Three Stage Voting and the Hugos
The Back Story
(see also Nicholas Whyte’s post here http://nwhyte.livejournal.com/2855654.html )
The procedural response to the Rabid Puppy and Sad Puppy campaigns has consisted of use of existing rules and the proposal of new rules. I’ll pick on three to start with.
- To vote you need to be a supporting member of Worldcon and to have paid the associated cost.
- No Award – the existing procedure that allows voters to pick ‘none of the above’ from a set of finalists. This was used by voters to deliver a very forthright rejection of the slates in 2015.
- 5-6 a rule change as a consequence of the Puppy campaigns – people nominate five works but six works become finalists. This change was voted on in 2015, ratified and came into force for 2017. It limits the impact of slate voting on the list of finalists. If it had been in place in 2015 and everything else had been the same, then in more categories there would have been at least one more non-slate finalists.
The first process limits the extent to which the Hugos are vulnerable to click-farms or other coordinated attacks by large numbers of spurious voters.
Together, those second two processes amount to damage control. No Award stops unpopular finalists winning by being the only option because of a slate. 5-6 makes it likely that there is at least one non-slate finalist that could win, instead of the whole category going to No Award. Neither, in themselves, would stop a repeat of the Sad Puppies 3 / Rabid Puppies 1 combo but they strongly disincentivise the use of slates as a way of trying to control the award.
The next two are more complex:
- EPH – a voting system for the nominations that limits the impact that one disciplined minority can have on the nomination process. By ‘disciplined minority’ I mean a large group of voters who vote in a similar way. That could be people voting on a slate or it could just be an natural grouping of people with very similar tastes. It doesn’t stop a slate from getting works onto the ballot but it helps other works make it as well.
- EPH+ – a tweeking of EPH to improve the extent to which slate voting has limited impact.
My feelings about EPH have been that it is a good thing in itself, regardless of the slates. More generally it further disincentivises slates. Its down sides include some mathematical complexity which leads to some lack of transparency with the counting process.
The question in 2016 was whether together these would be enough. With the Rabid Puppies still impacting the 2016 ballot and analysis suggesting EPH would not be completely sufficient to prevent impact from a well disciplined slate, there was concern that other measures were needed. More pressing was the tactical change from the Rabids – their nominees came in three kinds:
- Hostages – plausible, popular works nominated to prevent a blanket ‘No Award’ vote against slates hard.
- Promotional – Castalia House works nominated as a publicity stunt by the publisher and chief Rabid Puppy.
- Trolling/griefing works – works intended to mock the Hugos or defame individuals or otherwise nominated in the hope of causing consternations.
Those three categories were overlapping.
The voters rose to the occasion again. The hostage plan didn’t work and at least one of the trolling attempts backfired when the inimitable Chuck Tingle was adopted as popular hero. However, despite the failure of Rabid Puppies 2 to make a lasting impact, it still suceeded in being a pain in the arse for everybody. In other words the griefing element of the campaign was a reward for the Rabids.
What to do? There was much discussion but a consensus arose around a proposal called Three Stage Voting or 3SV. In short, adding an additional round of voting after the initial nomination period. This round would allow votes to remove works not regarded by voters as legitimate possible finalists.
The Current State of Affairs
I believe it is fair to say that on the whole Sad Puppies were motivated by a genuine desire to gain some award recognition. Vox Day and the Rabids were more motivated by general mischief. However, both were motivated by the promotional/publicity-stunt elements of getting nominated for the Hugos.
Larry Correia’s capacity to motivate a significant number of fans to buy supporting memberships and nominate led to there being a significant Puppy voting block. Correia’s withdrawal from the Hugos has led to a decline in that block. More overtly Sad Puppies such as Sarah Hoyt have overtly stated that spending money on such memberships is a waste and essentially giving money to people the pups dislike.
In short, the immediate cause of the Puppy Kerfuffle has gone. The Sads have ceased to be a factor. However, that doesn’t mean others may not try similar antics in the future.
Of the rules above, I think the top two effectively stopped the Sad Pups – it’s just the effect was not immediate nor was it obvious to the organisers of Sad Puppies 1,2 & 3 what would occur. The effort of a slate, combined with membership costs, combined with the likelihood of a humiliating loss to No Award was enough to make the whole Sad Puppies campaign as a slate an unattractive option. 5-6 and EPH sealed the deal.
Longer term 5-6 and EPH(+) also mean that if another naive slate campaign arises from some other quarter, the damage done while that campaign follows the same cycle will be much less.
That’s all great as far as it goes. The 2017 finalists were rich and varied and good in all categories but…
The griefers are still there and their motives are not the same. The Rabid Puppies managed to get some griefing-style and promotional style nominees on the ballot. That impact was diluted by quality works but it was still present. The griefers don’t need to win to feel that their behaviour has been rewarded. Their goal is not a Hugo nomination but to create ill-will and to make people deal with their crap.
So 3SV is the Answer?
As things stand the only option on offer to deal with the griefing element is 3SV. There is no viable alternative that I’m aware of. A strong admin role during the nomination phase could theoretically remove griefing nominees but that idea is a non-starter: Worldcon/Hugo admins do not want that power and the consensus I’ve seen from Hugo voters is that such a proposal would never pass.
3SV is more acceptable because it passes that strong-admin style decision to the voters. In doing so, it potentially deals with other issues such as eligibility questions.
But, does it actually do the job?
It certainly would be another check against slate voting but I really think that is a solved problem. The issue is does it deal with griefing?
Nominally it does, in so far as it can prevent offensive works becoming finalists. However, does that solve the problem? Put another way which of these is the actual problem with the griefers:
- Works nominated by griefers becoming finalists?
- Hugo voters having to deal with works nominated by griefers?
3SV is a barrier to 1. but by its nature it makes 2. more feasible.
3SV will give voters an opportunity to reject specific works/nominees from the top 15. They won’t know how many nominations those nominees will have got or their ordering.
Now to get in top 15 for Best Novel in 2017 took about 200+ votes. A not insubstantial barrier but obviously significantly less than it takes to get an actual nomination. In less popular categories the number needed to get in the top 15 is substantially less – in many it is less than 100 and in some around 30 to 40.
So what, you might think, these nominees won’t be finalists and get no bragging rights or special status. I think that is missing the point. Works with abusive or insulting titles (for example) could be more easily gamed into the top 15 and by doing so griefers will get their psychological reward of Hugo voters acting to vote those works down. This may seem like an extraordinarily petty motive but the existing motives for the Rabid to spend more effort getting finalists on to the ballot are no less petty.
Appearing on the initial list will get more attention than appearing on the post-ceremony top 15 list. Although the two lists will contain the same works, the initial list is a list of potential winners in a way that the post-ceremony list isn’t. For the purpose of Castalia House’s publicity stunt motives, it may be more than sufficient to encourage nominations. For the same cost as getting one nominee as a finalist, the griefers can get multiple nominees on the initial list.
But ‘can’ is not the same as ‘will’ and its heart this is a psychological game rather than a procedural exercise. Are people like Vox Day or other people who might act in bad faith more motivated by the thought of gaining a single finalist and less motivated by a system that would deny them that? In which case 3SV could be a success. However, I think that they are primarily motivated by a desire to cause upset and dissension – in which case I fear that 3SV would simply a way of lowering the bar for their antics at precisely the time when they may have fewer resources to engage in them (where ‘resources’ are voting members of Worldcon).
Yet maybe that whole line of reasoning is a mistake. Acting in any diretion solely on the basis of the likes and dislikes and strange motivations of trollish-ideologues is, arguably, an error itself. A better approach is to place your own interests first and make those paramount. I think EPH is an example of that because I think it is a good thing in itself – yes, the Puppy antics gave the impetous to introduce it but it is defensible as a thing in its own right. For those who DON’T think EPH is defensible in its own right, the argument still holds – you’d be right to oppose it regardless of whether it was an effective vaccine against hydrophobia.
So is 3SV a good thing, in a circumstance in which there are no griefers? It does possibly pass questions of eligibility of borderline SFF works to the membership but, I’m not sure it will be an effective tool in that regard. Otherwise – does it do anything useful? I’m genuinely open to suggestions there – I’m not sure it does but that maybe just that I haven’t thought it through entirely.