A coin flip would be so much easier.
There’s very little in the way of policy. It’s up to each team lead / system owner what they feel should be backlogged. Basically - it all comes down to, what do you think you will reasonably get to and what do you think is a priority. If it’s not - then backlog it. We have finite resources and can’t fix everything. So things like crashes we can reproduce, regressions, problems with new systems are examples of things we try to fix ASAP. Problems with older systems, that have been there forever that have known workarounds, or would just be really really hard to fix right now without a lot more resources, tend to get backlogged until an opportune time presents itself to tackle that issue.
Which for me is where I take into account votes for items, e.g. Unreal Engine Issues and Bug Tracker (UE-4659), was backlogged as of about a week ago. It had been backlogged for probably about 1.5 years Why? Well - we knew that the render transform stuff was going to have limitations with the clipping system in Slate, and we knew that fixing that meant completely redoing Slates clipping system and changing the default behavior. This would of course break all existing customized uses of Slates clipping system, so big nasty hairy problem. There were more important things to tackle at the time with UMG and Slate, so that issue was backlogged.
I removed it from the backlog about a week ago, because I finally had a some time to work on redoing Slates clipping system, and had some good ideas on how to do it. Still, as expected it’s a really difficult problem - I have over 200 files checked out right now to make this change,.
We simply can’t fix everything, so it comes down a judgement call. Additionally if we don’t touch a bug within 6 months, a script automatically backlogs it, unless someone touches the bug, and resets the clock.