σ² = 4, n = 100, ε = 0.5
bound σ²/(nε²)
0.16.
The weak law of large numbers says the sample mean converges in probability to the true mean: for every , as .
This page is easier if you are comfortable with:
Chebyshev's inequality applied to (variance ) gives (Chebyshev's inequality). For a fair coin, , , : the bound is , while the exact binomial probability of or heads is 0.0569 (the binomial distribution). To make the bound at most 0.05 we need .
For a fair die (), , : the bound is . The bound is crude (compare the central limit theorem, which gives the actual normal-shaped fluctuation), but it proves the principle: relative frequencies settle down, which justifies the interpretation of probability as long-run frequency (relative frequency).
The law is the basis of Monte Carlo methods, whose accuracy is described by the Monte Carlo error, and of estimation by sample averages (mean squared error). The strong law is a stronger statement: the averages converge for almost every infinite sequence.
Averages settle near the true mean as the sample grows, which justifies estimating by sampling.
More tosses make a large deviation less and less likely.
These examples are worked by hand, step by step, so you can compare your own working.
Bound for 100 fair coin tosses.
How large must be so that the bound is at most 0.05?
Bound for 100 rolls of a fair die.
A few situations come up again and again. Knowing them saves time and helps you spot answers that cannot be right.
bound σ²/(nε²)
0.16.
bound
0.04.
n ≥ σ²/(ε²·0.01)
1600.
n = 100
bound 0.04.
The sample mean converges in probability to μ. Chebyshev gives a quantitative bound: P(|X̄ₙ − μ| ≥ ε) ≤ σ²/(nε²).
σ² = 4, n = 100, ε = 0.5. Find σ²/(nε²).
Find the bound for n = 400.
Find n making the bound 0.01.
Find the bound for ε = 1, n = 100.
σ/(nε²) ✗
σ²/(nε²) ✓
Variance.
σ²/(nε) ✗
σ²/(nε²) ✓
Squared.
X̄ = μ ✗
close with high probability ✓
Not equality.
same ✗
probability vs almost sure ✓
Different modes.
Try each one on paper, then open the answer. Use the solver above to see the full working.
0.09
1
80
0.01
80
0.4
Unlimited fresh questions with instant marking. Choose a level, work it out on paper, then type your answer.
Solve: …
Sample means converge in probability to the mean.
By Chebyshev’s inequality.
σ²/(nε²).
It gives probability, not almost sure convergence.
Chebyshev’s inequality.
More in Advanced statistics and stochastic processes and the formula sheets.