In page 107 of "Meyn, S. P., & Tweedie, R. L. (2012). Markov chains and stochastic stability. Springer Science & Business Media",
a distribution $\Gamma$ is said to be "spread-out" if some convolution power $\Gamma^{n*}$ is nonsingular with respect to Lebesgue measure.
I find there are few straightforward explanations on this definition. Can anyone provide some intuition?
