Omniversail guide
Reviewing material right before you forget it produces far more retention per minute than reviewing it while it is still fresh.
Memory decays on a predictable curve: quickly at first, then slowly. If you review something just as it is about to drop out, you reset the curve and it takes much longer to fade again.
That is why cramming feels productive and then evaporates, and why a short daily review beats a long weekly session even when the weekly session covers more material.
Choosing what to review by hand is where most self-study systems die. The list is always too long, the temptation to postpone is always there, and the items that need review are rarely the ones you would choose.
An adaptive system removes the decision. It uses your actual performance on each item to estimate when that specific item will be forgotten, and puts it in front of you at that moment.
Drills install new material and build automaticity. Review keeps installed material from decaying.
They have to be interleaved. A day of pure new material is a day of maximum forgetting; a day of pure review is a day of no progress. Omniversail schedules both against the same model so the trade-off happens automatically.
A large queue is normal, not a sign you have failed. It usually means you added a lot of material recently, or that earlier material was left too long.
The fix is to keep the daily session small and consistent. The queue is drained by daily repetition, not by one heroic session, and its length is not a measure of how well you are doing.
For vocabulary, review is close to the whole job, because the words are the content.
For grammar and patterns, drills do the installing and review does the maintenance. If you can already produce a pattern without thinking, it should leave the review queue and stay in the drill rotation.