Roast Development Explained: DTR, Underdeveloped, and Baked

The fix: Track DTR (first-crack-to-drop ÷ total time; convention 15–25%): grassy-sour = underdeveloped, extend development with steady heat; flat-bready = baked, keep momentum and shorten the roast. One change per logged batch.

The first distinction that separates roasting beginners from roasters: roast level (how dark) and roast development (how completely the flavor chemistry finished) are different axes. A light roast can be beautifully developed; a medium roast can be underdone inside its brown shell. Most disappointing home roasts fail on development, not level — and the fix lives in one ratio and two taste signatures.

#The five stages, with the chemistry attached

1. Drying (green → yellow) — moisture leaves; nothing flavorful yet. Rushing here sets up unevenness later: the outside races ahead of a still-wet core.
2. Browning (yellow → light brown) — the Maillard reactions begin: sugars + amino acids building hundreds of flavor compounds. This is where complexity is born; the bready-toasty smell phase.
3. First crack — accumulated steam pops the bean structure audibly. A landmark, not a finish line: the bean at first crack is not yet finished coffee.
4. Development phase — everything after first crack. Caramelization deepens, acids transform, sweetness rounds. How long you spend here, relative to the whole roast, is the development decision.
5. (Second crack and beyond) — structure breaking down, oils surfacing, roast flavor overtaking origin flavor: dark roast territory, where development questions give way to "how much roast taste do I want."

#DTR: the one number roasters argue about

Development Time Ratio = time from first crack to drop ÷ total roast time. The community convention: 15–25%. A 10-minute roast with first crack at 8:00 and drop at 10:00 has a 20% DTR — squarely typical.

Treat DTR as a bookkeeping tool, not a law: it's how you make batches comparable in your log ("batch 12: 22% DTR, sweeter than batch 11 at 16%") and how the taste faults below get diagnosed. Different machines and bean densities shift the right number; your palate, recorded across batches, finds yours.

#The two development faults, by taste

Underdeveloped (dropped too soon after first crack, or rushed overall):

• Grassy, vegetal, peanut-raw notes; sharp sourness no brewing fixes
• Center of the bean visibly lighter than the surface when cracked open
• The classic first-time-roaster fault — excitement drops the batch the moment cracking ends
• Fix: extend development 30–60 seconds next batch; keep heat from collapsing after first crack

Baked (the sneakier opposite — not over-roasted, but development stalled):

• Flat, dull, breadlike; aroma strangely muted; neither bright nor rich
• Happens when temperature plateaus or climbs too slowly late in the roast — time passes but the reactions crawl. Long DTR with no momentum = baked, which is why "more development time" isn't automatically better
• Fix: maintain momentum — steady heat through development rather than coasting; shorten total roast time if batches drift past ~13–14 minutes

(And plain over-roasted — past your intended level into generic dark — is the third fault, but you can see and smell that one happening.)

#Applying it at popcorn-popper scale

You don't need probes and software to roast with development awareness:

1. Log times religiously: total time, first-crack start, drop time. That's enough to compute DTR and correlate with taste.
2. Hold heat steady through development — the era after first crack is exactly when entry roasters lose momentum (the popper's airflow cools as beans lighten). Listen: cracking should finish its arc, not trail off into silence ambiguously.
3. Change one thing per batch — drop 30 seconds later, or charge hotter, never both. Roasting is brewing's variable discipline all over again.
4. Taste at the proper rest (2–4 days) and write the verdict next to the numbers: grassy → more development; flat/baked → more momentum, less total time; ashy → you went past your level.

Development is the roasting skill: level is just when you stop, but development is what you built before stopping. Ten logged batches with crack times and verdicts will teach you more than any forum thread — and the moment a batch comes out grassy and you think "16% DTR, knew it," you've stopped making smoke and started roasting.