Skip to main content

Research · May 2026 · 20 min read

Why ADHD tools create loops — not momentum

We analyzed real ADHD behavior — Reddit threads, clinical research, first-person accounts — and found that most ADHD tools, including an earlier version of Resistaa, are built on a false assumption. Here's what the data showed and what we changed.

1. What actually happens when an ADHD brain tries to start

The problem is almost never task complexity. People with ADHD usually know what they need to do. What fails is the transition from knowing into acting — the moment where the task becomes real enough to begin.

Motivation for many ADHD brains is often easier to access through interest, novelty, challenge, urgency, or passion. "I want to go to the gym" contains none of these. It's importance-based motivation — which is often unreliable for ADHD task initiation, regardless of how much the person cares about the outcome.

Real user language

"I know what I need to do. I just can't start. It's like the engine won't turn on."

"My brain is screaming JUST START while my body refuses to move."

"Between telling myself 'You've got time' and 'Holy sh*t, where do I begin?' — all the unstarted tasks consume all my energy."

"I want to be the type of person who's consistent — but I never am."

The four real failure loops

The pre-launch drift: Person intends to go. Checks phone "just for a minute." Forty-five minutes pass. Window has closed. Tells themselves "tomorrow." They never entered a loop — they drifted out of range before initiation occurred.
The fake-start loop: Puts on shoes. Fills water bottle. Drives to the gym. Sits in the parking lot. Drives home. Documented verbatim: "Another time, I didn't even leave the parking lot; I spent the entire hour right there." Each preparatory step feels like progress, but none of them cross the commitment threshold.
The avoidance spiral: The task has been deferred multiple times. Each deferral adds emotional weight. The gym now represents every time they failed to go. Thinking about it triggers shame fast enough to suppress any initiation signal before it fires.
The abandoned workout: Person gets inside. Starts. Picks up phone during a rest. Loses track of time. Boredom spikes. "I just get so distracted and then it takes me so long." This is a different failure — initiation wasn't the problem. Sustained engagement was.

Real progress vs. fake progress

Fake progress is any action that feels like movement but doesn't change the distance to the actual goal or create irreversibility: organizing the desk, researching the optimal workout plan, putting on shoes and sitting back down, driving to the gym and sitting in the parking lot.

Real progress is any action that makes not continuing cost more than continuing: leaving the house, walking through the gym door, telling someone you're on your way. The distinguishing feature is not the size of the action — it's whether it changes the cost-benefit structure of stopping.

2. Why the standard model fails — including ours

Most ADHD tools — and an earlier version of Resistaa — operate on a single assumption: the user is stuck because the task is too big. Make it smaller, and they'll start.

This is wrong in five specific ways:

Micro-actions solve for the wrong problem: They address task complexity, not activation. The user already knows the steps. "Put on your shoes" doesn't fire the neurochemical initiation signal — it gives the person something small to do while the signal continues to not fire.
Completing micro-actions provides fake relief: Each completed step can briefly relieve urgency. The person feels like they're making progress, but they haven't crossed the commitment threshold. Relief decays, they return to the stuck state — now with shoes on.
Escalation rules assume linear momentum: If the system generates step 3 because step 2 was completed, but the user's internal state hasn't changed, step 3 fails for exactly the same reason step 2 would have failed if skipped. The user can complete an action robotically and loop back to the couch.
Physical actions miss the emotional layer: ADHD task initiation failure is, in most cases, emotionally mediated. Shame, perfectionism, identity threat — "if I can't do this right, I'm the kind of person who doesn't follow through" — are the real blockers. No physical micro-action addresses this.
The model can't distinguish five different stuck states: "I can't go to the gym" can mean: never left the house, drove there but can't enter, has been meaning to go for weeks, or is genuinely depleted. Each state requires a different intervention. Generating the same category of response regardless of state produces repetition, not momentum.

3. The new model: five states, not one

The 7 friction patterns Resistaa already identifies (task paralysis, rejection sensitivity, time blindness, etc.) describe why it hurts. They're still correct and still matter.

What was missing: a second layer describing what is happening right now. These are not why-dimensions — they're operational states that determine which intervention type will actually work.

Pre-initiation freeze

Feels like

"I want to, I know I should, but I literally cannot make my body move. It's like there's a weight. The engine won't turn on."

Actually doing

Scrolling. Making tea. Reorganizing. Opening and closing tabs. Not consciously procrastinating — the brain is running avoidance routines without full awareness.

What works

Activation — something that changes internal state before changing task proximity. Body doubling, environmental change, urgency injection, commitment device.

Fake-start loop

Feels like

"I keep doing things but somehow I'm not doing the thing. I've put on my shoes three times."

Actually doing

Completing preparatory actions that don't require full commitment. Clothes on, bag packed, in the car, in the parking lot — nothing crosses the commitment threshold.

What works

Commitment — something irreversible that changes the cost-benefit structure of stopping. Texting someone. Paying for the class. Leaving the phone in the car.

Avoidance spiral

Feels like

"I keep meaning to but it never happens. The more I don't do it, the worse it feels to think about it. I'm the kind of person who doesn't follow through."

Actually doing

Active avoidance. The task now carries the weight of every prior failure. Thinking about it triggers shame fast enough to suppress the initiation signal before it fires.

What works

State shift — don't address the original task directly. Reframe to remove accumulated shame loading. "Not going to the gym. Just driving somewhere."

Partial engagement

Feels like

"I started but I'm not really there. I'm going through the motions."

Actually doing

Physically present but attention has drifted. Phone out during rest periods. Time-blindness has eroded the session. Half-doing without intention.

What works

Re-engagement — novelty injection, challenge frame, or exit permission. "Do the next 3 sets as fast as you can. Then you can leave if you want."

Burnout stall

Feels like

"I just can't. Not today. Everything feels heavy. I'm not even pretending anymore."

Actually doing

Genuinely depleted. Not performing avoidance — actually spent. Often follows a hyperfocus crash or accumulated emotional load from repeated failures. Looks like a freeze but the resource pool is empty.

What works

Recovery — not the original task. Honor the stall. Micro-success in a completely different domain. Do not escalate.

4. The intervention system: six categories

Instead of "generate a micro-action," the system selects an intervention category based on the detected state, then generates the right action for that category. Not all outputs are small physical steps — the right intervention depends entirely on which state the user is in.

Activationfor: Pre-initiation freeze

Change internal state before changing task proximity. Leave the room. Body double. Set a timer with a consequence. Don't plan anything.

"Open a separate note and write the one sentence you are avoiding. Do not send or submit it."

Commitmentfor: Fake-start loop

Create irreversibility. Make not-continuing cost more than continuing. Remove the retreat path.

"Text someone: 'I'm walking in now.' Send it before thinking."

State shiftfor: Avoidance spiral

Don't address the original task. Remove shame loading by reframing or replacing the goal entirely.

"Stop trying to go to the gym. Walk in any direction for 4 minutes."

Pattern breakfor: Any detected loop

Switch the strategy, not just the wording. If a loop keeps repeating, the next intervention should change the format, stakes, or environment instead of offering another version of the same advice.

"Stop drafting in the email app. Make a three-line private outline in notes: point, risk, next step."

Re-engagementfor: Partial engagement

Novelty injection, challenge frame, or explicit exit permission. Initiation succeeded — the problem is sustaining attention.

"Do the next 3 sets as fast as you can. Then you can leave."

Recoveryfor: Burnout stall

Do not retry the original task. Build a micro-success elsewhere. The resource pool is empty — pushing harder depletes what little remains and deepens the shame load.

"Today is a rest day. Pick one 2-minute task in a different domain."

Loop detection rule

If the user returns without marking progress, uses emotional language (shame, "I always," "I never"), or has been trying for weeks — the system is forbidden from repeating the same intervention category. It must switch. This single rule breaks most repetition loops.

5. Concrete comparison: "I want to go to the gym but I can't start"

Old system

1.Put on your workout clothes.
2.Get your water bottle ready.
3.Pack your gym bag.
4.Set a specific time to leave.

The user completes all four and is still on the couch — now with a bag packed and a plan. Each step looked productive without crossing the commitment threshold. The system has no way to detect this. It generates step 5 anyway: another preparatory action into the same loop.

New system

State detected: pre-initiation freeze → Activation

"Don't plan anything. Just get into your car and sit there with the engine on. You don't have to drive yet."

Environment shifts without full commitment. Partial irreversibility. No decision load about what to do at the gym.

If user returns still stuck → Commitment

"Text someone right now: 'I'm on my way to the gym.' Send it before you read the next sentence."

Irreversible. Social accountability activated. Creates mild urgency without requiring willpower.

If loop detected → State shift

"Stop trying to go to the gym. Walk outside in any direction for exactly 4 minutes. Don't bring your gym bag."

Removes the gym — and its accumulated shame — from the frame entirely. If they drive there afterward, unforced. If not, the failure pile didn't grow.

The principle everything else follows

The tool is not a step generator. It is a state-change engine. The size of the action is irrelevant if the action is in the wrong category for the user's current state. An effective unblock system must identify which state the user is in before prescribing anything — then match the intervention to the state, not to the task.

Sources

Try Resistaa — built on this research

Try your first unblock free.

Try Resistaa →

Discussion

Comments are moderated before publication.

0/1200
Why ADHD Tools Create Loops — Not Momentum