I've crafted and played literally thousands of Lippy decks.
When Cerys got added I tried it a few times but it's not my taste of deck, not braindead but boring to play more than 2-3 times.
Lippy offers a lot of deck building types:
You can play around creating new cards/special synergies(Alzur, Master mirror, Idarran, elf and onion soup, Cossimo) to later reuse some of them.
You can use Lippy to use the combo of a combo deck to win R1, bleed R2 and reuse combo in R3(this is the Cerys deck, but you can play Iris + Dracoturtle, or other combos like Sukrus Arnaghad)
Another playstyle is to use all or most of the discard package and add many tech or control cards and use it to pointslam/card select, to have the answers at hand when they matter and discard them if they don't to maybe play them later.
None of those variants are truly good. In the long run you'll brick too much, your Lippy can be useless, you can draw poorly or topdeck horribly, encounter decks that will run over you with ease, and is so predictable by Roach, Cerys, etc... not to mention your plays are mostly in the same order always. You MUST use Cerys before discarding anything, the same with roach or worse even with Knickers, you must play a gold before you discard any card. You also need to play Cerys R1 to have an extra shieldmaiden in R3, in other Lippy decks you need to play your create cards in R1. It has to play almost always in a predictable order.
It also is pretty costly, you need Lippy (10prov) + Roach(9prov now) +/ Knickers (8p), Morkvaark(8p) also Oneiro(13p) is a Must, this makes any Lippy deck less powerfull overall, relying on dumping the full power of the deck each round or lose. You must control the round because you can't fit any more power or engines since your deck is weakened by the Lippy combo cards. So long R3 you will lose vs 90% of decks.
So all in all when Lippy decks work can seem unfair or unstoppable, but that is not 100% of the time. In fact, Lippy is a pretty mediocre(if fun) card, unique enough to make decks around that work often enough to win about 50% in the best case.