I’m going to write for a bit about the history of the Hugo rules in the WSFS Constitution, and what I think they mean. But I should make it clear that these are not official rulings of any sort, which only get made by the Business Meeting presiding officer, and the Hugo Administrator. And the Hugo Administrator’s rulings are, in the end, the only ones that count, because the Constitution delegates the final decisions to the Worldcon committee, and the Worldcon committee, in every year since it was permitted to do so, delegates that power to the Hugo Subcommittee. And that delegation must be total; the convention committee doesn’t get to override the decisions by the Hugo Subcommittee.
Section 3.2.11: The Worldcon Committee is responsible for all matters concerning the Awards.
Section 3.12: Exclusions. No member of the current Worldcon Committee or any publications closely connected with a member of the Committee shall be eligible for an Award. However, should the Committee delegate all authority under this Article to a Subcommittee whose decisions are irrevocable by the Worldcon Committee, then this exclusion shall apply to members of the Subcommittee only.
I’m going to try to quote sections of the Constitution for reference. The current version is online at http://midamericon2.org/what-is-wsfs/ along with other rules, and the minutes of last year’s Business Meeting, and a set of videos taken of the meeting for those who want to follow it completely. But be aware that the meetings run for almost ten hours (I think it was probably the longest Business Meeting in history). And, for historical reference, I’d point people to the archive of past Business Meeting minutes at http://www.wsfs.org/bm/archive.html
I’ve made the assumption that if you’re reading about the Hugo Award rules, you’re probably generally familiar with the overall structure. But, for those who aren’t, here’s a quick summary. The Worldcon (an SF convention that’s been held annually since 1939, except that we skipped 1942-5 during WWII), annually, gives out an award called the Hugo Award for work in the field. The rules for governing the Worldcon are made the by the members of WSFS – but the membership of WSFS is anyone who has an attending or supporting membership in the Worldcon (a supporting membership gives you the right to vote for the Hugos, and the upcoming site, but not to attend). So the membership gets to set its own rules – but is limited by what I consider the most important part of the Constitution:
Section 1.6: Authority. Authority and responsibility for all matters concerning the Worldcon, except those reserved herein to WSFS, shall rest with the Worldcon Committee, which shall act in its own name and not in that of WSFS.
In general, the areas reserved to WSFS are the rules for administering the Hugo Award, selecting the site of the Worldcon two years from now, and holding a Business Meeting which can amend the rules. For the last four decades, amending the Constitution requires a motion to be passed by one year’s meeting, and ratified at the immediately following year’s meeting. Pretty much everything else is up to this year’s committee.
It’s a balance between three bodies, each with its own authority. The WSFS, through the Business Meeting’s power to amend the Constitution (which is binding on the Worldcon committee and the Hugo Subcommittee/Administrator) has the final say, but it’s only through a Constitutional amendment that it can authoritatively change the rules. Otherwise, it has no power over either of the other two bodies. And the Convention committee and the Hugo subcommittee each have their own distinct spheres. The Hugo subcommittee has total control of the balloting – determining what is eligible, counting the votes, etc. – but only by administering the rules in the Constitution, and using judgement only when it’s absolutely necessary, and deferring to the will of the voters and/or creators where it seems possible (“Is this 38,000 word story a Novella or a Novel?”, for example – which usually will end up being put in the category where it gets the most nominations). But it has no control over any of the other aspects of the Hugo Award – in particular, the ceremony is entirely under the control of the Convention committee, where the only responsibility of the Hugo subcommittee is to ensure that the rockets have plaques with the right names on them, and that the envelopes contain the correct nominees on the outside, and the right winners on the inside (and possibly that the tech staff have the right name on the Powerpoint slide with the winner on it, if there is such a presentation).
And, speaking personally now, as a past member of the Hugo subcommittee (most recently at Loncon 3, in 2014), and knowing many of the past administrators as friends, the administrators hate to make judgement calls, because they’re often tricky, and the other viewpoint is often very much supported by a reasonable interpretation of the rules. To pick an example (which, since it was three dozen years ago, is probably not one that will inspire a resurgence of the argument), the 1981 Hugo administrator put Warhoon 28 in the Best Related category. Warhoon 28 was the 28th issue of Warhoon, a long running fanzine, and one that won the 1962 Hugo Award for Best Fanzine, which would seem to mean that the fanzine, Warhoon, was being nominated. But Warhoon 28, unlike all of the other issues of that fanzine, was a special tribute issue to Walter Willis (one of the great fanwriters in history), and collected all of his major fanwriting in a 600+ mimeoed page hardbound book. So the administrator chose to put that issue/book in Best Related, rather than putting that fanzine in Best Fanzine. It was a perfectly reasonable decision – but clearly one where he could have reasonably interpreted the rules differently, and decided on Fanzine. It, like other judgement calls, is not something that an administrator wants to make, since very reasonable arguments can be made as to why the administrator got it wrong.
Now to get to the Hugo rules. In general, the Hugo-specific rules are in current Article 3, although general rules about counting votes, etc., can be found in Article 6, and some stuff scattered elsewhere. But Article 3 is the part that tells what the categories are, and what goes in which category.
For categories that are distinguished by length (Dramatic Presentation, and the fiction categories) the Administrators are given some flexibility in moving a work between categories, so long as the length is close to the boundary (generally, it’s a 20%/5000 word margin). So a 42,000 word story, which would normally go in Novel (the Novel/Novella boundary is 40,000 words) could be put into Novella by the Administrators (who will generally look at where the nominators tended to put it, although they’re not required to). But note that Neil Gaiman’s “Coraline”, which was published in book form, and looked like a children’s novel, was nominated (and won) in Novella, since it was less than 35,000 words, and therefore could not be moved.
The baseline for the current Constitution, and the current set of Hugo categories, was set in the 1976 Business Meeting (in Kansas City, in the same hotel as this upcoming Worldcon). We had ten categories: four fiction categories, differentiated only by length, dramatic presentation, professional editor (which was a recent change at the time, since the category had been professional magazine, but got changed to editor in recognition of the fact that there were fewer magazines, with new SF showing up for the first time in books rather than the magazines, and so we changed it so we could recognize book editors), professional artist, and three fan categories (fanzine, fan artist, fan writer – where fan artist and fan writer were defined as people writing/drawing for fanzines). And professional was simply defined by print run – a professional publication was one with a print run of 10,000 or more. Since the professional magazines, and professionally published books, all had those kinds of print runs, that rule gave the right answer.
As we started to see more non-fiction books about SF, and books of/about SF art, we added a new category, first awarded in 1980, for Best Related Non-Fiction Book. As part of the general trend to recognize electronic works, which don’t necessarily get published in the form of books, it’s now become the Best Related Work category, and so it no longer needs to be a book (or even an e-book). The key thing which distinguishes a work eligible in that category from other categories is in the end of 3.3.4, “… is either non-fiction or, if fictional, is noteworthy primarily for aspects other than the fictional text, and which is not eligible in any other category.”
When it became clear that, during the late 70s, we had three fanzines whose circulation was many thousands, while most fanzines were having circulations in the low hundreds (when you’re printing and mailing physical fanzines, and generally they were available for free, there were real limits on circulation, depending on people’s budgets), we split out semiprozines, just to get them out of the fanzine category. And we tweaked the rules somewhat, so that there were more contenders than just the three that we moved out of fanzine; if it were only that, then semiprozine wouldn’t be a viable category. We were starting to see the beginnings of small run fiction magazines, and serious academic small circulation magazines, and the semiprozine rules put those into the new category, so it was a category offering reasonable choices.
As electronic publication for books and magazines, and as blogs and other similar formats became the locations for more of the works in the field, it became clear that “print run” was a poor concept for distinguishing between professional and non-professional works. So we changed it to a pure financial rule:
3.2.10: A Professional Publication is one which meets at least one of the following two criteria:
(1) it provided at least a quarter the income of any one person or,
(2) was owned or published by any entity which provided at least a quarter the income of any of its staff and/or owner.
An important consideration is to note that professional/non-professional refers to a publication, and not a person. A person can be both a professional writer/editor/artist, and a fan writer/artist, or publishing a fanzine. You don’t stop being a fan just because you’re a professional. For example, Terry Carr (who won a Hugo Award for his fanzine in 1959, best fan writer in 1973, and best professional editor in 1985 and 87) was the Fan Guest of Honor at the Worldcon in 1986 (and was thrilled to get that honor – he was a fan until the day he died). If it’s being done for free, then it’s fan work, and the person is eligible in the fan categories. The clearest example was Jack Gaughan, who won both Professional Artist (for his work in the professional magazines) and Fan Artist (for the work he donated for free to fanzines) in the same year – 1967.
Starting with the 2003 awards, we split “Dramatic Presentation”, with the split being primarily by length (a running time of 90 minutes), but with the wording telling the Administrators that TV shows should generally go into BDP-Short, and movies into BDP-Long. The feeling was that TV shows (particularly Buffy, at the time, which was getting lots of nominations, but didn’t appear on the final ballot) were at a disadvantage in the nominating process, since their nominations would be split among lots of different episodes, whereas the movies wouldn’t split them. So that while a single TV show, if you added up all the episodes, would get more than enough nominations to make the ballot, but since you couldn’t add them up, the TV shows never appeared on the ballot.
And later we added Graphic Story, as those works became more significant in the field (and we thought that enough members were interested in them so as to be able to nominate on an informed basis). We also split Editor, based on editing novels and editing shorter works. And we added Fancast, as podcasts became a popular medium.
The recent results with Game of Thrones, and the Wheel of Time, bring up the question of the serialization rule:
3.2.4: Works appearing in a series are eligible as individual works, but the series as a whole is not eligible. However, a work appearing in a number of parts shall be eligible for the year of the final part.
This rule has been a part of the Constitution since the beginning. It comes from the fact that most SF used to be first published in magazines (books were much rarer), and novel-length works were too long to be published in a single issue (three and four part serializations of novels were quite common). So the question is, when is the work eligible, and the decision was that if a work was really a single serialized work, then it was eligible when the final part appeared. The problem is determining when a work really is a single work, and when it’s a series. And the policy that Administrators have generally followed is to ask the creators, “Is this one work, or is it a series.”, and go with whatever the creator decides. However, once the decision has been made, it’s irrevocable, so if a creator decides that a part of a work can be nominated, then it indicates that it’s a series, with the individual parts eligible, and not a serialized work, with the entire work eligible. So that Game of Thrones Season 1 was determined, by its creators, to be a single work. But, in later years, the creators decided that the full seasons were not a single work, and that therefore individual episodes, and not a season were eligible.
I’d also like to say a few things about the two-phase selection mechanism. Right now, people who are members (as of Jan 31 of the appropriate year) of the current Worldcon, or the immediately preceding or following one, are eligible to nominate. However, only the current year’s members are eligible to vote once the final ballot appears, and they only need to be members by the time they vote.
In the current set of rules, the finalists are determined by simply adding up the individual nominations from each member (each of whom can make up to five unranked nominations), and the five works (plus ties) receiving the highest number of nominations appear on the final ballot. This has the simple result that the works with the most people who want them on the final ballot are the ones that appear there.
Once the final ballot goes out, a different voting scheme is used to determine the final winner. The votes are ranked (unlike the unranked nominations), with a ranking from 1 (the individual voter’s highest choice) to wherever the voter stops caring about preferences among the remaining choices. The winner must win with a majority. If no candidate receives a majority on the first ballot, then the least popular choice (the one with the fewest first place votes) is eliminated, and its votes distributed to the second place choices on ballots that have them. And the process is repeated, as necessary, until a work receives a majority (or a tie, in which case multiple works receive the award). What this tends to mean, except when a choice is so overwhelmingly popular that redistribution is unnecessary, is that the choice that is least unpopular tends to win, since unpopular works tend not to get the second, third, etc. place votes redistributed to them, so they fall off, even if they got the most first place votes. If a choice is initially preferred by a majority, then no redistribution happens, so my general description is unnecessary. To pick a recent, uncontroversial choice, of the 1058 voters in the BDP-Short category for the 1939 Retros, the Orson Wells broadcast of War of the Worlds had 813 first place votes, and no redistribution was necessary.
The Retro results for 1939, with complete breakdowns, are at http://www.thehugoawards.org/content/pdf/1939HugoStatistics.pdf
And the formal explanation of how the elimination procedure is done can be found in
Section 6.4: Tallying of Votes. Votes shall first be tallied by the voter’s first choices. If no majority is then obtained, the candidate who places last in the initial tallying shall be eliminated and the ballots listing it as first choice shall be redistributed on the basis of those ballots’ second choices. This process shall be repeated until a majority-vote winner is obtained. If two or more candidates are tied for elimination during this process, the candidate that received fewer first-place votes shall be eliminated. If they are still tied, all the tied candidates shall be eliminated together.
And section 3.11.1 says that if you get a tie at the end, then everybody gets a Hugo Award.
One final note is that there are two amendments concerning the nominations process which received first passage last year at the Sasquan Business Meeting, and which will be up for ratification at MidAmericon. The simple one to explain is called “4 and 6”. Under the current rules, each person can make up to five unranked nominations, and the five (plus ties) works receiving the most nominations become the finalists. This proposal changes the maximum number of nominations a person can make to four, and has six slots on the final ballot. The discussion on that appears on pages 23-25, 78-82, and 128-129 of the minutes.
The more complicated one to explain is called “E Pluribus Hugo”, and I’m not even going to try to explain it, since a minimal explanation would double the size of this. The proposal, which appears on pages 25-35 of the minutes (with an explanation of how it works by the proponents), and the debate, on pages 35-36, 73-78, and 129-134 really needs to be read in order to understand it, and the arguments raised for and against it. Those minutes can be found at:
I realize that, by reading the minutes, people can determine my feelings about both proposals (and why I feel that way), since I participated in the debate on both. However, in this article, I’m trying to be a disinterested source of information about the rules, and not a partisan. So I’m not going to debate those in this forum – I’ve made my opinions and reasoning public already and the volume of discussion on those proposals can get very large.
I hope this helps bring a historical perspective of how the rules are structured, and why/how they are interpreted. I’ve also had my viewpoint about some of the other rules in some of the other threads here, particularly in the thread on Best Editor and Best Professional Artist, and I tried not to repeat them here, in the interests of keeping this reasonably shorter.
Thanks for offering this informative historical view.
RE: the distinctions between fanzine/ semiprozine/ professional publications
I’ve found the Semiprozine Directory to be a fantastic resource to figure out what’s what, as there are publications that seem to change eligibility year on year.
I read this post, and found it to be informative, detailed, and good motivation to ….do something else with my life….
That’s NOT a criticism of the piece or the author. It’s just that the frabjous fingdangle of rules and variances drive this ADD person to OH LOOK! A SQUIRREL!
So I wrote a review of Jeff Duntemann’s most recent release, “10 Gentle Opportunities.” Really, I did! It’s right here: https://www.amazon.com/review/R12664CVDDWTGO/ref=cm_cr_rdp_perm?ie=UTF8&ASIN=B01AQ1549E
I’d love to know if anyone ran the 2015 nominations through the “E Pluribus Hugo” process to see if it would have yielded the results its advocates want to see.
Last year’s Adminstrator had hoped to provide an anonymized version of the nominations so people could test out EPH on it. But his initial attempts weren’t sufficiently anonymized — and he hasn’t been able to generate it yet.
I expect that he might well be able to report things back to next year’s Business Meeting, presumably before the final ratification vote on EPH takes place.
I’d rather see a decent analysis matrix that shows that EPH and punishing the voters (4 of 6) would prevent or mitigate the black swan events that cause their proposal. Right now, all I’ve seen is attempts to duplicate previous years’ results and a stunning unwillingness to consider the edge cases that are the reason for their proposal.
Well done, sir!
I have to ask, is E Pluribus Hugo expressly designed to make it easy for Vox to No Award the Hugos next year as well as this year? Because if Vox gets 300 people to nominate only him in all the categories, then he will be on the final ballot it seems. With him on the final ballot, No Award gets broad support from the Vox haters, and high points from the 300 people who nominated him to start with. Hell, I doubt it would take 300, I bet 200 would do it no problem. Has anyone told theses people Vox is a game designer? Or it is the liberal = stupid issue?
EPH seems to be proof against true slate voting, i.e., Straw Puppies. (It would be completely irrelevant to Sad Puppies IV, though.)
What I don’t know whether it’s been simulated is whether you can generate individualized slates that are just different enough to evade the EPH filter and yet still swamp the nominations.
Lots of people have tried different simulations and there are a number of different software implementations of EPH. I made a ‘toy’ version of EPH using Excel (only formulas, no macros/VBA) and it pretty much works as stated. Really it is just another version of elimination voting (like AV aka ‘an Australian ballot’ as used in the final voting) but without having to put preferences on your ballot.
It doesn’t stop slates and if, say, somebody can organize enough people to all vote the same way then they can still sweep the board. The difference it makes is that ‘somebody’ needs more people for it to be ‘enough’.
Now notably it doesn’t actually matter if there is an organized slate or not. If there a whole bunch of nominations for a category that are very similar for any reason those votes will tend to have less impact on the final ballot than they would with the current system. So it should lessen the extent of any group of people with very similar taste dominating a category (e.g. think of Doctor Who fans and Best Dramatic Presentation Short) but increase the chance of that group getting ONE thing on the ballot. So, for example, if lots of people who really like Baen Books voted the chance of a category being nothing but Baen goes down but the chance of at least one Baen goes up. So regardless of last years events (and people’s opinions of them) EPH should tend to make the spread of finalists a bit more varied.
The thing I’m wondering about is the underlying assumptions in EPH with respect to slate voting. In particular, does it assume strict slate discipline — every item on the list gets the same number of votes? Does it assume no crossover with non-slate voters? To what extent will it fall to the Amazon effect — “people who read this also read/liked that”?
Maybe it’s the physicist in me, but I’d like to think someone would run the EPH algorithm on some real-world data to see whether or not it blows up before it’s enacted into policy.
It really doesn’t make any assumptions about slate voting at all. I think part of the issue with how EPH has been discussed is that because it has been presented as a remedy for slate voting it sounds as if it has some sort algorithmic slate-detector element to it. But it just simply doesn’t.
The anti-slate element is more of a side effect of the elimination process. The idea with these kinds of voting systems (e.g. the ‘Australian Ballot’ style system) is to reduce the impact of wasting a vote on a candidate you like but who might not be very popular.
So imagine the US ran the presidential nomination process using the current Hugo nomination system. Candidates from all parties are in play and everybody just writes down say 3 candidates that they like and at the end of the process 3 candidates get nominated for the general election. Assume there are nearly equal numbers of Democrats and Republicans all voting (plus some people who might say nominate HClinton and JBush or Bloomberg or whoever). Now even though there are equal numbers of Democrats and Republicans, because there is a bigger field of Republicans the nominations are more widely spread and so you could end up with the final three candidates being HClinton, Bloomberg and Sanders! Now this could happen even if there were actually MORE Republicans nominating than Democrats because the Democrats are less varied in what they are nominating (just because there are fewer possible choices).
Applying EPH to this weird fantasy election and the results play out differently. The Republican votes are more spread out among candidates but as low scoring candidates are eliminated, the weight of the remaining votes increases boosting the chances of the most popular of the set of Republican candidates (whoever that might be). So the final set of 3 candidates then ends up having at least one Republican and at least one Democrat.
EPH won’t go into effect until next year’s nominations & voting. This year’s balloting is operating under the same rules as last year.
In that scenario the nominees would be Vox and 4 other candidates who were not Vox or on a Rabid Puppies slate. Voters might vote for No Award above Vox but they would also be more likely to vote for at least one of the other 4 non-Vox nominees. It is unlikely that No Award would win the category (although it might beat Vox). No Award only wins the category if:
1. It is the top choice overal just like a regular nominee (e.g. many categories in 2015)
2. It isn’t the top choice but more people preference No Award higher than they preferenced the top-choice. This specific outcome didn’t happen in 2015.
However, you are right that EPH doesn’t prevent an organized group of nominators getting a single work onto the ballot. What it does is make it less likely that such a group would dominate the whole category. i.e. it ensures a mix in the event of a faction (or a group of competing factions) trying to dominate the nominations.
Just filled out my nominations ballot. I have some picks that I’m iffy on and a few slots currently left blank, but overall I’m satisfied with what I’m nominating. Since we’ve got two months to change our mind, I suppose I might add or change a few things if I find some works that are worth it, but probably not much.
The only reported EPH simulation on modern real world data was from a 2014 Hugo Administrator, who stated at the last BM that “there was a change.” (One can guess that one change was a decrease in the number of Dr Who episodes making the shortlist, but I don’t think that was confirmed publicly.) A 2015 Administrator stated he would “work with” the EPH team to test the 2015 data. (I don’t know if it was done.) The EPH sponsors argued there was plenty of time to test it extensively before the ratification vote. No update since then.
In my opinion, the fatal flaws of EPH are the unsubtle shifts in its signals to voters about the purpose of the exercise. The old way was an honor system. Members were treated as a community coming together to honor its authors. EPH assumes members will “cheat” (what constitutes “cheating” has no unambiguous definition) and must be thwarted. But has no way to prevent such “cheating,”, particularly in the smaller and more arcane categories, so it effectively dares them to go ahead and try. Recipe for disaster.
The second fatal flaw is the shift from Best=most excellent to Best=most popular, and “more fans will be happy if as many of them as possible get at least one of their picks shortlisted.” Maybe we won’t always succeed in choosing excellence, but shouldn’t we strive to? That means saying “here’s my best effort to identify the best things of the year, but I’m prepared to be outvoted if the best effort of others means that other picks than mine rise to the top.”
With EPH, if a lot of people who nominate Authors A and B also nominate Author C, and A and B get the most nominations, then C doesn’t deserve to be shortlisted. Why should Author C be punished for having a fan base that overlaps with the more popular authors? That has nothing to do with excellence and everything to do with giving a piece of the pie to each of several squabbling interest groups.
My picks almost never make the final ballot and I still vote – in hopes they will sometimes show up on the long list and remind those authors that a bunch of us are rooting for them. The award isn’t there to make you or me happy about our selfish choices. It’s to honor the authors most loved by the community as a whole. If you take that spirit away, and replace it with something else, it will be something else. Hope it’s not too late.
The stated goal of the old system was never best=most excellent. It was best=least unpopular. That is stated on the Hugo page for the voting rules.
That’s how the scheme works, but with the nominations and final award balloting using different rules to try to get to the most excellent.
In the nominations phase, the attempt is to get the list of finalists down to the most excellent 5 nominations (where, in this case, the judgement of the nominators is considered the way to pick the most excellent). Once the potential winners have been winnowed down from the thousands of eligible works that appear in many of the categories, to a list of five, then the final balloting, which tends to produce the least unpopular of the excellent works as the winner (although if something is sufficiently better, the minds of the voters, then it can win even if it is unpopular, as well).
There is no ideal voting system for taking a set of individual preferences and producing a group preference. All you can do is come up with a system that doesn’t fail too often. Arrow’s Theorem is never your friend.
A different way to think about it is that the current system assumes that you’ll be twice as happy if two of your picks are on the final ballot. EPH assumes that you’ll be equally happy as long as at least one of your picks makes it. It is reasonable to say that there’s a bigger difference between one and none than there is between one and five. (Unless your motive is to destroy the awards, of course, in which case getting five is all that matters.)
I spoke to an administrator who told me had had personally run EPH on the real 2015 data and that he was disappointed that it didn’t make more of a difference. I played with the data that was released (just top 16 or so) and realized that in the best Novelette category, there were so few non-slate nominations that even with EPH the slates probably would have swept it. I shared that with the EPH folks http://nielsenhayden.com/makinglight/archives/016283.html#4216987 and they agreed that EPH really does need people to make more nominations (and to nominate 5 per category) if it’s going to work well.
I like EPH. I voted for it at Sasquan, and I’ll vote for it again at MidAmeriCon II. But people need to realize that it’s not going to work as well as they hope without an increase in nominations,
Just a reminder, the nominating electorate is so small that 300 people disrupted it. The Hugo controversy drew the eyes of forums and sites who have and can organized fundraising drives with numbers of participates that dwarf the 5-7k final voting population of last year’s awards. And, yes, at least one of those sites has mulled over pushing a slate of their own. So, imagine a repeat of last year, with a faction 5k strong influencing the nominations. Even if Puppies, Kickers, and neutrals stood side by side to repel boarders, the 5k have the numbers to sweep aside the historic 1 to 2k nominators. And that’s just some minor forum online. God help you all if Blizzard, PewdiePie, or Rooster Teeth took an interest in the Hugos. Y’know, sites whose audiences are in the millions. EPH might be able to handle a 25% increase in nominating electorate, but the real problem is that the Hugo electorate is an amoeba to the elephant that is science fiction and fantasy’s audience across all media. Grow the electorate so that 300 or 3000 newcomers don’t have inordinate influence.
But what do you think people are going to nominate en-masse? The bigger the number of people nominating the more likely that the finalists will tend towards a sort of mean-popularity. Fans of PewdiePie aren’t all going to have the same taste in books, it will be a mix of stuff and the stuff that is most likely to be in common won’t be that different than some other quasi-random grouping.
The impact of Group-X or Y joining in the nomination process depends very much on the extent to which there is SMALL VARIATION in what Group X likes within Group X.
Model it in several ways. As a random distribution of nominees with no overlap with the core fandom, as a massive mini slate around one or two nominees and no other nominatioms, as minislate plus random distributions of the remaining nominations, and a hard slate of five nominations of joke caandidates. Call this last one a 4chan raid if you must. Model each for an order of magnitude larger than the Sad Puppies.
EPH was a reaction to a perceived brigading by a group that was 25% of the “normal” group. It needs to show that it can mitigate the effects of that and similar, but more severe, events. Right now, it just seems to be a more complex way to duplicate current results while failing the “can I explain this to a five year old” test.
What needs to be done prior to a final vote, is to subject EPH to stress testing. Does it work if 4 of 6 further reduces the diversity of nominees that comments here indicate it needs? How can it be gamed? How can it be broken? What are the likely unintended effects of using EPH? Can you mitigate these unintentional effects or are they of low enough risk as to be neglible? What reassurances do we have that EPH isn’t more problematic a cure than what it is supposed to prevent. Basic due diligence, in other words. If minor mods or the scrapping of 4 of 6 are needed, let’s learn that prior to implementation.
Otherwise EPH is the oh so common solution in search of a problem.
EPH helps remediate a problem where the works that become finalists are actually unpopular with a majority of voters. Scenarios in which a majority of voters vote for stuff are not scenarios which can be ‘fixed’ with any reasonable voting scheme.
If enough people join and vote for stuff that might not otherwise win then voting method reform can’t do anything about that (and arguably that isn’t actually a problem).
Can it be gamed? A little bit. EPH helps somebody who wants to get a specific work onto the ballot but that is a less problematic circumstance then somebody getting five unpopular works onto the ballot (by ‘unpopular’ I mean works that are likely to lose in the final vote).
The current system is a bit like having a US Presidential Primary system were it is possible for the final set of candidates to be all Democrats or all Republicans.
The current system (finalists = people with most nominations) functions oddly if there is a large minority of voters with very similar votes. It doesn’t matter the ‘why’ of that, the particular motivating case was the two Puppy campaigns but it is easy to crunch the numbers and see that the system can produce very odd results when you have a any kind of large majority voting in very similar ways. By ‘odd results’ I mean an outcome where the finalists are not works well regarded by the majority of votes or an outcome were a type of work liked by many voters doesn’t appear at all in the set of finalists (which reflects the original complaint by many here).
Imagine for Best Dramatic Presentation Short Form in 201X. Amazingly there is a new series of Firefly the year before and every Browncoat fan is super excited by the revival and nominate five episodes each. In the current system if there are enough Browncoats (there doesn’t need to be a majority) then Firefly episodes can sweep the board and everybody else has nothing but Firefly to choose between. EPH makes it more likely in the circumstance that while Firefly gets represented in the finalists other stuff does too. Note there is no ‘bad behavior’ in either scenario, just fans nominating what they think is excellent.
It doesn’t assume that C ‘doesn’t deserve to be shortlisted’ anymore than the current system says that D ‘doesn’t deserve to be shortlisted’ – where D is the work that gets nominated in EPH rather than C. Assuming that both D and C are works that really should be in the pool of finalists is an argument for increasing the number of finalists rather than for either system for tallying the votes. Whether you use the current system or EPH if good stuff isn’t making it onto the final ballot then there is a problem.
The obvious example is The Three Body Problem. It only just scraped onto the ballot this year because Larry Correia and Marco Kloos withdrew their books – but it proved to be a well deserved winner of the Hugo for Best Novel.
I’m specifically going to reply to your comment, “Whether you use the current system or EPH if good stuff isn’t making it onto the final ballot then there is a problem.” — since I have strong disagreements with that claim. And I don’t want to solve it by letting enough more works onto the ballot so that there’s no good stuff that isn’t there.
The Business Meeting has always felt (and expressed so rather explicitly during debate), that, in order for a category to be considered healthy, that “It’s an honor just to be nominated.” must be true. If there is so little available that there aren’t significant good works which don’t make the ballot, then it’s not an honor to be nominated. And the fact that, in general, the Business Meeting has felt that way is why publishers have put “Hugo Award Nominee” on book covers — because the field has many good works that don’t make it onto the ballot, and making it onto the ballot is a sign that this work has had to overcome significant competition to even make it on.
If there’s a category where there’s not a lot of good stuff that doesn’t make it on, then it’s a category that should be abolished.
Thanks for the article.
By ‘good stuff not getting on the ballot’ I mean examples like Three Body Problem. I don’t mean a particular category not having any good stuff on it all but rather potentially exceptional and Hugo winning works not getting to be finalists.
Brian Z was asking about the case of work ‘C’ being penalized because it was popular with fans of A and B. I don’t think that is a fair way of looking at EPH but the other side of the issue would be what work does work C displace?
However I see your point about expanding the number of finalists in a category could lead to a devaluing of the status of being a finalist. I agree that simply being a finalist should, in itself, be an honor. In the Best Novel category I don’t think there is much danger that the honor in being a finalist would be devalued with more finalists. I don’t know about anybody else but I’m having a hard time getting it down to five! 🙂 The BDP categories this year have lots of great stuff to nominate. I’ve just finished doing a survey of all the semiprozines listed at the Clarke’s World Semiprozine directory and wow, again I’m going to struggle to get my own choices down to five.
Over at File 770 this post is being praised with faint damnation:
“steve davidson on January 30, 2016 at 1:50 am said:
I thought Ben’s piece on MGC was “appropriate”. Ben is a fixture of fandom and of Worldcon. Puppies (of whom many write for and read MGC) claim they are being excluded from fandom – (excuse me, from the teeny-tiny, ineffective, meaningless, agenda-ridden, corrupt, puppetized, clueless, controlling, elitist, meaningless, politically-correct, cult-like, unfun Trufans who have manipulated the Hugo Awards for the past twenty years) – and Ben is including them via the process of revelation.
Outreach is a good thing, even when it is likely to be a futile gesture.”
I was amused by how quickly JJ pivoted from mocking me for refusing to go say what I think on the puppy blogs to mocking me for “predictably” coming to spread my lies here.
Hi Ben. I found out about this essay from the F770 blog. It’s very informative but I’d like to correct a slight misconception about the inclusion of Warhoon 28 into the “Best Related” category. You wrote that:
…the 1981 Hugo administrator put Warhoon 28 in the Best Related category. Warhoon 28 was the 28th issue of Warhoon, a long running fanzine … which would seem to mean that the fanzine, Warhoon, was being nominated. But Warhoon 28, unlike all of the other issues of that fanzine, was a special tribute issue to Walter Willis. … So the administrator chose to put that issue/book in Best Related, rather than putting that fanzine in Best Fanzine.
The late George Flynn, who was in the know about what actually happened, provided that that Warhoon did not receive enough nominations for it to be a finalist that year in the “Best Fanzine” category. But it did receive enough nominations to make it as a “Best Related” finalist. And since there had only been a single issue that year, the Willis appreciation issue (#28), that individual issue was moved over to “Best Related”. George did not mention if the issue had received nominations in both categories, but it’s not that much of a reach to believe that it did. In the end, it made no difference as it was not the winner.
George was a font of information about things of fanhistorical interest. It’s been more than a decade since his passing, and he is missed.
Like George, I spoke to the Administrator that year, who told me (as you expected) that it had received nominations in both categories. So he had to decide which category to put it in.
As I mentioned above, Administrators hate having to make those kinds of judgement calls. So they fall back on three general principles:
(1) If, during the debate adopting or ratifying that section of the Constitution, there’s a ruling by the chair on the meaning, then it’s binding. If there’s no ruling, then if there are speeches by the prevailing side on what this section means, then it’s strongly indicative that this is what it means (speeches by the losing side don’t count, since the meeting may well be disagreeing with that interpretation).
(2) If there’s a set of precedents that are similar enough, then, in general follow the precedents.
(3) In general, try to figure out what the intent of the nominators are, and follow it if it’s a reasonable enough interpretation of the rules (and that’s where the issues come in).
In this case, possibilities (1) and (2) weren’t available. The category had first been introduced as a Special Category in 1980, and was being repeated again in that year. (Side explanation: A committee can create up to one category not in the Constitution that gets treated like any other Hugo Award, but only for that year. In this case, the 1980 committee used its category for this, and the 1981 committee followed suit. And the 1980 Business Meeting had passed an amendment adopting it as a permanent category, and that amendment would be ratified in 1981 — but that wouldn’t take effect until the 1982 convention’s awards.) So there really was no precedent to fall back on, and nothing like this had been discussed in the debate. So, since the voters seemed to want it on the ballot, and there was sufficient ambiguity in the rules that it was clear that it could be put in Related Book, then that’s where it went. But there was a great deal of discussion as to whether a numbered issue of an ongoing fanzine, independent of how it was bound, could be considered a “book”.
Administrators really hate to make judgement calls. People remember them forever, and there are often good reasons why the decision was wrong, as well as good reasons why the decision was right.
This was an excellent history and overview of the Hugo Awards process. Ben has attended more Worldcons that almost anyone alive today, has been deeply involved in the process of crafting the World Science Fiction Society’s rules for decades, and contributed extensively to A Short History of the Hugo Awards Process at TheHugoAwards.org.
Thanks for writing this! And thanks to MGC for hosting!