So, this year was my third time as Hugo Award Administrator, and my sixth year of being involved with the Hugos in general. And it was by far the weirdest.
I was the Hugo Award Administrator in 2017 and 2019; Deputy Hugo Administrator in 2020 and 2022; and WSFS Division Head for a few months in 2021. This involved Worldcons on three different continents – Helsinki (2017), Dublin (2019), New Zealand (2020, though conducted virtually in the end), Washington DC (2021), Chicago (2022) and Glasgow (2024). Each presented their own problems. I’ve written up the 2017 experience here and here, and 2019 here and here, and more briefly 2022 here; 2020 and 2021 were too painful to write up in full, for somewhat different reasons.
I see from my records that my discussions with Esther MacCallum-Stewart, the Glasgow 2024 Chair, about taking on the role of WSFS Division Head began as early as February 2020, though not confirmed until November 2021. I got my team lined up fairly early – Kat Jones as Hugo Administrator from the beginning, with Cassidy as her deputy and Kathryn Duval as my deputy from October 2022, and Laura Martins joining the team also at an early stage. For the other parts of the WSFS Division, Jesi Lipp came on board as Chair of the Business Meeting in September 2022 and Naveed Khan and Thomas Westerberg for Site Selection in April 2023. Bridget Chee also volunteered to wrangle Hugos on the ground in Glasgow. So far, so good.
But several things went very awry in the process. The first of these was software for administering the Hugo voting and counting. This has been the subject of Monday morning quarter-backing from people who are unaware of the constraints under which a Worldcon operates. The first thing to understand is that for the convention, the registration software is critical to the functioning of the entire organisation. The Hugo and Site Selection stuff is a secondary issue, and no matter how hard you may try and push, the fact is that the convention will be embarrassed if the Hugo stuff doesn’t work properly, but will go bust if the registration stuff doesn’t work properly.
For various very good reasons, which I don’t intend to go into here, the Registration software often tends to get largely or completely rewritten for each convention. WSFS’ aim under my watch has generally been to get the Hugo software talking to the registration system and up and running at an early stage in January for nominations and in April for the final ballot. This does not always work. 2020 was the worst case in point, but it’s a bit of a bare knuckle ride every time.
This year, we had an external service provider who did indeed produce a good registration solution. But their Hugo work fell short of expectations, and we had to resort much more hastily than I would have liked to volunteer efforts – a combination of Kansa, the venerable but creaking back-end for tallying nominations written by Eemeli Aro in 2017 and subsequently updated by David Matthewman, and NomNom, a bespoke solution for the front end of nominations and for both ends of the final ballot written specially this year by Chris Rose. This was precious and valuable volunteer time, and I cannot thank them enough.
Even so, the launch of Hugo nominations voting glitched very badly when it turned out that there was a serious software problem linked to the registration system interface that needed to be resolved, and we had to stop the voting a day after it had started and restart several days later. Henry Balen was also crucial to managing the relationship with the software provider at this point.
In the meantime, we faced much bigger reputational problems affecting the Hugos as a whole. I attended Chengdu Worldcon in 2023 and enjoyed a lot of it. But the data from the final ballot vote, released at the start of December, looked distinctly odd. And when the data from the nominations count was released in January, I was dismayed to see that it was clearly very flawed. The numbers simply did not add up, and were clearly not the output of a genuine ballot count.
The other immediate flaw that caught public attention was the disqualification of several finalists without explanation. It has been generally surmised that this was because of overt or implicit censorship from government authorities. I do not know if this is true, but it seems unlikely to me to be the full story. Babel, by R.F. Kuang, which was disqualified from the ballot, actually won a Chinese-organised prize in Chengdu a few months later. On the other hand, John Chu’s story “If You Find Yourself Speaking to God, Address God with the Informal You”, which is about a gay relationship in Taiwan, was allowed to stand as a Hugo finalist.
Personally, I am very disappointed with the behaviour of those who led the Hugo administration in 2023. There is no excuse for the breach of trust with voters and nominees, or for the damage that was done to the Hugos as an institution. The lack of transparency around the decisions is an additional reason for frustration, but the basic point is that the vote very deliberately failed to reflect the wishes of voters.
I have reflected on what I might have done if I were in the position of being instructed that certain Hugo nominees were not allowed on the ballot by local law. (NB that it is still not clear that this was what actually happened in 2023.) The Hugos and Worldcon must of course obey local legislation. But I would have wanted very clear professional advice before taking any such steps, preferably advice that could be published. And if put in an ethically difficult position, I like to think that rather than proceed, I would have resigned – as I did under milder circumstances in 2021.
A silly proposal went to this year’s Business Meeting suggesting that there should be a committee performing broad surveillance of how Worldcons choose WSFS software solutions. This is a bad idea. It adds bureaucracy to a fraught process, blurs lines of responsibility, and ignores the big issue for the convention itself (the registration software). If you think that the biggest problem with the 2023 Hugos was quality control of the software, I have a panda sanctuary that I’d like to sell you. But I will co-operate in good faith with the committee that has been set up to look into these things.
Unfortunately as more details came out about what had happened at Chengdu, it became clear that Kat Jones could not continue as 2024 Hugo Administrator and she resigned. After some reflection, I took on the role of Hugo Administrator myself, doubling up on my existing job as WSFS Division Head. I would not in general recommend this; there are good reasons why these two jobs are generally done by two people.
Translation was a big issue. The many WSFS members of Chengdu were all entitled to vote in 2024 Hugo nominations, and we commissioned Sophia Xue (Xue Yongle) from Shanghai to translate all the relevant materials. I had met her, ironically, because of the 2023 Hugo disqualifications, as a result of which she had qualified for the final ballot, and we had had a long chat at the Hugo ceremony in Chengdu. She caught a number of flaws in the Chinese translation of the WSFS Constitution that had been done the previous year.
I have described the work of the Research Team in the Administrators’ Report and elsewhere. It was essential to the exercise. We had to disqualify three potential finalists, all Chinese (though one also qualified in another category) and another five declined nomination, a couple of them rather late in the process. There were some other tricky calls – Astounding finalists with self-published early work, tallying votes for two of the three volumes of a non-fiction nominee. Northanger Abbey was not a tricky call, but I included it in the final report for amusement.
I had had high hopes of producing a video announcing the final ballot with a professional production company and a well-known Scottish actor. The costs, however, were simply unfeasible, so we ended up with a video of self-recorded announcements by various speakers, prefaced with an introduction filmed by my son of me at the Atomium, north of Brussels, which is as science fictional a backdrop as you can get in our part of the world.
I was particularly pleased that we managed to get Geoffrey Gernsback, the oldest great-grandson of Hugo Gernsback, after whom the Hugos are named, as one of the announcers. He supplied us with a photograph of himself as a baby with his great-grandfather. Assembling the whole thing was a mammoth task accomplished without visible seams by Meg McDonald and James Turner.
Looking at the categories, ten out of twenty included Chinese finalists writing in Chinese and ten did not, so we decided to ask Sophia Xue to read the names of the former while I did the latter. (We then used the same audio for the ceremony, which startled me when I heard my own voice booming into the auditorium.)
We also held a town hall meeting for the Hugo finalists a couple of days before the announcement was made, partly for transparency but also to help finalists to make the most of their nomination status in local media.
People sometimes ask why we do not open voting on the Hugo final ballot as soon as it is announced. The simple fact is that there is not enough time. As noted above, we had a couple of very late withdrawals. Better to do the last coding twiddles when we know that the candidates are settled. We did the announcement at Eastercon on the Friday evening, including also details of those who had declined or were disqualified, and posed with those finalists who were present.
At Eastercon we also unveiled Sara Felix’ tartan rocket:
I should mention also the Hugo Helpdesk team, led by Terry Neill backed up by Rosemary Parks, and the Hugo Packet team consisting of Dave Gallaher, Jed Hartman and Scott Bobo, who ensured important elements of the user experience. The Packet in particular needs perhaps a bit more attention from WSFS as a whole; it is an important element of voter expectations, but is nowhere to be found in the Constitution.
I am out of chronological order here, because an early and easy decision in the entire process was to commission Iain Clark to design the trophy base. In general I prefer to go to known creators; running a competition absorbs time both from artists and administrators. Iain’s design is simple and pleasing, and leaves no doubt about the geography of the convention.
Hugo voting ended on 20 July, 55 years to the minute after Neil Armstrong and Buzz Aldrin first stepped on the Moon. I was on a plane from Munich to Brussels at the time: my flight from Los Angeles the previous day had got delayed landing and I had to argue my case for Lufthansa to get me home that night. I landed knowing that we had more issues to deal with.
In about two thirds of the categories, an early lead was established by the eventual winner. We watched the inflow of votes for the rest with interest. In early June we became aware of something unusual – a particular finalist, who we will call Finalist A, was picking up something like ten nominations each day, at a time when most others were picking up just a couple. On closer examination it became clear that almost all of Finalist A’s votes were coming from newly purchased memberships, which were behaving quite unlike memberships controlled by real people. We took no immediate action – as Napoleon said, never interrupt the other side when they are making a mistake – but as history records, we disqualified 377 of those votes, and Finalist A therefore did not win in their category.
My hypothesis, based on the data that I have, is that a well-resourced fan of Finalist A hired a ‘vote farm’ to get a Hugo for their favourite creator. A notice was published in some public or semi-public online or meatspace forum, inviting people to make a quick buck by buying a Glasgow 2024 membership and sending the sponsor proof that they had voted for Finalist A. The sponsor then paid their money back plus a bonus. The greedier ones put in multiple memberships in alphabetical sequence, or in one case giving names which were translations of the numbers from 13 to 17.
For reasons which should be obvious, I am being circumspect about the precise details. I don’t want to make it easier for the next person who decides to do this. I do want to emphasise that the evidence points against Finalist A being involved or knowing about this in any way. I should also add that I have briefed and will brief future Hugo administration teams at greater length.
No other outcome was changed by the disqualification of the 377 votes – actually a couple of close finishes in other categories were widened. We found that disqualifying votes in NomNom is a rather tedious activity, which is just as well – it forces you to be absolutely certain that you are getting it right.
Out of time for today; coming soon, my experience of the convention itself.
Continued here.