Just a reminder again, the word variance means UP and DOWN.
As I mentioned, the problem is ignoring that fact and assuming when you go UP, you can just blow it all and then complain about it when it goes down
(which is the main faulty argument for PPS)

Again, as I said, with small pools, it can become unmanageable, but with even a few 100 PH it's really not that bad.
300 PH is currently a bit under 2 weeks per block - which can sometimes be 10 weeks or 2 days.
I have to pay 6 months of power where I mine, so not really sure how that could be considered a problem.

To use some math to back that up,
(which is really where the problem lies, even the big pools seem to have no idea about the statistics of bitcoin, let alone miners):

What's the chance of getting only 50% or worse blocks over 6 months?
Well assuming it's an average of 14 days per block, that's roughly 12 blocks.
So CDF-Erlang of 0.5 for 12 events is 2%

How about getting 80% or worse blocks over 6 months?
The CDF-Erlang of 0.8 for 12 events is 25%

Now 300PH isn't a tiny miner - yes in todays terms there are a lot of multi EH miners, but someone with that size (300) setup,
if they are already paying 80% of their BTC to power, then they really need to consider what they are doing ...

But again this is all due to the fact that miners really have no idea what they are doing,
if a small pool just had a few 300 PH miners e.g. 900 PH, those numbers suddenly become smaller, and in the bad case example - 50% - ever so unlikely:

So CDF-Erlang of 0.5 for 36 events is 0.01%
CDF-Erlang of 0.8 for 36 events is 11%

But again remember that variance is both UP and DOWN, so incorrectly assuming it's only down, is again, the basic PPS (faulty) argument.