Time & Sourness

So, time equals sournesslink

I have come to understand that the sourness to a large extent depends on the fermentation time; the longer you let it ferment, the more sour it will become.

Baking sourdough bread is in this respect truly a two-stage rocket, both the duration it takes for the starter to peak, as well as how long you let your dough proof, will have an impact on your bread's resulting sourness & character.

You can get your starter ready in 4 hours, or you can let it simmer for 12.

You can proof your bread for 7.5 hours in a warm and cozy environment, or you can let it proof for 48 hours in the fridge.

If you want only a mild sourness in your bread, combine a rapid starter fermentation with a short proofing time.
If you want a more sour bread, let your starter ferment for longer, and give it a long proof in the fridge.

And experiment with different combinations to find what works best for you.

Hydration & Sourness

Hydration level equals sournesslink

Not all "sour" is the same, or so they say.

The theory is that high-hydration starters restrict oxygen. This forces the bacteria to produce creamy, yogurt-like lactic acid rather than sharp, vinegary acetic acid, giving your bread a sweeter, milder profile.

One caveat though: a stiff starter also tends to favour yeast over bacteria, so the overall tang of the bread can still end up milder than the acid type alone would suggest. Hydration nudges the flavour of the sourness more than its amount.

Temperature & Sourness

Temperature equals sourness toolink

And supposedly, temperature is the bigger lever of the two.

Lactic acid bacteria thrive in warmth, roughly 27–32°C, while acetic acid producers prefer cooler conditions, roughly 10–18°C. So a warm ferment will lean creamy and mild, and a cold ferment (like a long fridge proof) will lean sharper and more vinegary.

Put together with the hydration story above, you get a rough two-by-two:

Warm + wet tends toward the mildest, most yogurt-like sour. Cold + stiff tends toward the punchiest, most vinegar-like sour. The other two corners sit somewhere in between.

The Math

Putting numbers on itlink

A starter feed is the same fermentation problem the rise model already solves, so the timing below is built on it rather than on a one-off formula — the water temperature is then its own calculation.

The Starter (Inoculation) Weight Formulalink

The rise model relates fermentation time to inoculation and temperature as $t = k \cdot S^{a} \cdot e^{\,b \cdot T}$, where $S$ is the starter percentage and $T$ the temperature. To feed for a target time $n$ at your room temperature $z$, we solve that same relation for the inoculation:

$\%S = \left(\frac{n}{k \cdot e^{\,b z}}\right)^{1/a}$

We keep that result inside a practical band of 5–40% seed. When the band can't reach your target time on its own, desired time spills over into the temperature instead — which is what sets the water temperature below.

Once we know the required starter percentage relative to flour, we find the exact Flour Weight ($F$) needed to yield your total output target ($y$):

$F = \frac{y}{1 + \frac{x}{100} + \frac{\%S}{100}}$

The Desired Water Temperature ($T_{water}$) Formulalink

This one borrows the Desired Dough Temperature idea from baking tradition, adapted for a starter feed. The desired time decides a target temperature $T_{\text{target}}$ for the freshly-stirred mix, and we pick the water temperature that lands it there. The flour and the seed both sit at room temperature $z$; there's no mixer, so friction is left out — a slack batter stirred by hand adds no meaningful, repeatable heat. Weighting each component by its mass $m$ and solving for the water gives:

$T_{\text{water}} = z + \left(T_{\text{target}} - z\right) \cdot \frac{m_{\text{total}}}{m_{\text{water}}}$

The result is clamped to a safe 4–40°C: warm water gives a freshly fed starter a head start, but past about 40°C it does more harm than good — which is where the biology comes in.

Why the temperature is cappedlink

A sourdough culture is two organisms working together: the bacterium Lactobacillus sanfranciscensis (the acid and flavour) and the yeast Candida milleri (the lift). They don't share the same temperature tolerances. Gänzle, Ehmann & Hammes measured them directly¹: the bacterium peaks around 32–33°C and stops growing at 41°C, while the yeast peaks near 27°C and is already dead-stopped by 36°C.

So heating a feed toward 40°C is counter-productive — past roughly 32°C you're shutting down the very yeast that makes the starter rise, long before the bacteria are in trouble. That's why the target temperature is capped at 32°C and the water at 40°C. It's also a guard rail for the model itself: the $e^{\,b \cdot T}$term keeps predicting “faster” as temperature climbs and never captures the fall-off above the optimum, so we don't let it run past where the biology turns over.

Feeding

Feeding your starter for a target timelink

Plug in the four inputs and the calculator returns the exact weights of active starter, water and flour to use, plus the water temperature and the target temperature it's aiming for.