There's a real tension at the heart of every learning platform: when do you stop letting the kid practice and make them prove it?
Most platforms pick one of two extremes. Neither works well.
Extreme 1: never test
Apps in this camp let the kid drift forever in 'practice mode'. The kid feels good, the parent feels good, but no one really knows if anything is sticking. Three months in, the kid is still on the same hex with no closure.
Extreme 2: test on a fixed schedule
ALEKS and a few enterprise platforms do this. Every N hours of practice, a Knowledge Check pops up — unannounced, often without warning. If the kid does badly, progress can roll BACKWARD. Reviews of these systems are full of words like 'meltdown', 'anxiety', and 'I cried'. Forcing high-stakes assessment on a 9-year-old without warning is not pedagogy. It's stress engineering.
What we picked instead
Rodybee uses a 'thermometer' — a continuous multi-signal check that runs after every practice session. It looks at four things together:
(1) Are all the skills in the current block actually mastered? (accuracy AND fluency, see the previous post)
(2) Are the kid's last 10 exercises in the block strong (≥90% accuracy and on-target time)?
(3) Has it been at least 24 hours since the last block test in this program? (prevents speed-running)
(4) Has the kid done at least 3 distinct sessions in this block? (prevents passing on a lucky streak)
If all four hold, the system tells the kid: 'I think you're ready for the block challenge. Want to try it?' The kid can say yes, or 'maybe later'. No pressure either way.
What if the kid keeps saying 'maybe later'?
If readiness keeps holding for many sessions in a row (the exact number depends on a parent setting: permissive=30, balanced=15, strict=5), the test becomes mandatory. But — and this is the key — it doesn't pop up cold. The kid gets a conversational message ('You've been ready for a while — let's pick a day to do it before [date]') and a 3-day window to choose when. Not 'now or else'. 'Before this date, on the day you pick'.
What if the kid stops being ready?
If the kid's accuracy drops or they take longer to answer, the system stops nagging. The mandatory window closes. Practice continues. The thermometer goes down. We will never force a test on a kid who isn't actually ready — that's the worst version of testing, and the one most likely to cause real harm.
Where the parent fits
The parent has the override that the kid doesn't. From the dashboard, the parent can manually trigger a checkpoint when they see their kid is ready but stalling out of habit. They can also set the system's default insistence: permissive (30 sessions of grace), balanced (15), strict (5). Different kids need different defaults. Same family, different ages — same dashboard, different sliders.
Why this is different
We treat the test as a moment of confirmation, not a gatekeeping event. The hard work happens in the daily practice. The test just verifies what the practice data already strongly suggested. When the data is right, the test is calm. When the data is wrong (kid passes by luck or fails despite readiness), it's a signal to fix the data — not a verdict on the kid.
Self-Determination Theory has a phrase for this: autonomy-supportive instruction. Kids learn more, retain more, and ENJOY more when they feel they have agency over the pace. We try to honor that without abandoning the rigor that makes mastery actually mean something.