Instead of taking a list of the Top 15 nominees (semi-finalists) and saying "Vote No on any of these nominees you think do not deserve to be a Hugo Award Finalist," this system takes the Top 15 (ordered in such a way as to not give away how many first-round nominations each got) and says, "Choose up to five of these nominees. You may write in one additional nominee, which counts as one of your five choices." The five nominees from the Semi-Final round (potentially including write-ins) are the Hugo Award Finalists.
(Again, "Top 15" could include ties, or if you prefer, any nominees receiving at least N% of the first-round nominations. I'm not wedded to Top 15 other than it's an existing classification we've used for years to generate the post-ceremony Long List.)
As with 3SV, we don't contact the semi-finalists in advance here. They can decline nomination after they've been announced if they so choose. Any semi-finalist who withdraws or is found to be ineligible will not appear on the Final Ballot regardless of how many votes it gets. Announcements of ineligible works would probably be made in public. Withdrawals are up to the decision of the individual, but would be included in the post-ceremony statistics.
Let's look at how the Semi-Final Ballot would have looked last year. Try to clear your mind of what actually appeared on the final ballot and the post-Ceremony Long list. I'm ordering it in reverse alphabetical order by title in order to try and confuse the actual number of nominations each work received. Here is your list of 15 Semi-Finalists. Vote for up to five (including not more than one write-in). Do not rank your choices; place an X by any candidate you would like to see on the Final Hugo Award Ballot.
__ Words of Radiance (Brandon Sanderson)
__ Trial By Fire (Charles E. Gannon)
__ The Three Body Problem (Liu Cixin)
__ Skin Game (Jim Butcher)
__ My Real Children (Jo Walton)
__ Monster Hunter Nemesis (Larry Correia)
__ The Mirror Empire (Kameron Hurley)
__ The Martian (Andy Weir)
__ Lock In (John Scalzi)
__ Lines of Departure (Marko Kloos)
__ The Goblin Emperor (Katherine Addison)
__ The Dark Between the Stars (Kevin J. Anderson)
__ City of Stairs (Robert Jackson Bennett)
__ The Chaplain’s War (Brad Torgersen)
__ Ancillary Sword (Ann Leckie)
__ ____________________________________ (Write-In)
During the Semi-Final voting, Larry Correia withdraws Monster Hunter Nemesis and Marko Kloos withdraws Lines of Departure. These positions are not back-filled. Whether the public knows about the withdrawal is up to the individuals.
Only members of the current Worldcon may vote on the Semi-Final Ballot. The five works with the most votes in this round (including any write-ins) are your Final Ballot, which is voted upon by the same Instant Runoff Voting system we've used for 40+ years.
What this system does do is to concentrate members' attention upon a smaller list of potential finalists. I think most people are more comfortable picking from a list than having to make free-form nominations, which is why there usually are more final-ballot votes cast than nominating ballots.
There's a single write-in space primarily to allow people to let off steam. I personally doubt too many works would ever make it to the finals through this write-in space. I would not want more than one write-in space because it defeats the purpose of concentrating the members' attention on the most likely finalists obtained from the initial nominations.
Could griefers still get all five of their favorites onto the final ballot? Maybe, but probably only if the remaining 80% (by most estimates) of the members ignore this round. Given that voting on this round is easier than the existing first round of nominations, I think that unlikely.
Should we expect people to have read all fifteen semi-finalists? No. This is still a nominations round. You're still making up to five equally-weighted nominations. If you've read and liked any of these Top 15, endorse them with a nomination. If you still have a work that is your top favorite, write it in; it might still make it, although I doubt it unless there's an amazing groundswell of interest in it.
This "Double Nomination" voting system means that the members vote for things, not against them as the 3SV system does. It also removes the need for a complex quorum calculation to prevent griefers from gaming the second round. As long as the griefers continue to be only about 20% of the electorate, it seems unlikely to me that they can dominate the field at this point. If they somehow grow to an actual majority, then they win, and they deserve to do so. The Hugo Awards are not actually about minority representation, despite what some people have falsely claimed. They are designed to reflect the wishes of a majority of the members who participate.
Again, I didn't come up with this. I take no credit for it. But if someone wants to propose it to the WSFS Business Meeting, I will help you draft it into the correct technical form if you want me to do so.
ETA, May 5: There is a suggestion from David Goldfarb on Making Light that instead of "vote for 5," this would could use approval voting, in which you could vote affirmatively for any or all of the semi-finalists. (And presumably also for your one write-in.) The five that got the most votes (not marking a candidate is the same as voting No) would go onto the shortlist. This idea has merit as well, I think.