Practice that meets every student where they are, then moves with them.
Each student starts at a target reading level. Free activities run statelessly (no prior-attempt history), so the starting level comes from the assignment defaults: typically the middle level for the activity, with a low or high override available. You can pin the starting level for the whole class through the Assign modal, or let the default ride.
Upshifts happen only when a student is perfect so far. At 40% of the way through, a perfect run bumps the student up one level (if a higher level is available). At 80% complete, a still-perfect run bumps up a second level. Downshifts happen when a B (80%) is no longer mathematically possible. Either way, the next item is already retuned by the time the student gets to it. There is no waiting until the end of the assignment to recalibrate.
Most adaptive systems just keep pushing harder content at students who are missing items, which is how kids learn to hate reading practice. Ereading.AI watches the math. The moment a B (80%) is no longer mathematically possible, the activity steps down a level. Drops are staged. A student who started at level 4 steps to level 3 at the halfway mark, then to level 2 at the final threshold. A student who started at level 3 steps down once to level 2. Nobody crashes to the bottom in one move, and the experience does not collapse on the student.
If a student bumped up early and then started missing items, the system rolls back gracefully. If the max possible final score falls to 94% or below, they drop one level and further upshifts are locked. If they had moved up two levels and the max possible drops to 90% or below, they fall all the way back to the starting level and stay there. The goal is to reward strong performance with harder material, but pull back before the student crashes.
The final score is what passes back to the LMS and the gradebook today. On paid teacher attempts, difficulty tracking (which level shifts happened and where the student finished) is captured on the attempt record while the result PDF is being generated. Surfacing that tracking in the teacher results view is on the roadmap, not in the UI yet.
Pick a skill, assign an activity, see how it grades itself.
