Next Number in Sequence
One sequence per line, terms split by comma or space (negatives, decimals and large integers are supported) - the next terms are inferred automatically.
Calculation Result
Download CSV
| No. | Known terms | Model used | Basis | Formula / recurrence | Next term (predicted) |
|---|
Introduction to the tool and how to use it
Put the first few terms of one sequence per line (comma or space separated; negatives, decimals and large integers are all supported) and the tool infers the next term, or several terms ahead.
Models are tried in priority order: arithmetic / geometric → finite-difference polynomial (order ≤ 6, extrapolated by a difference engine) → second-order linear recurrence an = p·an-1 + q·an-2 (solved exactly and by least squares) → well-known sequences (Fibonacci, Lucas, primes, squares, cubes, triangular numbers, factorials, powers of 2, Catalan numbers) → odd/even grouping. Every model is verified against the known terms and only fully consistent ones are accepted; the first one that passes wins. Turn on "Show other candidates" to list the other passing models, each marked "Candidate".
Parameters: terms to predict (1-50), max terms used per line (default 100, up to 500), strict mode (all known terms must match, or allow the tolerance given by the decimal-places setting), decimal places (output precision, also the tolerance in loose mode) and the output separator.
If no model fits, a "first-difference continuation" gives a reference next term clearly marked "uncertain" - it simply extends the last difference and is not guaranteed.
Limits: up to 300 lines, 500 terms per line and 50 predicted terms per run; anything beyond is silently trimmed. Integer sequences print as integers (never 13.000000) and integer literals are handled exactly with big integers. To identify the pattern itself only, use the sequence pattern detector.
Models are tried in priority order: arithmetic / geometric → finite-difference polynomial (order ≤ 6, extrapolated by a difference engine) → second-order linear recurrence an = p·an-1 + q·an-2 (solved exactly and by least squares) → well-known sequences (Fibonacci, Lucas, primes, squares, cubes, triangular numbers, factorials, powers of 2, Catalan numbers) → odd/even grouping. Every model is verified against the known terms and only fully consistent ones are accepted; the first one that passes wins. Turn on "Show other candidates" to list the other passing models, each marked "Candidate".
Parameters: terms to predict (1-50), max terms used per line (default 100, up to 500), strict mode (all known terms must match, or allow the tolerance given by the decimal-places setting), decimal places (output precision, also the tolerance in loose mode) and the output separator.
If no model fits, a "first-difference continuation" gives a reference next term clearly marked "uncertain" - it simply extends the last difference and is not guaranteed.
Limits: up to 300 lines, 500 terms per line and 50 predicted terms per run; anything beyond is silently trimmed. Integer sequences print as integers (never 13.000000) and integer literals are handled exactly with big integers. To identify the pattern itself only, use the sequence pattern detector.
Message board
All messages →-
No one has spoken up yet — want to go first?