A learning method built on four ideas that work — deliberate practice, mastery gating, spaced repetition, and rubric-aligned feedback. Five to ten minutes a day. Compounding. Measurable.
Anders Ericsson's deliberate-practice research is unambiguous: reps without targeted feedback engrain whatever you happened to do, including the wrong things. Every drill in DET DUDE has a single skill target and a measurable success criterion. You only advance when the structure is repeatable.
“Practice does not make perfect. Perfect practice makes perfect.” — Vince Lombardi
Public DET prep gives you 2,000 prompts on day one — and a 100-score plateau by week three. We invert it: the same prompt drilled until your tense, structure, and signposting are automatic. Only then does the full library open. Volume without structure is noise.
Five to ten minutes a day on verb tenses and conditional thinking. Feels like a warm-up, not a class. Seven days running unlocks your next-tier features for free — the only DET prep where showing up pays you back. The habit is the product.
Generic “your essay is fine” feedback is useless when you need 130. Every signal we give a student is mapped to the DET rubric: phoneme accuracy, lexical range, syntactic complexity, discourse markers. You see exactly which dimension is holding the score down.
Same skill — English — totally different scoring engine. The methods that work for the IELTS leave points on the table here.
The DET difficulty engine raises and lowers item complexity based on live performance. Static IELTS-style drilling does not train the recovery skill needed when an item suddenly jumps two CEFR bands above you.
You speak alone, to a camera, with sixty seconds. There is no rapport, no follow-up, no recovery. Generic conversational practice does not build the cold-start fluency the DET specifically scores.
DET writing items reward verb-tense variety, conditional clauses, lexical sophistication, and discourse markers — not the content of your argument. A student writing a brilliant essay in plain present tense scores lower than a mediocre essay with rich grammar.
Yes/No vocabulary, Read Aloud, Fill-in-the-Blank — these are scored on accuracy plus latency. Slow correctness loses points. Generic vocabulary lists do not train the recognition speed the DET demands.
We did not invent the underlying learning ideas. We engineered them into a DET-shaped delivery system.
The single best-supported finding in skill acquisition: structured reps with immediate, specific feedback. Our drill loop is built on this. Each rep targets one rubric dimension and ends with a measurable score.
Lexical and phoneme cards resurface on a spacing curve that accelerates forgetting-resistant memory. Your weak phonemes appear more often. Mastered ones drop to maintenance frequency. No flashcard waste.
You do not advance until you hit a 7/10 threshold on the current question type. Bloom showed mastery gating moves the median student two standard deviations above the lecture median. The same effect carries over to language acquisition.
We run your speech through a forced-alignment model that scores you phoneme-by-phoneme against the DET prosody rubric. Generic accent feedback says “unclear”; we say “your /θ/ scored 0.42 — drill this minimal pair.”
Vocabulary suggestions in the Answer Upgrade Engine are drawn from a CEFR-graded thesaurus, not academic word lists. We push you from B2 to C1 to C2 in measured steps — never asking for a word your reader does not expect at your level.
Our LLM grader is constrained to the published DET writing rubric: lexical resources, grammar range, task achievement, coherence. It cannot reward an essay for being “interesting.” Scores correlate r=0.91 with human raters.
You cannot skip ahead. Each item lights up only after its drill hits 7/10 mastery. This is the order — and the reasoning.
Pure recognition tasks. Builds the lexical surface area everything else stands on. Fast wins, daily measurable accuracy.
Adds time pressure and audio decoding. This is where the phoneme inventory starts to matter — both for recognition and production.
You produce, but the prompt scaffolds. Trains the cold-start fluency loop without the open-ended pressure of full writing.
The high-leverage items. Adaptive difficulty, full open production. By the time you reach this phase, the underlying grammar muscles are already trained — the prompt is the only new variable.
The two ungraded-but-reviewed items that universities actually read. Plus the operational layer: camera setup, proctor rules, ID handling, invalidation avoidance.
Group classes are taught by named, accredited DET instructors — every session capped at 99 students, every Q&A handled by a human. Async expert review is written by the same coaches. The AI grader does the volume; the coaches do the nuance. We will never ship a feature that pretends an AI is a person.
Take the free diagnostic. Thirty minutes. You walk away with a real score baseline and a Miyagi plan calibrated to your gap.
Start free diagnostic