A review schedule is only as good as the evidence it is built from. These five habits all feel like extra effort, and each one feeds the schedule something it cannot use — leaving you with review dates that describe a version of your memory that does not exist.
1. Replaying the same round to get a better score
Answering a word again within ten minutes of the last answer that counted is recorded but does not raise its mastery or move its review date. The answer is still in your working memory, so getting it right proves nothing about Thursday.
Without that guard, three replays would walk a set to full mastery and schedule it a fortnight out having never once been recalled cold. The guard protects you, but the effort you spend on the replays is still spent, and it buys nothing. Come back later instead.
2. Deleting a set and re-importing it
People do this to tidy up a messy import. It works, and it throws away every scrap of study history in the set, because mastery attaches to the word records you delete. The re-imported words are new words as far as the app is concerned, all at mastery zero.
If the problem is formatting, fix the file and import as a new set, then delete the old one only once you are sure. If the problem is a handful of bad entries, edit them in place — an edit keeps the history, a re-import does not.
Getting the file right before the first import
3. Studying one set at a time, set by set
It feels organised: finish set one, move to set two. The cost is that set one's due words go unanswered for as long as set two takes, and the schedule was built assuming you would come back.
The review queue is built across all of your sets at once rather than per set, precisely so that you do not have to remember which set has words waiting. Using it is nearly always better than working through sets in order.
4. Grinding for XP
XP only ever goes up, so it rewards volume, and the fastest ways to earn it are not the best ways to learn. Scramble pays 3 per word and timed Match can double a round — both are excellent modes, and both will happily give you a big number on words you already knew.
If you want to know whether a week of study worked, look at how many words moved up in mastery, not at how much XP you earned. The two often disagree, and mastery is the one that can go down.
What XP measures, and what it does not
5. Letting weak words pile up
A word you have ever answered wrongly stays flagged as weak until it reaches full mastery, and weak words are eligible for review whether or not their date has arrived. Accumulate thirty of them and they will fill the review half of every session on their own, squeezing out new words entirely.
That is the system working, not failing. But it does mean the answer to "why am I not seeing new words any more" is usually a backlog of words you keep missing, and the way out is through them rather than around them.