Explore convergence, divergence, and the precise epsilon-N definition of limits
A sequence converges when its terms close in on a limit and stay there. Set the tolerance, watch sequences settle or fail to, and find order hiding inside chaos.
Pick a sequence, drag the epsilon slider, and press Find N. The smaller the tolerance, the further out N has to go.
Current sequence: aₙ = 1/n
Limit: L = 0.0000
The heart of rigorous analysis: for a sequence to converge to L, we must be able to trap all sufficiently late terms within any tolerance epsilon of L. Drag the epsilon slider to see how the required N changes -- smaller tolerance demands going further into the sequence.
Key insight: Convergence means that for every epsilon > 0, no matter how small, there exists an N such that all terms beyond index N lie within epsilon of the limit. The universal quantifier on epsilon is what gives this definition its power.
Choose a sequence and press Play to watch its terms settle toward the limit. Compare how quickly different sequences get there.
The harmonic sequence converges to 0
Limit: L = 0
Monotonicity: decreasing
Watch sequences approach their limits in real-time. Observe how different sequences converge at different rates -- some approach quickly, others more gradually. The animation reveals the "settling down" behavior that defines convergence.
Key insight: The rate of convergence varies widely. Geometric sequences like (1/2)^n converge exponentially, while 1/n converges much more slowly. Both satisfy the epsilon-N definition, but the required N differs dramatically.
Pick an unbounded, oscillating, or chaotic sequence, then switch on the comparison with 1/n to see how it differs from one that converges.
Terms grow without bound as n increases. No matter how large a number you pick, the sequence eventually exceeds it.
This sequence: Diverges to +∞ (unbounded)
Not all sequences converge. Explore the three main ways a sequence can diverge: growing without bound (unbounded), oscillating between values without settling, or behaving chaotically within bounds. Compare with convergent sequences to see the difference.
Key insight: A sequence can be bounded yet still diverge (like (-1)^n), showing that boundedness alone does not guarantee convergence. Monotonicity plus boundedness does -- that is the Monotone Convergence Theorem.
Choose increasing or decreasing and press Animate. The terms are squeezed toward the supremum or infimum and never pass it.
Every bounded monotone sequence converges to its supremum (if increasing) or infimum (if decreasing).
Current sequence: aₙ = 2 - 1/n converges to 2.00
A powerful result: every bounded monotone sequence must converge. An increasing sequence bounded above is "squeezed" toward its supremum; a decreasing sequence bounded below approaches its infimum. This theorem is fundamental for proving convergence without knowing the limit in advance.
Key insight: The Monotone Convergence Theorem relies crucially on the completeness of the real numbers. In the rationals, a bounded increasing sequence can "converge" to an irrational gap.
Generate a bounded sequence, then press Step to halve the interval, keeping the half with more terms. The terms you keep form a convergent subsequence.
Every bounded sequence has a convergent subsequence.
The Bisection Method:
Every bounded sequence has a convergent subsequence. This demo visualizes the bisection proof: repeatedly halving the interval and selecting terms from the more populated half extracts a subsequence that must converge. This theorem connects boundedness to convergence in a deep way.
Key insight: Bolzano-Weierstrass is equivalent to the completeness of R. It says that bounded sequences, even chaotic ones, always contain hidden convergent patterns.
Spot a mistake or have an idea?