Over the long weekend the World Science Fiction Society Business Meeting met several times. It was a marathon, much longer than is usual at Worldcon, because there was a lot to discuss. So much so, in fact, that I’m not going to attempt to cover it all in one post. This post, which is quite long enough, will be devoted solely to measures that affect how the Hugo Awards are conducted. The majority of these measures were intended, in some way or another, to combat the efforts of Pox Day and his Rabid Crypuppies.
Of course I was not actually at any of the meetings. However, video of the meetings has been posted online, and I am particularly indebted to Mx. Rachael Acks for their superb live-blogging of the meetings.
I want to start with the theory that the Puppy Problem will go away because they will get bored of pumping their money into voting when it is clear that they can’t actually game the final ballot and will those always end up getting thumped with a No Award. Pox may think that by putting people like Neil Gaiman on his slate that he can claim to have “won” because so many other people love Neil’s work too, but Neil has made it very clear that he dislikes the Puppy project just as much as anyone else.
An analysis on the Chaos Horizon website makes the claim that the Puppies are already running out of steam because Pox got 437 nominations for Editor: Long Form in the first round of voting, but only 165 first preference votes. They attribute this to large numbers of Puppies having participated in the nominating stage on the strength of the two-year eligibility rule for that part of the voting, but being unwilling to splash an additional $50 on a supporting membership of MidAmeriCon II to be able to vote on the final ballot.
This analysis is somewhat flawed because the voting methods are different. You get five equally weighted votes in the nominations, but have to rank them in the final ballot. There’s no guarantee that every Puppy voter will have put their Glorious Leader in first place when other Puppy Picks are available. Pox was eliminated fairly quickly in Editor: Long Form, so we can’t gauge the level of support he had in preferences. However, in Fanzine, where far fewer people cast ballots, the Castalia House Blog ended up with 298 votes in the final round of redistribution.
Still, 298 is a lot fewer than 437, and I’m not entirely surprised that some of the Puppy supporters have given up. Pox may be able to convince himself that he’s winning, but his fans will be a lot less invested in his infallibility. As the Puppies will have to pony up cash again to nominate next year, we may see rather fewer of them.
I want now to look at a couple of measures that were not specifically aimed at the Puppies, but which are relevant to the way in which the votes are counted and which will feed into the discussion later. Both are measures which received first passage at Sasquan last year and were ratified this year, so they will be in operation in Helsinki.
The first measure removes the requirement for finalists to have at least 5% of the vote. This is the rule that resulted in us having fewer than 5 finalists in Short Story for several years because the nominations were so diverse. One of the arguments for keeping it was that we may end up with lots of finalists because of a big tie for 5th place. That now may not happen, for reasons I shall explain later.
The other measure is the so-called Nominee Diversity motion, which is specifically aimed at the Dramatic Presentation: Short Form category. Basically it means that if there are lots of Doctor Who episodes on the final ballot, they must all be written by different people (or possibly not — see updates below). The salient point here is that WSFS members have got fed up of having a category for TV episodes in which almost every finalist comes from the same show (that’s definitely right). This too has relevance for later discussion.
The first set of specifically anti-Puppy measures to be discussed were those intended to limit participation in the voting pool. Currently you can participate in the nominating stage if you have a membership in the current Worldcon, the previous one, or the next one. There were motions to cut this down to just the current and previous years, and to the current year only. The former was passed (and requires ratification next year), the latter was not.
I’ve been a bit concerned about these motions because I don’t like the idea of us retreating into elitism as an anti-Puppy measure. We should be encouraging participation, not discouraging it. Having seen the debate, I think we did the right thing.
The right to nominate if you are a member of the following year’s Worldcon is a relatively new thing, and it is now opposed by experienced Hugo Administrators who cite problems with the system. There’s no doubt who is a member of the previous year’s Worldcon. No one is going to be joining it after it has happened. Next year’s convention, on the other hand, has a growing membership, and that leads to a requirement to coordinate closely with that event to ensure that everyone who has a right to vote can do so.
That would be OK if it meant that lots of new people were getting to vote, but most of the people who join a Worldcon more than a year in advance are WSFS regulars who join every year. We were told that including next year’s members resulted in the franchise being extended to very few additional people, and it wasn’t worth the effort.
Restricting nominating rights to just the current year is a different matter. From the debate that was very clearly an attempt to make it more expensive to vote. Also, as Lisa Hayes cogently pointed out, many people join Worldcon the current too late to participate in the nominations stage, or even the final ballot. The two year system gives those people a stake in the Hugos, which is a Good Thing.
Well done, Business Meeting, in both of those cases. We seem to have got it right.
Now we get to the serious changes in the voting mechanism. First up was a new proposal, Three Stage Voting (3SV), which I discussed before the convention. At the time most of the negative points I had seen made were about the additional load placed on the Hugo Administrators, which I addressed. At the Business Meeting three new groups of objections emerged.
The first was that the proposed calendar simply didn’t allow enough time for everyone to read and judge 15 semi-finalists in each category. That’s a fair point, and I’d like to see if any adjustments could be made to give people more time. However, it is also true that no one has enough time to read every eligible work before the nominations stage, and yet we still have it, so I’m not sure that this objection holds water.
Some people were also concerned that 3SV could be used by the Puppies to kick deserving finalists off the ballot. I’m not worried about this. The threshold to remove a potential finalist is quite high, and the Puppies have proved twice now that they just don’t have the numbers to game anything other than the nominations.
The final objection was that 3SV is purely a negative proposal designed to kick Puppy Picks off the final ballot. Effectively it is a proposal to move the process of No Awarding forward in time so that we end up with 5 finalists in each category, each of which people are happy do not need to be thumped with a No Award.
No one (expect possibly Pox and his cult followers) particularly likes how contentious the Hugos have become. But using No Award is a wholly negative action and we have been doing that a lot lately. 3SV doesn’t make the process more negative, it just moves the negativity to an earlier stage of the process.
At this point I should note that when I first wrote about 3SV I did not fully understand how it was proposed to work. I apologize profusely for this. Kevin has since put me right, but there was no time to go into the detail before the convention as a complicated amendment would have been required.
I thought that in stage 2 of 3SV we’d be allowed to up-vote works on the long lists as well as down-vote them. It seemed obvious to me that we would want to do that. However, the current proposal only allows for down-voting.
Allowing up-voting as well would have two benefits. Firstly it would allow people to change their votes based on items on the long lists that they had not heard about before. It is by no means unheard of for a work that got the fewest nominations to win the final ballot, because being on the final ballot exposes it to a lot more readers. Allowing up-voting in stage 2 would enhance that effect. In addition, if we had up-voting, the second stage would be much less of a solely Puppy-kicking exercise.
Kevin tells me that the proposers of 3SV decided not to allow up-voting because they feared that people would demand that the Voter Packet include all 15 items in each category on the long list, which would be a massively complex operation. I think that was foolish, for two reasons:
- People will demand to have these works in the Voter Packet anyway. Indeed, someone at the meeting did so.
- As I noted the other week, part of the value of 3SV is crowdsourcing eligibility checking. The items on the long list would not have been checked for eligibility so we have a solid excuse for not putting them all in the Voter Packet.
Talking of eligibility checking, Jonathan Cowie pointed out that the nominations statistics for this year show 83 nominations for “The Coode Street Podcast (Jonathan Strahan)” and separately 66 nominations for “Coode Street Podcast (Wolfe/Strahan)”. This is an obvious error. Thankfully, even if the two had been combined, the show would not have made the final ballot, but at least one mistake like that has been made in the past that did affect who got to be a finalist. We need extra eyes on eligibility checking.
Obviously allowing up-voting would be a greater change, and therefore can’t be tacked on during the ratification process next year, but I think it is worth doing. More to the point, I think that WSFS members will demand it, once they get to see all of the good stuff on the long lists that they wish they had nominated.
3SV was passed by the meeting, which then proceeded to discuss ratification of E Pluribus Hugo (EPH), the complicated statistical procedure intended to combat slate voting of the type practiced by the Puppies.
Up until recently, we had no idea what effect EPH would have. This year we had actual test data. There are two reports, one looking at the 2014 and 2015 Hugo Awards, and one looking at this year’s data (they also look at the Retro Hugos from 2014 and 2016, but those tests are not as useful because a lot fewer people vote). The reports are quite long, but the salient points are as follows:
- EPH will not get rid of all of the Puppy Picks on the final ballot. At best it will remove one or two.
- EPH will also kick off some non-Puppy works. Indeed, in 2014 it would have kicked off one of the eventual winners.
- The effect of EPH on the Dramatic Presentation categories is currently unknown because it was not tested on them.
The effect of EPH on this year’s ballot was quite encouraging. It didn’t get rid of all of the Puppy Picks, but it added enough good material to allow a contest in almost every category. That is, there were at least two finalists worth voting for in most of the categories. But not all of them. It made no difference to Fan Artist, and in Fanzine Lady Business again finished sixth. Given the validation of EPH, Black Gate may not have withdrawn.
The effect on previous years was less good, and in particular it is notable that EPH made five changes to the final ballot for 2014, a year in which the Puppies were not active. In particular it kicked Galen Dara out of Professional Artist, a category that she went on to win.
What is going on here? You may well ask. Surely the point of EPH is to defeat slate voting. It should not make any difference to the ballots if there were no slates in operation. But it does, because it detects what we might call “natural slates”. That is, if a whole bunch of people happen to nominate the same works, EPH will penalize that, even if no collusion took place. Algorithms have no political bias. They do their job in all cases.
One thing we can say about this is that it has proved the Puppies have a point. While I am satisfied that there was no SJW slate in operation in 2014, nevertheless EPH found that some categories did indeed suffer from “group think”, and it acted to produce a more diverse set of finalists. I like the idea that Joey Hi-Fi would have been a finalist in 2014.
So I have no objection to the detection of “natural slates”. Politically, however, I suspect it will be a minefield. If, next year, when EPH is used on the actual voting, people who are not on the Puppy slates get eliminated by it, I think that there will be an outcry. Fandom at large is expecting EPH to get rid of all of the Puppies, and no one else. It will not do either. People are not going to be happy.
Another potential issue here is the effect that EPH will have on Helsinki in particular. Finnish fans will presumably want to vote for Finnish works. Because there are a lot fewer Finnish writers than non-Finnish ones, there will be much less diversity in their nominations. I suspect that EPH will see the Finnish votes as a slate and kick some of the nominees off. That too will make some people unhappy, including me.
By far the biggest problem with EPH, however, is its effect on the Dramatic Presentation categories (and my thanks to Martin Easterbrook for alerting me to this). None of the tests presented to the Business Meeting included these categories. Apparently it was too time consuming to do this. So my first question is, what about the poor Hugo Admins for Helsinki who will have to run EPH on those categories? It sounds like they are in for a lot of work.
In addition I note that the Dramatic Presentations are two of the highest profile, and highest participation, categories in the Hugos. If there is a problem with them, there will be serious outrage. If I were being offered a new piece of software and the tests I was shown deliberately excluded two of the most important cases I would have been very worried indeed, and would have sent the people trying to sell me the software packing until they had done all of the tests I wanted.
So what will be the effect of EPH on the Dramatic Presentation categories? I quote from Rachael Acks’ live blog here: “Dramatic presentation will be the most changed category.”
Why? Because these are the categories that have the least diversity in nominees. To EPH, both Dramatic Presentation categories look like slates have been at work, and it will act to diversify the finalists.
To some extent it will be right. The Dramatic Presentation: Short From category has had slate voting of a sort for as long as it has existed. Fans of each series get together online to decide which episode(s) to nominate. No one has cared too much in the past, partly because it is only one category, partly because the slates are not run by arseholes determined to wreck the Awards, and partly because there are several competing slates each year.
However, as we have seen from the passage of the Nominee Diversity amendment, WSFS members have become fed up of Dramatic Presentation: Short From being dominated by Doctor Who every year, and they have acted to try to stop that. EPH will, I think assist with that process. Which is perhaps a good thing.
Except that the Doctor Who fans won’t like it. If they see some the episodes they nominated getting kicked off by EPH I suspect that complaints will be made. In fact, given the amount of dummy-spitting I have seen because of Jessica Jones winning the category this year, I suspect that very vociferous complaints will be made.
Something else that concerned me about EPH is what was said by the people doing the testing. Jameson Quinn was still in favor of EPH, but it is his baby. Dave McCarty was much more reticent. Dave was one of this year’s Hugo Administrators, he’s a past Worldcon chair, and like Kevin and I he is a member of the Hugo Awards Marketing Committee, the people charged with dealing with any bad publicity that the Hugos get. This is what Dave had to say in the testing report:
The changes to the Ballot and Long list are not easily verified and for people reviewing the detailed results at the end the only way to check that the process is working correctly would require access to secret nomination data and significant time. The difficulty in verification means that to check any result requires time which is NOT available to award administrators when it is time to close the nominating and prepare for the Ballot announcement. These are significant hurdles for a process that is generally designed to be open and democratic.
All of which makes me quite worried about the effect that EPH is going to have. Fortunately there is a clause in the motion implementing it that allows us to suspend it for a year at the Business Meeting. So if things go very badly pear-shaped in Helsinki we can keep voting to suspend it until the BM can vote to remove it again.
Having said all of that, EPH does have one unexpected benefit. If you have looked at the testing reports you will see that, whereas the traditional nomination system always gives integral numbers of votes, EPH gives fractional numbers. That’s a result of the algorithm. This will, I think, make it much more difficult for ties to happen, which will remove one of the major worries over the effect of dropping the 5% rule.
Finally, on Puppy issues, the Business Meeting ratified the 4 and 6 proposal. Or rather they didn’t, because they voted to change it to 5 and 6 and ratified that. Yes, they watered it down, making it less effective as an anti-Puppy measure.
Why would they do that? Well partly because they felt it was safe to do so. They had already passed EPH, so they felt the Puppy threat was being dealt with. 5 and 6 is a very different thing to 4 and 6. It is basically a proposal to increase the number of finalists on the ballot. It is a “share the love more widely” exercise. Business Meeting attendees are a bunch of softies at heart, and they tend to like any proposal that means giving out more prizes. They also dislike having anything taken away from them. 4 and 6 would have reduced the number of works they can nominate in each category, whereas 5 and 6 does not.
Overall I think we have made progress this weekend. I am fairly confident that next year’s final ballot will have at least one, hopefully two, deserving winners in each category, even if Pox and his cult do decide to try to piss in our beer again. But I also expect a fair amount of controversy over the fact that EPH does not do what people expected it to do. It will not get rid of all of the Puppy Picks, and it will kick off some non-Puppy finalists.
It’s the Hugo Awards. There will be drama.
Updates: A couple of small but significant changes here. I mis-spelled Jameson Quinn’s name, for which profuse apologies. Also supporting membership of Mac II was $50 (though it is down to $40 for Helsinki and San José).
The other update is a bit more complicated and pending. The Nominee Diversity motion isn’t as clear as I, with my editor hat on, would like. I read it one way. Kevin, with his knowledge of the intent of the proposers, says it means something different. I need to talk to him more about that. What I did miss is that you can have two similar works on the ballot, not just one.
Normally I get Kevin to proof this stuff, but he’s driving home from Worldcon and so has no time to look at anything in detail.
I’m a sponsor of 4 and 6, and the person who moved to amend it to 5 and 6. It’s not quite right to say that the meeting voted to make it less effective as an anti-Puppy measure: in combination with EPH 5 and 6 is actually a more effective anti-Puppy measure than 4 and 6. Going from 5 nominations to 4 wouldn’t hurt the Puppies at all, since we’d expect to knock one of their nominations off anyway, but it would weaken the opposition to slates a bit. The argument I made to the meeting was that 5 and 6 was better than both 4 and 6 and 5 and 5 in several ways, *including* fighting slates.
Well I’ll take your word for that, but like so much about EPH it isn’t immediately obvious why that should be so.
Also, if we do end up suspending EPH, we only have 5 and 6 to fall back on, not 4 and 6, right? Because that would take two years to bring in.
A few comments.
The reason 5 and 6 works better than 4 and 6 when combined with EPH is, as I understand it as a layman, that EPH is a more data-intensive approach than FPTP (which was in use previously). More data makes EPH perform better. (This is perhaps best understood when comparing 1 (nominee) and X (finalist) versus 2 and X; for 1 and 1 EPH will devolve back into FPTP.)
I very much agree about the problem with that the DP categories weren’t included in the EPH runs, and I haven’t been able to work out a satisfactory reason why. Maybe skip for prior cons, but this year?
I doubt the Doctor Who (or, say, Game of Thrones) contingent would cry havoc over not dominating the ballot. As long as one or two episodes of those shows end up on the ballot, I think they will be satisfied. It’s more about the show than the exact episode. Heck, I even doubt there has been collusion about the voting among the various fandoms, rather than the usual spirited and ongoing discussion about favourite episodes.
Last, I’m not that concerned about changes in the finalist list. To me (being Swedish), FPTP is a fundamentally unjust and unfair system. EPH is a marked improvement towards proportionality. The trouble is rather that people have been taught that FPTP is fair, when it’s not. The last few years I hope have been a wake-up call about that. What matters isn’t the exact work that ends up on the ballot but that they are all perceived as worthy finalists and winners.
Excellent explanation of why 5 is better than 4. Now I get it.
Also I agree totally about FPTP. I get very tired of people campaigning to use FPTP for the final ballot because they think it is “easier” or “traditional” or even “the way America does things”.
As to the rest, you have doubts but I have facts. Media fans have always been aware of the FPTP nature of nominations, and therefore the need to agree on which episode(s) to nominate. It even happened with Babylon 5 fandom back before we had the split.
For a long time there was also a lot of foolishness about the need to have only one episode on the final ballot to avoid “splitting the vote”. JMS famously could not get his head around how the final ballot worked. I think we have mostly got over that since Doctor Who has been so dominant, but I still see it surface occasionally.
I think the short time period on 3SV is a good thing. It’s not supposed to be a vote on the merits of the work. It’s supposed to be a way to get rid of obvious trolls and griefers. Things like “SJWs Always Lie”, and, of course, VD himself, who will presumably never appear on the ballot again once 3SV takes effect.
In fact, that might be the first critical test of 3SV. If VD manages to end up on the Best Editor list after it goes into effect, I think we can, quite simply, say that 3SV has failed.
If 3SV can’t get rid of Pox that means he has the support of at least 40% of the electorate, which would be rather worrying.
I take your point about Puppy-swatting being the main purpose of 3SV. I just think that once fandom has 3SV it will want to use that second stage for other things.
The idea of 3SV with upvoting has actually been discussed quite extensively: Hampus Eckerman had a proposal along those lines, which he later withdrew. People demanding items in the voter packet is certainly not the only reason it was dropped.
Here’s why I think it wouldn’t work: you would have to read the things you upvoted. You wouldn’t have to read everything (though no doubt some people would feel they ought to), but you would have to read what you voted for. And you would have a very short time in which to do so.
When people are confronted with this, they generally say ‘wouldn’t you have read a lot of them already?’. But I don’t think you necessarily would. A lot of people don’t read short fiction in the ordinary course of life, don’t read fanzines or semiprozines, don’t listen to fancasts, don’t follow the work of artists, and, while they do reads novels, do so quite a long time after they first appear. And if we want to expand the Hugo votership, we need to draw in more such people. When these people want to vote for such things for awards, they have to deliberately go in search of them. If the things that came up on the longlist didn’t agree with the things they had already searched out, a lot of them would be new, and they wouldn’t have time to read them.
If we had infinite time, I agree upvoting is a much better idea. (With downvoting, a work that is not downvoted keeps the advantage it entered this round with, so slated works that are not horrible enough to downvote are still more likely to reach the final round; with upvoting this isn’t so.) But I don’t think we can do it in the time allowed.
Time is certainly an issue. I’ll give you that. We all wish we had more of the stuff.
However, what you are saying is that no one should be allowed to upvote because they may not have the time to read new stuff. I’m saying that some people (by no means all) will read new stuff and will want to upvote it. I don’t see why they shouldn’t be allowed to.
I suspect you will also get people arguing that if you have to read something to upvote it then you must also read something to downvote it. And if you don’t have time to read to upvote, then you don’t have time to read to downvote either.
Of course I know the Puppy stuff is fairly obviously horrible, but it is hard to be consistent about allowing voting one way and not the other.
Ah – I think that’s a different proposal from the one I had seen discussed. On the one I knew of, there was a longlist of 15, you got the chance to upvote up to five, and the ones that got the most upvotes reached the ballot. Ordering in the initial nomination round ceased to signify, except in so far as it determined what was on the longlist (which is a good weapon against slates). So if you didn’t upvote, you had no influence on the final ballot at all.
Whereas on your proposal you can either downvote or upvote – is that right? If a thing gets a certain number of downvotes it’s removed from the ballot, supposing it was on it (which the voters at this point won’t know). So what do upvotes do? Is the idea that if a thing gets enough upvotes it’s added to the ballot, supposing it wasn’t on it? Does the ballot then grow if enough things are upvoted?
I think the underlying idea was that the vote was on whether something should be allowed to be on the ballot, rather than just on whether it is good enough – it’s a way of dealing with abuse. And there doesn’t seem to be a converse of ‘so offensive it shouldn’t be allowed to be on the ballot’, so upvotes aren’t needed in the same way.
I don’t have a specific proposal as yet, but here are a couple of things I think will happen.
First up, consider a voter who knows nothing about fan art, and so leaves the entire category blank. On seeing the top 15 she Googles the nominees and finds a few she really likes. She wishes she could have nominated them.
At the other extreme, suppose our voter is a big reader of short fiction. She agonized for ages over which five stories to nominate. When she sees the top 15 she finds that only 3 of her picks have made the cut, but there are several others on the list that she very nearly chose. She wishes she could change her nominations.
How we handle that, I don’t know, but I think it will happen, and I think there will be demand for change. I also think it is a positive thing that will offset some of the bad vibes of 3SV being seen as just Puppy-bashing.
If I understand Nominee Diversity clearly, it would prohibit a situation such as BDP-Short Form in 2012 or 2013, when “Doctor Who” had three episodes on the ballot each year, and would restrict the series to two finalists, regardless of who wrote those episodes. So if it had been in effect in 1968, it would also have prohibited “Star Trek” from having more than two episodes on the ballot as it did that year, notwithstanding the fact that the five “Star Trek” finalists that year had five different writers.
While “Doctor Who” could potentially be affected by Nominee Diversity, I don’t think it was the primary target in the minds of all the people who supported that amendment. Some of them may have been motivated by John C. Wright’s receiving 3 of the 5 finalist slots last year for Best Novella (Nominee Diversity would have limited him to 2 finalists in that category had it been in effect).
Yeah, that’s how Kevin is explaining it to me. I’m just not sure that the wording says that unambiguously.
Good point about Mr. Wrong.
First: my name’s “Jameson”, “e” not “i”, like the whiskey. Thanks for noting that you misspelled it, but, heh heh, it doesn’t actually help if you misspell it in the correction too.
Second: I absolutely support an option to upvote during the 3SV period. I think you should only be able to upvote 1, or possibly 2, works, so that people don’t feel obligated to read the whole longlist in the limited time available. The ONLY reason I didn’t make that proposal this year is that I could see that the BM agenda was already bursting at the seams. I think this proposal should be made, and ideally passed, in Helsinki.
Third: on the BDP categories, I think that Dave made a mistake by not including them. In my paper, I left out BDP:Short because it involved a lot of work including a few judgment calls which I didn’t want to undercut the admins on. But he is an admin so he didn’t have that concern. I didn’t want to seem ungrateful in the meeting, and indeed I still am very grateful for all the work he did for the Hugos in general and for analyzing EPH in particular; but I think he got this wrong. I’ve said more about this on Rachael Acks’s blog and on File 770.
Aaargh!
Your name is mis-spelled in the testing report. Fixed here now.
I don’t think that limiting the number of upvotes will preclude people from feeling obligated to read the whole list. After all, loads of people feel obligated to read every novel published before they can nominate. We should perhaps limit it so that people nominate/upvote no more than 5 works per category. It is an open question as to whether they should be allowed to effectively change nominations, but I’d be OK with that.
And yes, your reason for not proposing upvoting this year is pretty much the same as mine. Let’s do it next time.
Jameson, can you clarify what in your view the effect of upvoting would be? (See my previous comment for a guess, but I may be quite wrong.) I’m finding it hard to assess the proposal without knowing that.
A point that few people mention is that if we stick with EPH, we can eliminate the restriction on just nominating 5 works. We can allow people to nominate as many things as they like. That’s a better way to satisfy the folks who agonize over which five good short stories were really the best.
Another point is that 3SV won’t work without EPH. Without EPH, Vox will simply produce three slates, tell people to vote based on the last non-zero digit of their phone numbers, and he’ll sweep all of the top 15 in most categories. With EPH, he’d need over a thousand followers to get more than 10 in almost any category.
Well we are going to do six anyway next year, but the problem with nominations is often that people only list one or two items in each category. Allowing more won’t help with that. But having a third stage will encourage them to vote for more things because they are not faced with a blank slate.
Your comment about 3SV is based on an assumption about the number of slaves Pox has at his command, and that there is no change in the behavior of other voters. Also I suspect there’s more thinking that can be done about exactly how it will work. Having said that, I’m fairly happy to have EPH help control slates. My primary objection to it is that I think there will be outrage among fandom when people see it removing what they think of as “good” finalists while not removing all of what they see as “bad” finalists.