Articles

Why there are no streaks in I Have No Self Control

By Michael Fuchs · Published 2026-10-08

In short A tester asked for streaks. My app has none, and no scores, badges or guilt messages either, and that is less of a scientific decision than it may look. In the studies that exist, streaks, points and badges mostly did what they promise, though almost all of those studies ran for minutes or weeks, and none followed adults with ADHD over time. Little of it reaches an app like mine. I leave streaks out for three reasons I can name: my own experience that they wear off, my expectation that a broken streak costs more than no streak, and the way the app is built, where asking for more time is a whole answer and not a failure. Scores, guilt messages and badges are missing too, for reasons close to these.

A tester asked me for streaks this week: a count of the days in a row spent inside the limit. Many habit apps have one. Mine has none, and no scores, badges or guilt messages either.

That is less of a scientific decision than it may look. I read what studies I could find, and most are small, short, or about something else: a task of a few minutes, a course at a university, children in a laboratory. Very little reaches an app like mine, used for months by people like me. Streaks come closest to an exception, and the exception runs against me: where they have been tested, they mostly did what they promise. Even with help from AI tools, I found no study of whether that lasts in an app over months.

So much of what follows is a matter of taste. This article says what the studies show and where they stop, and then gives my reasons. Some rest on studies. Most rest on my own judgement, and I mark which is which.

What streaks can do

In 7 studies, most of them online experiments, people who were shown an intact streak in a log of what they had done were more likely to carry on than people who were shown a broken one, even when both had done exactly the same [1].

A newer paper went further and paid for streaks. In 8 online experiments, most of which let crowd workers choose again and again between a dull paid task and a funny video, one group earned the same small reward for every task. The other group earned less for the first task, more for the next ones in a row, and dropped back to the lowest rate after a break. That second group was paid less for the same work, and still did more of it in each of those experiments. In the remaining one, people only imagined the work, and said they would do more [2].

Game elements without a streak have mostly done well too. In an online task with 273 people, those who saw points, levels or a leaderboard wrote more tags for the same 15 pictures than those who saw a plain version, though not better ones, and the groups did not differ in how motivated they said they were afterwards [3]. On the trading site of a Finnish university, the 1,579 people who registered in the year after badges were introduced were more active than the 1,410 who had registered the year before; once the growth of the site and the time each person had been registered were taken into account, that held for completed trades, comments and page views, and no longer for trade offers [4]. A meta-analysis of 16 studies found that gamified programmes raised physical activity. Of those studies, 6 looked again 12 weeks or more after the programme had ended. Taken together they showed no significant effect; with one of them set aside, a very small to small one appeared, which the authors read as a sign that the effect outlasts the novelty [5].

Even the study I know that looks at the dark side of streaks comes down in their favour. It interviewed 17 runners whose streak of at least 100 days had ended. Disappointment was the feeling they named most often, and several described a grief that went on for weeks and months. The authors still conclude that streak running has the potential to support lasting change like few other techniques, with a small risk of backfiring for some [6].

What the studies did not measure

Now look at how those tests were run. The paid streaks lasted one sitting of 5 or 10 choices [2]. The picture-tagging study took about 22 minutes, questionnaires included [3]. The badge study reports only totals per person, with no breakdown by time, so the paper does not show whether the lead held or shrank over the year, and its author notes that “novelty has not been directly measured in related studies” [4].

A few studies have followed a game layer over time. One of them tracked it in 2-week steps. At a Brazilian university, 756 students practised programming for 14 weeks on a system that was gamified for some year groups and not for others. For the first 4 weeks the gamified groups were clearly more active. Then the difference shrank: on one of the 3 measures it was not significant for 3 of the 2-week steps, on a second it was only marginally significant in one, and from somewhere between week 6 and week 10 it came back in part, but not to where it had started [7].

For ADHD there is less still. The nearest paper I found interviewed 14 adults who reported having ADHD about an app of 37 static pages, built by the authors’ own university together with a health service and an ADHD charity. Its one sentence on streaks is the authors’ summary: “Rewards, badges or streaks” for reading and for mindfulness exercises “were also noted as ways to improve engagement” [8]. No participant is quoted asking for them. The one who is quoted on a game element wanted lessons that open one step at a time [8].

What is established is narrower, and it is about children. A review of 22 studies with 1,181 children found clear evidence that reward improves performance with and without ADHD, and that response cost does too, where a wrong answer takes away part of the reward. For children with ADHD, response cost seemed to be as effective as reward. On 3 of those studies the review also found that children with ADHD seem to choose an immediate reward over a later one more often than other children [9]. On whether a reward wears off faster with ADHD, the review’s text discusses a single result, and it is about the body: in one study, the heart rate of children with ADHD got used to repeated rewards sooner [9]. The nearest experiment I read asked a different question, what happens when a reward stops. It had 111 children aged 8 to 12, 49 of them with ADHD, learn a rewarded response and then go 2 minutes without any reward. In the planned comparisons the groups did not differ [10].

So nothing I have read shows that streaks wear off for people like me. Nothing shows that they keep working either. As far as I could find, with AI tools doing the searching, the question has not been asked.

Why I leave them out anyway

The second reason is about what a broken streak says.

The third reason needs no study, because it is about how the app is built. It gives the apps you choose a daily limit. When the limit is used up, a curtain stands in front of them with two buttons of the same size. One says Not now. The other leads to more time, at a price you set earlier, a short walk or a wait, and paying it earns a ticket for more minutes. Both buttons are whole answers. A streak of days under the limit would make one of them the wrong answer: paying the price, which the app offers in good faith, would break it. The question on the curtain, whether more is worth it right now, would turn into a test, and the app into its examiner.

What else is missing

Scores and charts

The app shows no percentage of days kept, no weekly chart and no grade. The evidence here is thin in both directions. In a 4-week test with 31 people, each week brought another version of a modified Twitter app, and some versions offered a page of usage statistics and showed a pop-up after 20 minutes of use. Prompts during use picked up no significant difference in how much in control people felt, in a sample the authors call small, while a set of changes to the app’s feed made a small one. Asked week by week, people were more likely to say that the statistics page made them feel in control than that the pop-up did. The time they spent showed no significant change either way [14]. In a 2-week study, 25 students estimated their screen time each day and compared it with the real figure, and 20 of them finished. Several participants reported guilt and disappointment at the numbers, “even for what they considered meaningful activities”, while positive mood, asked at the start and at the end, rose by about 10% on average, in a study without a comparison group [15].

Guilt

The app never says that you failed. That is a stance, and the studies do not hand it to me. In an analysis of 1,529 store reviews of 61 self-control apps and browser extensions, one reviewer wrote of a punishment that it “almost says ‘you’ve failed’”, and another, of a browser extension, “I like the guilt trip it gives me if I attempt to get into facebook or other social networks”. In 9% of the reviews people asked for stricter or more persuasive tools. The authors filed 2% of the reviews under the bad feelings such tools can set off, and some of those reviewers welcomed them [17]. The smokers’ study above goes further against me: self-blame after a slip that followed a longer smoke-free stretch went with a later next slip, and its authors suggest that treatment should not try to talk people out of it. For guilt itself the same study found both directions, depending on whether a person’s patch held nicotine or was a placebo [12].

Badges and leaderboards

Nothing in the app can be collected, and nobody is ranked. I know of one study in which badges went badly, and it is small: 71 students in 2 sections of one course, not assigned at random and taught by one of the authors. Over the 16-week semester, motivation dropped once and stayed lower in the section with a compulsory system of 22 badges and a leaderboard, while it rose in the section without [18]. The picture-tagging experiment found no harm to motivation from points, levels or a leaderboard in its one short session [3].

What the app does instead

The day starts over. Each night the limit comes back in full, and nothing carries over in either direction. There is no record to protect and none to repair.

What you earn is yours for the day. A paid price gives a ticket for more minutes. If you do not use them, they lapse with the day, and the app says nothing about it.

There is room for bad days. When you are about to switch a limit off altogether, the app offers extra time from a small weekly reserve, as a smaller step than giving up. In a field experiment, people who were asked to reach a step target every day but had 2 emergency skips a week reached it on 4.00 days a week, against 2.83 days for people asked for every day without skips [19]. That was a goal to do more of something. Even with help from AI tools, I have found no study that tests a reserve for screen time.

What speaks against my choice

Streaks mostly work where they have been tested, and I may be giving up something that would keep people with the app. People ask for them; my tester did. Some users want a harder push than I am willing to build.

The sharpest point is against the curtain itself. The pop-up in the Twitter study told people after 20 minutes of use how long they had been in the app and offered a button to leave, and 6 of 11 interviewees found it more annoying than useful; one said, “I’d rather make that decision on my own, to close the app” [14]. The same person thought such a feature would be good for people who cannot stop, and only 6 of the 31 had wanted to use Twitter less in the first place [14]. My curtain differs in two ways: the limit is one you set yourself, and more time has a price. I know of no study that tests whether that is enough of a difference, and I have not tested it myself.

How I checked

I searched and read for this article with AI research tools on 8 October 2026. They read each paper in full, methods, results and limitations, and then a second pass held every sentence that cites a paper against that paper’s text. I cite 16 papers here for the first time. Five of them were open only in the authors’ own versions: two manuscripts [18; 3], an accepted manuscript [12], a dissertation reprint [9] and a preprint [15]. Three papers carry a published correction; all three were read, and none touches a result I use. Every number is traced to a passage of its source. The searches found no long-term study of streaks in an app and none in adults with ADHD; a study I did not find may still exist. If you find something I got wrong, the imprint has my address.

This site belongs to an iPhone app: a daily limit on the apps that pull you in, and a price on more time, both set by you earlier. See how it works