There have been ~391,000 games played since 1876. With two lineups per game.

Starting from the leadoff hitter, if you assume that there is a 50/50 chance of each subsequent player's name starting with a letter later in the alphabet then this resembles a random walk problem. Which would just mean that the chance of it happening would be 1 to something on the order of 2^8. Just plugging that in and assuming lineups are randomly selected there would be ~30 instances of that. But the probability isn't consistently 50/50 as you go to the next player in the lineup though, and managers typically use the same lineup multiple times, so if you lower the chances of getting a later name in the alphabet to like 33% and assume that managers use the same lineups 50 times you would expect it to have happened ~2 times!

I don't know how realistic any of these assumptions are, and the number times the same lineup gets used is probably very conservative, but I'm surprised that it isn't overwhelmingly likely. I wouldn't want to try to find an example.
 

Thanks for the input. Yes mathematically there's a very slim chance of this. I really thought someone would have done it on purpose at some point.

If no one can find documented evidence after a while, I should have a contest to see if anyone can find one by going through boxscores. If we think of an NL (or AL pre-73) team with an A leadoff guy and check games where a Y/Z was pitching, that would be a start.
 

There are some Early Wynn starts that are close. From when they had Aparicio at the top of the order. Check this one out.