You dropped a bagel into a bowl of water, it sank straight to the bottom, and every recipe you've read says that means it isn't ready. Maybe. The float test doesn't measure readiness. It measures density, and a cold bagel is dense whether it proofed perfectly or not. I've watched the winners of a measured batch fail this test cold, then bake taller than everything else on the rack. I've also chased a pass until I ruined a batch. So before you throw the dough out or leave it on the counter for two more hours, here's what the test actually reads, what it quietly does to your dough every time you run it, and the two checks I turn to when it fails.
What floating actually measures
A bagel floats when it holds enough gas to offset its weight in water. That's the entire mechanism. Fermentation supplies that gas, so floating tracks proofing well enough that the test earned its reputation honestly. But gas isn't the only input. Cold stiffens and compresses dough, so a ring straight out of an overnight cold proof carries a density the bowl can't tell apart from underproofing. Cold dough is dense dough. The test just answers a different question than the one you asked.
The test failed my best bagels
Earlier this month I ran a controlled proofing experiment for the sourdough troubleshooting guide: one mix, split into underproofed, overproofed, and control groups, everything tracked with calipers. Straight from the fridge, my control bagels, the ones on my normal schedule, dropped to the floor of the boil pot like stones. Seconds later they lifted off, and they finished as the winners of the whole experiment: 51 percent oven spring, the most open crumb of the twelve. The overproofed group did the same dance. The only bagel the bowl judged correctly was the deliberately underproofed one, which sat where it landed and never lifted. One right verdict out of three groups, and one of the two it got wrong was the batch that won. The whole caliper table and the cross-section photos are in Why Are My Sourdough Bagels Dense, Flat, or Gummy?
Chasing a pass is how I ruined a batch
Every guide gives a sinking bagel the same prescription: more counter time, then test again until it floats. I've followed that advice to its conclusion, on sourdough, and what it got me was an overproofed, watery mess of a batch. I kept testing and the bagels kept sinking, and somewhere in the waiting the dough blew past its window and started losing its shape. The batch never recovered. More time helps only when the dough is genuinely underproofed; a cold ring or a waterlogged one just keeps sinking while the clock does damage.
The test waterlogs your dough
That second problem is the one nobody mentions. The float test isn't free. Every dunk soaks water into the surface of the ring. Test the same bagel over and over and it starts absorbing its way out of shape, with slack skin and extra hydration the recipe never called for. Eventually it can't pass because you kept checking whether it would. Three rules keep the test from eating the batch: don't start testing until the dough already feels close, use room-temperature water, and don't keep dunking the same bagel. Rotate through the batch so no single ring gets waterlogged.
How I use it now
The float test is still my primary check at the end of bulk. What changed is what happens when it fails. I don't reach for more counter time by default anymore; I pick the bagel up and read it. Feel breaks the tie. Airy and light in the hand, and I'll ignore the sink entirely: those bagels go straight into the fridge for their cold ferment with the bowl's opinion noted and overruled. That call has a batch behind it: one that failed every float test I gave it and still came out of the oven tall. Heavy and tight like it just came off the mixer, and the dough really does need more time. The hand measures the same thing the water does, without adding any.
The boil is the better test
Once the bagels are cold, the bowl is useless, and the boil becomes the better instrument. Expect fridge-cold rings to touch bottom when they hit the water. What matters is whether they come off it. The lift only rules out one thing, but it's the thing you were worried about: underproofing. In my measured batch the winners lifted off within seconds. So did the overproofed group, but those had already confessed on the counter, slack and spreading, long before they reached the water. The underproofed one had already given its verdict in the float-test bowl, and the bake matched it: a tight, wet crumb and 19 percent oven spring against the control's 51. Sink is normal. Staying sunk is the verdict.
Where to go next
The experiment: the measured batch behind this page, with the full table, is Why Are My Sourdough Bagels Dense, Flat, or Gummy?
The failure family: if the batch baked flat and dense anyway, the ranked causes are in Why Are My Bagels Flat and Dense?
The gear: the cedar boards from that experiment are in the boards collection.
The float test won't be the last thing I put calipers to this fall. Sign up below and the results hit your inbox before they hit this blog.