Time Gates × Time Gates Gates · 01 of 03

Why the gate, and what may be required

Time Gates’ exercises open the household's screen time. That is the strongest engagement lever available and the fastest way to ruin the measurements underneath it, so the order in which things get built is set by one constraint: a gate may only require what the machine can verify on its own.

1.1

The finding that outranks every effect size

The research base behind Time Gates contains one result that matters more than any of its effect sizes. The best unguided app version of the best-evidenced rumination programme was null in 365 young people, and the authors' own explanation was too little engagement. In every comparable programme, adherence falls off after week four. Whatever Time Gates can plausibly do for a teenager depends entirely on them opening it on a Tuesday in week six, when the novelty has gone.

Time Gates Gates is the only lever in the house with that kind of reliability. Tying the day's first game to the day's session is the honest reason to build the two products into one household, and revision 2 of the household design already expresses it in the words the family has: the session is a task. It is titled Today's practice, before After school, and an unchecked task holds games when its zone begins, exactly as taking the bins out does. Effort unlocks; scores never do.

1.2

What the gate can break

A teenager who wants to be in a game in eight minutes will speed-run the drill. A speed-run is not a weak result - it is a poisoned one. It arrives in the data as a shortening reaction time and a rising rate of pressing when they should have held, which is indistinguishable from getting worse at the thing the exercise trains. Every number the product shows a family afterwards is then measuring the gate rather than the teenager.

That is the risk the whole sequence is arranged around. Build the things that can be verified first. Turn the gate on late, after a baseline exists that no incentive has touched. Credit effort rather than performance. And never make the gate something a teenager can fail.

The line that must not be crossed

The gate is spent by doing, never by doing well. No score, no accuracy threshold, no improvement trend and no model judgement may stand between a teenager and their evening. The moment a bad session can cost screen time, the exercise stops being a measurement and becomes a test to beat, and everything measured through it from that day onward is worthless.

1.3

The gate-eligibility test

An exercise can be built for any reason its evidence justifies. It can only become required if it passes all five tests below. This, rather than the evidence grade, is what sets the order of the sequence - several of the strongest exercises in the catalog fail it and are therefore never required, however good they are.

  1. Deterministic completion. A fixed rule decides "done", computed from trial counts, dwell time and timestamps. The steering model proposes and the rule engine disposes; the credit ledger is on the list of things the model may not touch.
  2. No hardware. No chest strap, no webcam, no floor space. A requirement that needs a sensor is a requirement that fails when the sensor is in a bag downstairs.
  3. No licensed or finite content. The research clip instruments are licensed per user and have a fixed number of items. A daily requirement burns authored content faster than anyone can write it, and burns the held-out sets that make measurement possible.
  4. Survives a bad connection. Task events queue on the device and replay, so a network fault delays a credit instead of destroying it - and while it is delayed, the gate opens rather than closes.
  5. Under ten minutes, and winnable. Past the time a teenager will spend without resentment, the gate produces an argument rather than a session.
1.4

How a session checks off a task

The credit path has four deterministic steps and the model sits outside all of them. It also crosses two devices and a rule engine, which is the part that needs care.

Two devices, deliberately. The phone carries the ritual, the breathing pacer and every sensor, because that is where the camera, the accelerometer and Bluetooth are. The computer carries the waiting and inhibition drills, because they are scored against a display refresh and a touch layer cannot give that reliably. The day starts on the phone and ends at the desk, and the proposal treats that walk as a small rep of the skill itself. The cost is that a day's credit now depends on two devices reporting, which is a materially weaker offline story than one device would give and makes failing open a real piece of engineering rather than a line in a document.

The mechanism is a task, and that is the whole of it. Gates already has a rule that puts effort before screen time: a task is due before a zone, and an unchecked task holds games when that zone starts. Time Gates’ session is one of those and nothing more - no special posture, no second mechanism, no Time Gates-shaped exception anywhere in the controller. The consequence is that the fifteen-minute warning ladder, the hold screen that offers the task rather than a wall, the shield whose primary button becomes Done, and every sentence around them already exist and already behave the way the child has learned. A task in general holds four groups - Games, Social, Video and Everything else. Education and Essentials are never held, so homework and Time Gates’ own Tasks page are never behind one. The training task is narrower than it is allowed to be: it holds Games.

Precedence, so nothing surprises anyone. A device's per-zone override applies after a parent's grant and before a task hold. A device that is set to open in a zone is still held by an unchecked task, and a parent's Add 15 still lands on top of it. Budgets are per child per zone, shared across that child's devices, and never carry over, so checking the task reopens Games at the minutes the schedule already gave it and never quietly mints extra ones. There is no snooze under a hold, because the task is the way out: the shield's own button checks it off and lifts the hold without a network call.

Where the rule engine lives is decision R8, and it is open. The credit is a record in the rule engine, one per date per child. Whether that engine keeps running on the existing box or is replaced by the local hub - that box reduced to a ciphertext mailbox, every peer compiling its own policy, a hand-written merge - is decided by R8, which the design deliberately leaves open until the trial is merged. What the sequence below assumes is only the shape of the answer, which the trial has already taught: a device that has fetched a policy has not applied it, and the two are reported separately; and an approval is not in force until the device has it, which is why the states are syncing, ready, saved for later, consumed and expired, and why only the live screen is allowed to say ready.

THE TEEN'S TWO DEVICES THE RULE ENGINE - WHERE IT RUNS IS R8 TEMPO GATES Phone ritual · state label 4 min breathing Computer if-then plan one timed drill Validity check trials · dwell · premature-release cap FIXED RULE Rule engine one day credit, per date, per child R8 IS OPEN The task is checked Today's practice, before After school PLAY REOPENS PASS EARNED RUSHED → THAT ONE DIDN'T COUNT, RUN IT AGAIN TWO INVALID TRIES → CREDITS ANYWAY, FLAGGED IN THE WEEKLY VIEW The steering model picks the drill, generates items, rates free text MAY NOT GRANT, WITHHOLD OR DELAY A CREDIT SUGGESTS ONLY Never held by a task Education · Essentials homework · Tasks
The credit path is four deterministic steps wide and the model sits outside all four. The last step is an ordinary task being checked off, which is the point: Time Gates adds no mechanism the household has not already learned. A rushed session is refused at most twice; after that it credits anyway and the refusal appears in the weekly view instead, because a teenager who cannot spend the gate is a teenager who stops opening the app.
Evidence grades throughout this sub-section are the exercise catalog's own and are never upgraded. Nothing here names a condition, and no exercise is described as fixing anything: the promise is practice at a skill, on things that look like the exercise.

Version 133 · Privacy · Terms of use · Appearance