Two people are arguing about coffee right now, and they’re both wrong. One swears dark roast hits harder because it’s bolder, bitterer, more “coffee-tasting.” Obviously that means more caffeine, right? The other read a headline two years ago and now corrects people at parties: actually, light roast has more caffeine, it’s science. Both of them are citing a real fact. The number that actually decides what’s in their cup is sitting somewhere else entirely.
How you measure your coffee, and how you brew it, control your caffeine far more than the roast level printed on the bag.
📋 Table of Contents
The Fight: Two Wrong Answers, One Bag of Coffee
The “dark roast is stronger” camp is confusing two different things: intensity of flavor and caffeine content. Dark roast tastes bolder because roasting develops bitter compounds and a heavier body. That’s flavor chemistry, a completely different mechanism from caffeine content, and people conflate the two constantly.
The “light roast has more caffeine” camp got closer, and picked it up from real reporting on a real phenomenon, but flattened a genuinely interesting mechanism into a slogan. The actual reason is duller and more useful than either side wants: it’s mostly about how you’re scooping your coffee, and it only gets interesting once you correct for that.
The Scoop That Lies: Volume vs. Weight
Here’s the part almost nobody accounts for. As beans roast longer and darker, they lose more water weight and expand. They get less dense and slightly bigger. Fill a tablespoon with dark roast and a tablespoon with light roast and you’ve matched volume. Mass is a separate question, and mass is what actually determines how much caffeine ends up in your cup.
America’s Test Kitchen ran this exact comparison and found it took two-and-a-half more tablespoons of dark roast than light roast to reach the same 1.5 ounces by weight. Scoop-for-scoop, light roast beans pack in more mass, and more mass means more of everything in that scoop, caffeine included. Measured by the kitchen scoop most people actually use, “light roast has more caffeine” checks out. Weigh the beans instead, and the gap mostly closes.
Scoop your coffee your whole life without ever putting it on a scale, and you’ve never actually known how much caffeine you were drinking. This is also why a recipe that calls for “two tablespoons” falls apart the moment you switch roasts. The same measurement failure wrecks your brew ratio on every method we’ve written a protocol for. A ten-dollar scale fixes both problems at once.
The Study That Actually Settled It
Weight-for-weight, does roast level still matter at all? Researchers at Berry College and Drexel University finally ran the controlled version of this argument, publishing their results in Scientific Reports in late 2024. They roasted the same Ethiopian coffee, both washed and natural process, to seven precise degrees, brewed each one on an AeroPress at a consistent 15:1 water-to-coffee ratio, and ran every sample through high-performance liquid chromatography to get a real lab number for caffeine, the kind taste alone can’t produce.
Caffeine concentration peaked in the light-to-medium range, ahead of every roast darker or lighter than it, in beans that had lost roughly 14-16% of their green weight during roasting and landed at 134-165mg per 8oz cup depending on process. Roast much lighter or much darker than that window and caffeine dropped on both ends. Extraction yield told a similar story, peaking around 12-14% roasting mass loss and declining measurably by the time beans reached the darkest levels tested.
The title goes to the roast level sitting between the two corners, and almost nobody’s rooting for it.
The Mechanism: Porosity Fighting Volatilization
The reason a middle roast wins isn’t mysterious once you separate the two forces actually at war inside the bean. Roasting increases porosity (the same study measured it climbing from 13.6% in green coffee to 45.8% in the darkest roast tested), and higher porosity means water finds its way into the bean faster and pulls more soluble compounds out during brewing. On that axis alone, darker should always win.
But caffeine is also a molecule that leaves the bean on its own once things get hot enough. The researchers identified sublimation losses becoming significant once bean temperatures pushed past roughly 400-420°F, a threshold that dark roasts blow straight through on their way to second crack. So darker roasts extract more efficiently from what’s left, but there’s measurably less caffeine left to extract by the time you get there. Light-to-medium roasts sit in the window where porosity has already opened up enough to extract well, before volatilization has burned off a meaningful share of the caffeine. That’s the whole fight, decided by two curves crossing at almost the same point.
This is the same logic behind why water chemistry moves the needle on extraction independent of roast level: extraction yield is doing a lot of the real work here, and roast color is just riding along on top of it.
The Bigger Lever Nobody Mentions
While both camps were arguing over a roast-level difference worth single-digit percentage points, they skipped past the one variable that actually moves caffeine by a factor of two: the plant. Arabica beans run roughly 1.2% caffeine by dry weight; Robusta runs roughly double that, in the 2.0-2.2% range. That gap dwarfs anything roasting does to a bag of the same beans.
Every Fight Club blend is built on Arabica, so treat variety, brew ratio, and your scale as the real caffeine controls, and leave roast level out of that decision entirely. Buying a darker roast of the same Arabica bean shifts caffeine by a few percentage points at most. A finer grind, a longer steep, or a higher dose extracts proportionally more of everything dissolvable in the bean, and that swing runs into the double digits, far past anything roast level alone can do.
Brew Method Still Outranks Roast Level
Even inside a fixed roast and a fixed bean, brew method decides more of your final caffeine number than roast degree does. A longer contact time, a finer grind, and a higher brew ratio all pull more total solubles, caffeine included, out of the same beans. That’s why the researchers standardized on one method (AeroPress, 15:1, ten minutes) before they trusted any roast comparison at all: change the brew and you’ve introduced a bigger variable than the one you’re trying to measure.
Put a light-to-medium roast like Colombian Choke Hold through a full immersion brew with a long steep and you’ll out-caffeinate a rushed, coarse-ground pour-over of a darker bag like Big Joe Blend by a wide margin, even though Big Joe is roasted darker. If you’re chasing a specific caffeine target, dial in ratio and contact time first, and treat roast level as the last adjustment worth making.
Curious about the density gap from the scoop problem above? Weigh out a level tablespoon of Ethiopian Jab against the same tablespoon of Big Joe Blend on a kitchen scale. The numbers won’t match, and now you know exactly why.
Home roasters get an extra lever here too: hitting a repeatable, specific mass loss is a level of control eyeballing color never gives you, and a production-grade home roaster makes tracking that number practical. The 14-16% window the study identified as the caffeine sweet spot is worth writing down if you’re profiling your own batches at home.
The Numbers That Matter
| Variable | What Moves It | Approximate Effect |
|---|---|---|
| Measuring by volume vs. weight | Dark roast beans expand and lose density during roasting | Up to ~60% caffeine difference per scoop, measured by volume alone |
| Roast degree, by weight | Porosity (helps extraction) vs. volatilization above ~400-420°F (loses caffeine) | Peaks at 14-16% roasting mass loss; drops toward both extremes |
| Species: Arabica vs. Robusta | Genetic caffeine content, unrelated to roasting | Robusta runs roughly double Arabica’s ~1.2% dry-weight caffeine |
| Brew ratio and contact time | How much total solubles the water pulls from the same grounds | Larger swing than roast level alone, even within one bean |
The Closing Bell
Both sides of this argument had grabbed a wrong number and stopped looking. Dark roast earns its bite from flavor chemistry, and light roast earns its reputation from a scoop that happens to hold more mass. The scoop lied to one camp, and roast-level tunnel vision blinded the other to variety and brew method, the two variables actually worth arguing about.
Next bag you open, put it on a scale before you touch a scoop, and if caffeine is genuinely what you’re chasing, spend your attention on ratio and contact time before you ever reach for a darker roast. That’s the fight that actually decides what’s in your cup.
Stop arguing about roast color. Start arguing about your scale.
