NETMaths

Is this true?

An irreducible chain with a stationary distribution converges to it

No — it is false.

The counterexample

The two-state chain that swaps deterministically (period 2)

½, ½) is stationary and unique, but oscillates between 0 and 1. Aperiodicity is required.

The kind of mistake this is

Execution slip

The idea was right. The computation was not.

Drill statements like this

Others that fail the same way

From Limit Theorems and Markov ChainsMarkov chains: classification of states, stationary distributions

ShareWhatsAppTelegram