Maybe it would help if upvotes also increased the amount negative an entry can be before getting axed.

That is to say that an entry with 0 upvotes gets dropped with 5 downvotes at -5, but an entry with 10 upvotes might get dropped by 20 downvotes at -10.

That way, generally bad entries which aren’t getting upvotes still get dropped rather quickly, but ‘hit or miss’ entries that are getting a lot of upvotes and downvotes stick around longer, and get a better chance to get a surge of upvotes, rather than getting dropped when random timing gives them five more downvotes than upvotes.

The simple math could be that a submission gets dropped at -5 or -upvotes, whichever is lower, but there’s any number of possible formula to work with, depending on how resilient entries should be.

If you wanted to be really complicated, the point at which submissions get dropped could slowly rise over the course of the day, so they could drop at -100 at the start of voting, and drop at -5 near the end, giving plenty of time for decent entries to make a comeback, but letting bad entries still get progressively dropped as the day progresses, so people who didn’t want to wade through crap could just vote later in the day.

All I know is that I came to vote and I saw four total entries. Maybe a lot of them weren’t good, but four out of over three hundred?