A function of the current state of a Markov chain can discard information needed to predict its next value. A sufficient condition preserving the Markov property is strong lumpability: all original states with the same recorded value must give the same transition probabilities to every recorded class.
For the two-column model in growth rate of two-column sticky deposition, the pair of column heights is Markov, but its maximum alone is not. Starting empty, the maximum-height histories and both end at height two. The former leaves unequal columns, so the next height increases with probability ; the latter leaves equal columns, so it increases with probability one. Each history has positive probability. Their different predictions at the same time and current value violate the Markov property. By contrast, the absolute height difference retains exactly the information needed for its own transitions.
Past exam of the mathematics course of the University of Cambridge 2014 ib Paper 4 9H b ii Solution Created 2026-09-24 Updated 2026-10-06
Starting from forces . The first nonzero value is consequently . Remaining nonzero at the next step forces , because . Hence, on an event of positive probability,The two values are unequal by assumption. They expose memory of the hidden sign: immediately after leaving zero the sign is negative, whereas after two consecutive nonzero observations it is positive. Whenever the relevant observed histories coexist with positive probability, the Markov property for fails, since its next-step law is not determined by its present value alone. In particular, the absolute-value projection does not preserve the Markov property for all initial laws; the strong lumpability condition would require .
There is a small qualification to a blanket conclusion about an unspecified initial law. The printed assumptions do not ensure that either conditioning event is possible, and a degenerate initial law may never reach the hidden states that cause the failure. For example, in the order , takeAll printed constraints hold, yet always and is a Markov chain. Thus the correct general conclusion is the transformed process is not necessarily Markov; the two computed probabilities give the intended failure when their histories are available.