Lemmerly rates your vocabulary the same way chess ratings work — with ELO. Words have ratings. Players have ratings. Both shift over time based on performance.
ELO is a rating system invented by physicist Arpad Elo for chess. The key insight is simple: your rating should reflect not just whether you win or lose, but who you beat.
If you beat a very strong opponent, you gain a lot. If you beat a weak opponent, you gain a little. If you lose to a strong opponent, you lose very little. If you lose to a weak opponent, you lose a lot.
This makes the rating self-calibrating and continuously accurate. It is not a test score from one day. It is a living number that reflects your true level over time.
Every word in Lemmerly starts with an estimated ELO based on its frequency, complexity, and origin. Rare, technical, or archaic words start high. Common, everyday words start low.
But the ratings do not stay fixed. Every time a word appears in a session, the outcome updates its rating. If many players answer it correctly, the word's ELO decreases — it is proving easier than expected. If many players miss it, its ELO increases — it is proving harder than expected.
The same applies to the multiple choice options — the distractors. If a particular wrong answer fools many players, that entire question becomes harder. The system learns continuously from millions of answers.
A word's ELO is not fixed. It is a consensus rating — the aggregate judgment of every player who has faced that word.
Your ELO starts at 800 when you join. After each answer, it adjusts based on two things: whether you got it right, and the difficulty of the word you faced.
Answer a 1600 ELO word correctly when you are rated 1400 — big gain. You beat an opponent stronger than you. Answer a 900 ELO word correctly when you are rated 1400 — tiny gain. You were expected to get it right. Miss a 1600 ELO word when you are rated 1400 — small loss. You were not expected to know it. Miss a 900 ELO word when you are rated 1400 — large loss. You should have known that one.
This means your rating converges quickly on your true level. In the first few sessions, ratings move fast as the system calibrates. After that, movement slows — each new answer is one data point among many.
Here is the subtlety that makes ELO more interesting than a static score: because word ratings shift over time, the value of past encounters changes too.
Suppose you correctly answered the word laconic six months ago, when it was rated 1450. That encounter gave you a certain ELO gain. But since then, many players have answered laconic correctly — its rating has drifted down to 1380 as it proved easier than expected.
If you face laconic again today, correctly answering it earns you less than it would have six months ago. The word got easier — so beating it counts for less.
This is exactly how chess works. If your regular opponent improves dramatically, beating them now is worth more than it was a year ago. The system is always current.
Your ELO is not a memory of past performance. It is a live estimate of your current vocabulary level.
Words you miss are added to your review queue. The review system uses spaced repetition — words are shown again at increasing intervals: 1 day, 3 days, 7 days, 21 days, 60 days, 180 days.
Correctly reviewed words advance to the next interval. Missed words reset to 1 day. Over time, words you truly know fade from the queue. Words you struggle with keep returning until they stick.
Review sessions use a modified ELO — correct answers give 40% of the normal gain (you have seen the word before), while wrong answers give the full penalty. This keeps the rating honest.
Eight titles mark your journey. Each title requires reaching a specific ELO threshold. Unlike other platforms, titles are not awarded permanently — if your rating drops below the threshold, the title changes accordingly. Your title always reflects your current level.
The highlighted level — Lex Expert — is where most serious readers and graduate students land. Getting above 1700 requires genuine mastery of rare, technical, and literary vocabulary.
Regular short sessions beat occasional long ones. Three sessions a day, consistently, will move your rating faster than one long session per week. The spaced repetition system is designed around this — it surfaces words at exactly the moment your brain is about to forget them.
Read widely. The Gutenberg book ratings on Lemmerly show you which books match your current level. Reading at or just above your ELO level — what linguists call i+1 input — is the most efficient way to build vocabulary naturally.