I suspect it's something like this: Company A owns some patents that it wants to troll with. They get entities B through Z to buy the patents and troll, with the contract saying that A gets a chunk of the winnings if they win. If they lose, B through Z are on the hook for expenses. So A gets to hide, first of all, and second gets insulated against financial losses, and third doesn't get tainted by a bunch of adverse legal rulings. (That is, think about how Prenda Law got tainted. If they came after you, you weren't going to settle, because you knew it was garbage. Well, in this setup, if B through F have been shown to be bogus, and G is suing you, you may settle rather than fight, because you don't realize it's all part of the same bogus collection of junk.)

As I said, that's what I suspect. The actual scam may be different.

Well presumably if there is a significant penalty for making a false patent assertion, then placing the assertion in the hands of a very small company would simply bankrupt that company in the case of the penalty coming due, and the real owner wouldn't be liable.

So essentially the scam removes liability from the (dubious) patent owner - it removes the downside to (allegedly) abusing the law, and subverts the intent of the legislation.

The hook for compelling discovery is probably standing. That is, if you’re not the “real” patent owner (for whatever definition of “real” the judge likes) then you might lack constitutional standing to bring suit for infringement.

The substantive issue the judge is interested in is whether they made false statements in their court filings. The attorneys seem to be misleading the court about the true owner of the patents, and who the attorneys actually represent.