Trying to recall an answer strengthens memory more than reading it again. In the classic experiments, students who were tested remembered more a week later than students who studied the material again.12 Every review is a test: the answer stays hidden until you ask for it. A new card is shown with its answer once, then tested.
The same study, spread over days, is remembered longer than in one sitting. A synthesis of over three hundred experiments on recalling words and facts found spaced study ahead of massed study, and the best gap growing with how long you need to remember.3
FSRS, an open scheduler, models each card's memory with three numbers: how hard it is, how stable it is, and how likely you are to recall it today. It sets the next review for when that likelihood falls to the retention you choose, 90 % by default.45 The September 27, 2026 benchmark snapshot covers some 727 million reviews from 10,000 Anki users,6 and Anki has offered it since version 23.10. RectoVerso runs FSRS-6 on your iPhone, checked against the reference implementation on 1,746 reviews, and against Anki 26.9.3 across five review histories, including lapses and same-day reviews. It refits its memory parameters to your own reviews once there are enough. This checks the scheduler, not the validity of using reveal depth to choose a rating.
Many apps ask you to choose Again, Hard, Good or Easy. In RectoVerso, you still say whether you knew the card, but you do not rate the difficulty. You open the phone, or slide the answer up, only as far as you need, then tap I know it, or I know the rest once part of it shows. The whole answer then shows, so you can check: Next card lets your claim stand, I was wrong takes it back. If you don't know it, tap I don't know. Once all of it shows, the buttons say I didn't know and I knew it. The largest share of the answer you revealed before answering is the effort, and it's recorded with the answer.
FSRS still takes one of its four ratings, and the effort picks it. A Wrong (I don't know, I didn't know or I was wrong) is Again, at any depth. I know it with nothing revealed is Easy, I know the rest part-way is Good, and a Right after 60 % or more of the answer is Hard, I knew it with all of it included. These are bands: revealing more within a band keeps the same rating. New and relearning cards use Good for any nonzero reveal, so they can finish their learning steps.
Letter cues are old: memory researchers have taught words by taking letters away as they're learned.7 In a word-recall experiment, adding letter cues helped compared with studying alone when the full answer was then shown, though less than diminishing cues.8 That result does not validate reveal-based ratings for math, images or other card types. Choosing a rating from how much of the answer you revealed is our design, and we haven't tested it in a study. Every review keeps its effort, how the answer was revealed and how long it took; FSRS's parameters are refitted to your reviews and kept only when they predict your answers better. The reveal-to-rating mapping itself is fixed, not personalized by that fitting. The records could inform a future revision; we do not yet know whether this method predicts recall better than self-rated difficulty.
Your retention score counts Right answers on reviews at least a day apart, once a card is past its first steps, even when you revealed the whole answer. It is not a measure of unaided recall alone. Cards marked Fragile revealed at least 60% of their answer on three of their last four reviews; Wrong answers count too.
The scheduler estimates recall; it cannot promise you will remember a card. Its default target is 90%, and scheduling variation spreads reviews across nearby days. The app records the information actually shown, including discrete steps such as letters or math terms.