Slightly less cynically perhaps (and those who know me will be crying alien impostor as they real this!) the view has long been that it's not possible for patent examiners to be subject matter experts in all areas of technology. As such they're quite unqualified to judge patent applications on the basis of intrinsic merit and thus fall back to the "does it look like something already on file" part of their job. After that, the validity or not of the patent will be tested in the courts. (Why yes Gloria, these people are all just lawyers... aaahh, there's that cynicism!)
So to create a trollable patent all you have to do is pick some arcane corner of technology, say computers or electronics, find something so obvious no one has bothered to write it or anything like it down before and create a patent for it. The canonical example for me is sadly the "XOR pixels to create a cursor so you can XOR again to remove it and repaint it somewhere else (often one pixel over)" patent which at one time or another every graphical display and window system vendor has been trolled with. But only if you think XOR is a fabulous new invention and not completely obvious and with significant prior art. Or maybe this... https://www.nytimes.com/2002/05/13/business/patents-patent-office-faces-huge-backlogs-extremely-technical-inventions-absurd.html (patent lawyer teaches son how to make a patent with a playground swing as an example, patent is granted).
The 'fix' is to find some non-lengthy-court-battle way for patents to be reviewed. Maybe to consult experts or academics, or to have a year window after the patent is granted for the public (i.e. any interested party) to offer comment that might lead to review. The current system has clearly become too unwieldy to survive much longer as it drives small companies to settle or cease trading and fills the courts with unnecessary litigation from organisations big enough to defend themselves.