Figure 4.

Simulation results compared to the lower and upper bounds of Theorem 2 on the number of distinct phrases resulting from LZ78-parsing of binary sequences of length l generated by sources with three different source probabilities $P\left(X=1\right)$. For every $P\left(X=1\right)$, one thousand binary sequences of length l are generated. Error bars represent the maximum, minimum, and average number of distinct phrases.