean 20%?
Picks that came true, by year
Under the hood it's a Bayesian model fitted to every Glasto lineup since 2000. Each act's chance depends on how much they've played the big festivals lately, their Glasto history, which festivals they're booked for that summer, and a personal "Glasto-ness" that some acts just have. Festivals are grouped by region, so ones we know little about (Open'er, Mad Cool) borrow from their neighbours instead of being guessed.
Most 2027 lineups aren't out yet, so for every festival that hasn't announced an act, the model works through the ways it could go (who's likely to be added, given how big they are) and averages over them. As announcements land, those guesses get replaced by facts. Tour dates around Glasto weekend and anything Emily Eavis says go on top. Ruled out by Eavis? Near enough zero.
Bookies' odds are treated as noisy evidence, not gospel. We collected their autumn "to headline" prices for every Glasto since 2013 and checked them against what happened: their favourites headlined 4 times in 11, and 12 of the 33 real headliners weren't in their books at all. So they go into the model as two inputs, whether an act is priced at all and the price itself, with weights learned from those 10 festivals. Tested on past years, they barely change who makes the bill but they're what lets the model spot headliners.
It also knows how famous each act is in general, from how often their Wikipedia page gets read in the year before (festival lineups alone make stadium acts like Lady Gaga look smaller than they are). And acts booked for the cancelled 2020 festival, like Taylor Swift, count that booking as Glasto history.
It isn't that clever. Predicting each year from earlier years only, its top 40 got about right a year, against for the simpler model it replaced. About 30% of Glasto's big acts didn't play any of the 40 festivals we track, so it never had a chance with them. Open'er and Mad Cool barely appear on Wikipedia, so we can't measure them; we count them like Werchter and Roskilde, since it's the same early-July run with mostly the same headliners. It's a bit of fun, not a betting system.