Example of skewed distribution with unexpected ordering of mean, median, and mode.

[brian.field]

Miller mentions at the bottom of page 50 that many practitioners inaccurately believe that the ordering of the mean, median, and mode is just that (from left to right) for negatively skewed distributions (and mean, median, and mode (from right to left) for right skewed or positively skewed distributions.) but that this is not always the case.

Can someone provide a counter example?

Thanks!