Sourdough pizza dough: calculating fermentation time by temperature

The short answer

The fermentation time of a sourdough pizza dough is computed from three variables. The inoculation rate (the weight of starter as a percentage of total flour): the higher it is, the faster, but not proportionally. The temperature: the rate follows a bell curve, not a straight line, slow below 15 °C, lively between 24 and 30 °C. The state of the starter: a starter at its peak counts as 1, a young starter about 0.8, a fallen starter about 0.7, a starter straight from the fridge without feeding about 0.4.

Hours, bulk and final proof together, for a starter at its peak at 100 % hydration, computed by the Fornello 0.24 engine (inoculation law at confidence C). The time is when the dough has received the rise a pizza maker expects; your dough confirms it by its volume and texture, not the clock.

A starter's speed is not a straight line

Between 20 and 30 °C you could believe that "warmer, faster" is a linear rule. It is wrong at both ends. Fornello describes the rate with the cardinal temperature model (Rosso et al., 1993): a minimum temperature below which nothing moves, an optimum, a maximum. The curve is normalised so that 20 °C is 1, and the rest follows: 4 °C is 0.105, 12 °C is 0.46, 24 °C is 1.30, 28 °C is 1.59, and the maximum sits around 35 °C. Beyond that it falls fast.

The engine applies the same curve to a starter as to baker's yeast. It is an acknowledged simplification, noted at confidence C: the typical sourdough yeast, Kazachstania humilis (long called Candida milleri), grows fastest around 27 °C and stops around 36 °C, while baker's yeast has its optimum around 32 °C (Gänzle et al., 1998; Salvadó et al., 2011). The bake journal exists to correct the gap with your own culture. In the range where pizza is made, 18 to 28 °C, the shape of the curve is what matters, and it is right.

Doubling the starter does not double the speed

It is the most widespread error in sourdough recipes. A culture inoculated at 10 % does not have half the power of one at 20 %: it simply has one more population doubling to do before reaching the same point. Fornello models that with a power law: the leavening power is equivalent to a dose of instant yeast proportional to the inoculation raised to the power 0.6. Concretely, 20 % starter is worth about 0.24 % instant yeast, 10 % is worth 0.16, 5 % is worth 0.10. Going from 20 to 10 % does not add 100 % to the time, but about 50 %, as the table above shows.

That equivalence is also what lets the rest of the engine (temperature, thermal lag, maturation, transposition from one oven to another) run unchanged on a sourdough. One source of truth, not a second model alongside.

The fridge: what it really does to a sourdough

At 4 °C the relative rate is 0.105. Twenty-four hours in the fridge therefore bring about 2.5 hour-equivalents of rise, no more. That is little, and that is the point: the cold hardly makes the dough rise, it makes it mature. Enzymes and lactic acid bacteria keep working at a pace yeast does not follow, flavour builds, acidity climbs gently, and the dough softens.

But a ball put in the fridge is not at 4 °C right away. Mixed at 23 °C, a 250 g ball takes about two hours to come down, and it ferments during that time at its real temperature. The engine integrates that descent: 48 hours in the fridge for a 250 g ball starting at 23 °C give 6.8 h-eq, against 5.0 if it were treated as cold from the first minute. A third more, which explains why so many "48-hour" sourdoughs come out past it.

The starter's state changes everything, and nobody notes it

A starter "at peak" has doubled or tripled in volume, its dome is rounded, it smells of yoghurt and fruit. A young starter (taken before the peak) has less active population: Fornello gives it a factor of 0.8. A fallen starter (the dome has sunk, the smell leans toward alcohol) is at 0.7, with more acidity. A starter straight from the fridge without feeding is at 0.4, and above all its strength is unpredictable: that is the case the calculator flags with a warning rather than a figure.

The feeding can be computed too. For a 1:5:5 ratio (one part starter, five flour, five water), the seed is one eleventh of the mix, so 3.5 doublings to return to peak: 8.7 hours at 24 °C, 11.3 hours at 20 °C, 7.1 hours at 28 °C. A 1:2:2 ratio (one fifth) peaks in 6.5 hours at 24 °C. The calculator gives the feeding in grams for the amount of starter your dough needs, with a margin to keep the next seed.

Acidity: temperature and hydration are your two levers

Two doughs at the same pH can taste entirely different depending on which acid dominates. Cold and a stiff starter push the bacteria toward acetic acid: sharp, vinegary. Warmth and a liquid starter push toward lactic acid: mild, yoghurt-like. Fornello estimates the pH from the maturation load and the acetic share from the average temperature and the starter's hydration, then names the expected character: sharp, balanced or mild. If you want a sourdough crust that does not bite, run your starter warm and liquid, feed it often, and do not exceed 24 hours cold.

The calculator below opens on a contemporary Neapolitan, in sourdough mode. Turn the inoculation dial: the engine tells you whether your dose carries your schedule, suggests the one that would, and shows the pH, the character and the feeding. Change the hours and temperatures of the schedule, and watch the suggested inoculation follow.

Three sourdough schedules that work

Loads computed by the engine for a 250 g ball mixed at 23 °C, thermal lag included. The inoculation shown is the one the calculator suggests for each schedule, rounded.