What? I don't see what's wrong with that. 'PLL' is a valid description of that step. It's like somebody saying 'I'm reading this book' and you saying 'No that's wrong, you're reading a novel'.
Unless I've the wrong end of the stick?
Here's how to make pseudo-COLL + EPLL skip non-lucky: Solve at least one edge along with the corners. I can already think of a couple of ways to achieve this right now, none of them too efficient, but it's feasible anyway.
If you haven't already figured out, with this approach, EPLL skip occurs 33.3% of the time (>20%).
What? I don't see what's wrong with that. 'PLL' is a valid description of that step. It's like somebody saying 'I'm reading this book' and you saying 'No that's wrong, you're reading a novel'.
Unless I've the wrong end of the stick?
If you mean permutation of the last layer in the most literal sense, then yes, it is a valid description of that step. But when most, if not all cubers talk about PLL, they are referring to the step with 21 different cases, not 4.
Simply put, EPLL skips occur 1/12 of the time, PLL skips occur 1/72 of the time, hence EPLL != PLL when we're talking about skips here.