Grade each flashcard by how much effort recall actually took, not by how you feel about the card or yourself. Again means you didn't produce the right answer before flipping the card. Hard means you got there, but it cost you: hesitation, a false start, or a need for the note field. Good means normal, correct recall, no real struggle. Easy means instant, no effort at all. Rate what happened in the second or two before you saw the answer, not how obvious it looks in hindsight. Mayura's scheduler runs on FSRS and uses your rating to set the next interval, so an inflated rating doesn't reward you. It just books the card for a date you're less likely to remember it.
You flip the card, there's a two-second pause where your eyes dart around like you're checking for cameras, and then you tap Good anyway. Mayu saw that pause. Rate the pause, not the pride.
Why do people rate their flashcards more generously than they should?
People inflate ratings because Again feels like admitting failure, even in a session with an audience of one. Guilt plays a part too: mark five cards Hard in a row and it feels like evidence you're bad at Telugu, which isn't what the rating means. Then there's sunk cost. You drilled a word for three weeks, so surely you should know it by now, and that "should" quietly turns a wobbly recall into a confident tap on Good.
Mayu's take: the scheduler doesn't see your face or your hesitation. It only knows what you tell it.
How do you tell Again from Hard in the moment?
Again is a total blank, a wrong guess, or a guess that only matched because you'd ruled out everything else. Hard is when you got there, but it cost you: a false start, or needing the note field. Watch for relief. A flash of "oh yeah, right" when you flip the card means you recognized the answer rather than recalled it, and recognition isn't recall. Picking a transliteration out of four MCQ options doesn't count as honest flashcard recall; generating it cold does.
Mayu's take: needing the mnemonic in the note field is Hard doing its job, not something to be embarrassed about.
How do you tell Hard from Good, and Good from Easy?
Good is correct recall at a normal pace, no real hitch, the baseline for a word you've studied but haven't mastered. Hard is also correct, but it took visible effort: a beat of silence, or a near-miss you corrected. Easy is correct and instant, arriving before you've finished reading the prompt. Already anticipating the answer while your eyes are still moving across the Telugu script? That's Easy. Had to go looking for it, even briefly? Good at best, probably Hard.
Mayu's take: "instant" is a low bar on purpose. If you have to think about whether it was instant, it wasn't.
What goes wrong long-term if you keep rating generously?
FSRS sets your next review date from a stability estimate, and that only works if your rating reflects what happened. Rate Good or Easy when the truth was closer to Hard or Again, and the interval stretches further than your memory can hold, so the card comes back already forgotten. That feels like a fresh failure instead of a normal one, and it repeats: cards you've told yourself you "know" keep lapsing, giving a false sense of progress even as you log real reps. Your retention target, 90% by default, only means what you think it means if the ratings feeding it are honest.
Mayu's take: an overdue card that surprises you isn't the app failing. It's usually last month's generous rating catching up with you.
What's a quick checklist to grade yourself honestly?
Run through a few questions before you tap a rating, especially on cards you've rated the same way out of habit for weeks. Did you produce the answer before flipping, not after? Was there a real pause, since anything longer than instant rules out Easy? Did you lean on the note field or a related word, which points to Hard rather than Good? Stuck between two ratings? Take the lower one. It costs a slightly earlier review; the higher one costs an inaccurate picture of what you know.
A few Telugu words double as honest self-talk here. Chaalu (చాలు, enough) is the gut check for Good or Easy: if a card is genuinely chaalu, bet on it cold. Inkonchem (ఇంకొంచెం, a little more) is the honest version of Hard, admitting a card isn't solid yet. Konnisaarlu (కొన్నిసార్లు, sometimes) is a card you only get right sometimes, without help, a Hard rating wearing a Good costume. Eppuduu (ఎప్పుడూ, always) is the real bar for Easy, not "usually." Pratiroju (ప్రతిరోజు, every day) is the habit underneath it all.
| Transliteration | Telugu script | English | When to use |
|---|---|---|---|
| chaalu | చాలు | enough | gut-checking whether a card really earns Good or Easy |
| inkonchem | ఇంకొంచెం | a little more | admitting a card isn't solid yet, honest Hard |
| konnisaarlu | కొన్నిసార్లు | sometimes | a card you only get right without help sometimes |
| eppuduu | ఎప్పుడూ | always | the real bar for Easy, not just "usually" |
| pratiroju | ప్రతిరోజు | every day | the review habit that makes honest grading worth it |
Try it
- Next session, say your answer out loud before flipping the card, so you get real recall instead of "I probably know this."
- Pick a card you've rated Good on autopilot for weeks. Next time there's any pause, rate it Hard and watch its schedule shift.
- Use the note field for real mnemonics, and treat leaning on one as a legitimate Hard trigger, not a cheat sheet.
- When an overdue card surprises you, ask if it's one you've been over-rating. That's your inflate-o-meter.
- For one week, default to the lower rating whenever you're torn, and see if your due count gets more honest.
FAQ
Is it bad to rate a card Again a lot? No, it usually means the scheduler is still calibrating your real recall for that word. Honest Again ratings beat a string of inflated Goods that fall apart later.
What if I'm not sure whether something was Hard or Good? Take the lower rating. A slightly earlier review costs minutes; an inflated one costs a schedule that no longer matches your real memory.
Does rating Hard instead of Good hurt my streak? No. Streaks and session counts track that you reviewed, not what you rated, so an honest Hard still counts as a completed review.
Why rate on a scale instead of marking right or wrong? A binary right or wrong can't show how close you were to forgetting. Again, Hard, Good, and Easy give FSRS the effort signal it needs.
Can changing my retention target fix over-inflated ratings? No, they're separate levers. Lowering the target reduces review volume, but only honest ratings fix a drifted schedule.
Try Mayu
Open your next review and grade what actually happened, not what you wish had happened.
