Spaced repetition is one of the most replicated findings in memory research. The principle — that reviewing material at gradually increasing intervals produces stronger retention than massed study — has been demonstrated in controlled settings since the late 19th century. What differs across available software is how each tool implements the underlying algorithm and how practical the workflow is for adult learners.
How spaced repetition software works
All SRS applications share a common structure: each item in a deck carries a difficulty rating and a scheduled review date. When a review session begins, only items whose date has arrived are shown. After each response, the algorithm adjusts the next interval — shortening it if the answer was incorrect, extending it if correct.
The specific formula varies by application. Anki uses a modification of the SM-2 algorithm developed by Piotr Woźniak for SuperMemo in 1987. The original SuperMemo application now uses SM-18, a substantially more complex variant. Both approaches use similar conceptual foundations but differ in how they model long-term memory and handle lapses.
The SM-2 algorithm in practice
SM-2 calculates the next interval based on a per-card "ease factor" — a multiplier that increases when answers are correct and decreases after failures. A card with an ease factor of 2.5 will, after a successful review, be scheduled 2.5 times further into the future than the previous interval. A new card shown today with an initial interval of one day becomes due in 2.5 days after the first success, then roughly 6 days, then 15 days, and so on.
The system is straightforward to inspect and modify. Anki exposes all SM-2 parameters in its deck options, which means experienced users can calibrate interval growth to match their retention goals. The trade-off is that tuning requires understanding what the parameters actually control.
Anki: the dominant open-source option
Anki is free on desktop (Windows, macOS, Linux) and on Android. The iOS version costs approximately €25 as a one-time purchase. Synchronisation across devices is handled through AnkiWeb, a free cloud service maintained by the developer.
The desktop interface is functional but not polished. New users often find the deck and note type system confusing, since Anki separates the concept of a "note" (the content) from a "card" (a specific question-answer pair generated from that note). A single vocabulary note can generate two cards — one for recognition, one for production — which doubles the review load but produces stronger encoding.
Deck resources for Polish learners
AnkiWeb hosts a public deck repository. Relevant decks for Polish-based learners include frequency-sorted vocabulary lists for English, German, Spanish, and French — the four most-studied foreign languages in Poland according to Central Statistical Office data from 2023. The quality of community decks varies considerably; decks with over 10,000 downloads and a review history of several years are generally more reliable than newer uploads.
Polish language learners — those studying Polish as a foreign language — will find fewer ready-made decks. The most commonly used approach among this group is building custom decks from frequency lists and sentence mining through tools such as subs2srs or Yomichan-style browser extensions.
SuperMemo: the original, still under active development
SuperMemo was created in Poznań by Piotr Woźniak in the late 1980s and remains in active development. The current version, SuperMemo 18, implements a memory model that attempts to estimate the probability of forgetting each item and schedules reviews accordingly. The underlying model — called the Two-Component Model of Memory — treats memory as having both a stability component (how resistant a memory is to forgetting) and a retrievability component (how easy it is to recall at the moment of review).
SuperMemo is Windows-only and has no free tier beyond a limited trial. The interface reflects its long development history and is less intuitive than Anki for new users. However, the algorithm is more sophisticated, and users who have used both systems over multi-year periods often report that SuperMemo produces better retention at lower daily review volumes.
Practical considerations for Polish-based users
SuperMemo is developed in Poland and the Polish-language community around it remains active. The developer's website (supermemo.com) contains extensive documentation on memory research, much of it written by Woźniak himself. For learners interested in understanding the theoretical basis of spaced repetition, this is a more detailed resource than anything Anki provides.
Alternatives: Memrise, Duolingo, and browser-based tools
Several applications use spaced repetition as one component of a broader learning format rather than as the primary interface.
Memrise
Memrise combines SRS with video clips of native speakers pronouncing and using vocabulary in context. The spaced repetition implementation is less configurable than Anki or SuperMemo, and the interval algorithm is not publicly documented. For learners who want structured vocabulary exposure without the overhead of managing a custom deck, Memrise offers a reasonable starting point. Free and paid tiers are available; the paid tier unlocks offline access and additional content.
Duolingo
Duolingo incorporates spaced repetition into its lesson sequencing, though it does not expose this to users as a review system. The primary design goal is daily engagement rather than optimised retention scheduling. Retention research comparing Duolingo learners to matched control groups shows vocabulary gains consistent with the time invested, but the gamification structure encourages short sessions over the kind of deliberate review that dedicated SRS applications provide.
Physical Leitner boxes
The original paper-based system developed by Sebastian Leitner in the 1970s uses physical card boxes numbered 1 through 5. Cards in box 1 are reviewed daily; box 2 every two days; boxes 3, 4, and 5 at progressively longer intervals. The system requires no software and produces tactile engagement that some learners find motivating. It is less efficient than algorithmic SRS for large vocabularies but remains a practical option for learners building a core set of 500–1000 items.
Which system fits which learner
The choice between SRS tools depends on three variables: how much configuration the learner wants to manage, whether they need cross-device synchronisation, and whether they are building their own deck or using existing resources.
- Learners who want maximum control over interval scheduling and access to the largest community deck library: Anki.
- Learners willing to invest time in understanding memory theory and who study primarily on Windows: SuperMemo.
- Learners who want curated content without building decks: Memrise or Duolingo as supplements to a primary SRS tool.
- Learners building a small core vocabulary without screens: physical Leitner box.
No single system is optimal for all use cases. In practice, many learners in Poland use Anki as their primary review environment while using Duolingo or Memrise for initial exposure to new material before it enters the review queue.
Common errors in SRS practice
The most frequently documented error is adding too many new cards per day. The SRS workload grows predictably: adding 20 new cards daily in a standard Anki configuration will produce approximately 100 daily reviews after three weeks and 200 after six weeks. Learners who do not anticipate this growth often encounter an unmanageable backlog and abandon the system.
The second common error is creating cards that are too complex. A card asking for a full conjugation table produces less reliable encoding than six separate cards, one per verb form. The principle — one simple fact per card — was articulated by Woźniak and is consistently supported in the research literature on item difficulty and SRS performance.
Sources and further reading
The foundational papers on spaced repetition and the forgetting curve are available through standard academic databases. Ebbinghaus's original 1885 work has been digitised and translated. Wozniak's documentation of SM-2 and subsequent algorithms is published on supermemo.com. The Anki manual is maintained at docs.ankiweb.net and covers all algorithm parameters in detail.
For learners in Poland, the SuperMemo website and the AnkiWeb platform are the primary reference points for software documentation.