Friday, July 03, 2015

Hugos Round-Up

Multiple topics for this post. First up is my "How did I do?" on whether or not I found the slate works to be worse than the non-slate stuff and to answer Denise's standard question (Any of the nominees on the Puppy's slate?). Then up is my overview of the E Pluribus Hugo proposal to depower slates in the Hugo Award Nominations with a lame story/example provided.

I also thought I'd mention that I'm finally listed as a supporting member on the Sasquan membership list, which makes me very happy. I wish I could attend, but life being what it is, I'm just glad I was able to become a supporting member.

Just as a reminder, you can go to Renay's 2016 Hugo Spreadsheet of Doom or the Hugo Nominees 2016 Wikia to look for works that will eligible for a Hugo in 2016 so you can read them for possible nomination. Both the spreadsheet and wiki allow submissions, as well, so add your own favorites to get them a little more widespread attention.

The main content is after the cut for those of you viewing this on my main page (I'm trying to spare my usual 30 or so readers, who may not be all that interested).

How Did I Do?

So, let's work out for this week what puppy works there were in my reviewed categories and what I thought of them for the Hugo ballot.


There were three slate nominees in the Fancast category. All three were nominated by both the sads and the rabids. The slate nominees were "Adventures in SciFi Publishing", "Dungeon Crawlers Radio" and "The Sci Phi Show".

I ranked "Tea and Jeopardy" and "Galactic Suburbia Podcast" at the top of my ballot; neither was on a slate. I put "Adventures" third and left the other two off the ballot. As with the first four categories, this one sorted itself out into puppies and non-puppies by quality, in my perception.

Fan Writer

Ug. This category really got me upset. All the hate in three of the writer's works: it was pretty obvious those three were puppy nominees. Sure enough, Dave Freer was nominated by the sads while Cedar Sanderson and Amanda S. Green were nominated by both.

A slight surprise to me was Jeffro Johnson, who was also nominated by both slates. Johnson's was the only entry that I thought was actually focusing on things I wanted to read about. However, the writing was definitely disturbing in places and turned me off. I don't really consider it a Hugo winner, so it was below "No Award" on my ballot.

The other nominee, Laura J. Mixon, was the only non-slate entrant. I also put her below "No Award", but her writing was interesting and well-researched instead of hateful and childish. Had she included writing that was about science fiction or fantasty instead of only works about fandom, I probably would have ranked her higher. I'm still considering the category definition, and she may go higher before the deadline to finalize the votes, I'm not sure. I may give her stuff a re-read just to see. I've read some compelling arguments that works about fandom should be considered in the category.


The Novella category also disappointed me. Only one story was worth reading, and that one was hardly great. Just decent. I was not surprised to find out that all five nominees were from the slates. "Big Boys Don't Cry" and "One Bright Star to Guide Them" were nominated by both slates. "Pale Realms of Shade" and "The Plural of Helen of Troy" were both on the rabid slate.

The only story worth reading of the bunch, "Flow", was only on the sad slate. I ranked "Flow" below "No Award" and left the rest of the nominees off my ballot. The more I think about "Flow" in retrospect, the less I liked it as a story. I don't think it'll change my vote, but the story isn't one that holds up well on reflection.


No Award topped this category on my ballot, but there was a non-slate nominee. "The Day the World Turned Upside Down" was not on a slate, which actually doesn't surprise me. It was an ok story, but I don't consider it Hugo-worthy.

The other four nominees were all on both slates. None of them were award-worthy. A couple of them weren't awful and even edged into "good" territory, but they just aren't the caliber I expect from Hugo works. Not even close.


While there's some non-slate stuff that I also didn't find Hugo-worthy, so far none of the slate nominees have impressed me enough to put them above "No Award".

I expect that to change once I get to the dramatic presentations, as those all seem to be more mainstream choices. With all the literary awards, the puppies nominated their friends. They couldn't really do that with the dramatic awards, which are all professional works with major distribution and fandom. What they did do is keep any experimental and strange works off the ballot in the dramatic categories, which is bad enough.

E Pluribus Hugo

As soon as it became apparent that slates can, indeed, control what gets onto the Hugo ballot, the greatest mathematical minds in fandom got to work on trying to figure out a way to prevent slates from dominating while still keeping the openess and diversity of the Hugo awards intact. To say that E Pluribus Hugo is an elegant solution to a disturbing problem is an understatement. The way the math works, it will literally target slates without changing anything else.

There is a lovely set of PDFs and Power Points that explain the concept far better than I ever will be able to, but I'll try to sum up.

The proposal takes all the nominees and figures out how to eliminate the least popular by proportional voting. The problem with slates is that 200 people all voting for the same five candidates can easily overcome 1,300 people voting for a large variety of candidates. So if you go strictly by number of nominations, the slates will almost always win.

To counteract that, nominations are given a value. As the least supported nominees are eliminated, the remaining nominees on each ballot get a higher value. This means that people can still nominate five choices, but when one of their choices is eliminated from the running due to lack of support from other fans, their other four choices get a higher value, and thus have a lower chance of being eliminated.

It defeats slates because the items on slates tend to compete against each other... having the same number of nominations for the same number of candidates. Only when a set of candidates has nearly the exact same number of nominations each does this mechanism make a difference. Then the difference it makes is that it takes the slate nominees down to a single one, instead of allowing them to have all five.

Did I mention that there's a fantastic PDF put together by Keith "Kilo" Watt that really explains this far better than I can?

The more I look at the problem, the more I'm convinced that a technical solution (instead of bright rainbow-y fan force taking back the Hugos) is going to be required to override the people who have decided to make it their mission in life to destroy the Hugo awards. I believe E Pluribus Hugo is probably the best technical solution anyone has come up with. People who see a flaw with it need to go to the thread about it and point it out so it can be fixed.

Here's my clumsy effort to explain it more simply. I'm sure reading the PDF will get most people much further...

Infinite Pies

At the Infinite County Fair, the great pie contest has always been a highlight noted by the general public. Nominees for best pie sometimes see their sales go up, and the winners always post that they've won on their bakery doors. Judging in the pie contest is open to everyone for a small fee, but not everyone goes to the fair, and those who do don't always vote in the pie contest, which has two phases.

In the first phase, everyone who wants to and has paid the fee can pick three pies they think are excellent, the best of the year, from all the pies they have tried in the past year. The fair organizers then take all the suggestions and tally them up. The top three vote-getters are on the final ballot, and are invited to provide pies to the judges (although they don't have to). Then the voters pick the top pie from that final ballot in the second phase, ranking them by preference.

In 2014, only 50 people bothered to nominate pies. Because there are so many to choose from, the highest recommended pie (Julie's blueberry silk with mangoes) got only 7 recommendations. The second highest recommended pie (Jamal's chocolate cinnamon delight) got 6 and the third (Mama B's old-fashioned cherry) only got 4. Every other pie got 3 or less recommendations. Jamal's pie was the eventual winner of the run-off between the three pies. This is completely normal, because of the low level of actual involvement. Although lots of people in Infinite County eat pie, many of them like seeing the curated results and then use the pie contest to find out what pies to try. Other people don't realize they are allowed to vote, so have never tried. However, some folks in the town of Curville feel like their pie-makers are being snubbed.

As a result, a group of ten people from Curville got together and decided to run a slate for the pie contest in 2015. All ten people would nominate exactly the same pies, and hope that was enough to get them on the ballot. They picked Granny F's traditional apple, Susie's 'better than old-fashioned' cherry and Dan's dogfood pie. All ten people paid the fee and nominated the same things.

In the meantime, the usual group of 50 people nominated their picks. The top pick for them was Elayne's awesome avocado pie, which got 8 recommendations. Mark's molten chocolate and caramel surprise got 6 recommendations and Nevaeh's nutty cream pie got 5 recommendations. Granny's pie, which was on the Curville slate, got a single vote from one of the usual people who nominated.

But none of the usual picks got on the ballot. Instead, the slate nominees dominated. Granny's pie was ok, Susie's pie was not actually better than old-fashioned and the dogfood pie was an insult to everyone who likes pie. But the Curville folks were delighted and talked about how they had "taken back" the pie contest from some imaginary cabal. They started to insult the pie-makers they had pushed off the ballot, claiming those pie-makers were making pies only to be healthy and not to taste good. Then they complained when some of the usual voters said maybe an award shouldn't be given this year, an option available by the rules but rarely used.

The 50 normal voters were dismayed, especially when several of the Curville voters made it clear they'd never even tasted the pies they nominated. As more and more people heard about the incident and started to pay attention, they formed opinions based on how much they knew about the contest in the first place, who they'd heard the news of the incident from and their existing opinions about the people involved. Some paid the fee to vote on the final pie and to nominate pies for next year. There was much uproar in pie-loving circles.

Some more mathematically inclined pie-lovers decided there must be a technical way to stop slate voting without changing the rest of the results. What they came up with is a proportional voting system that pretty much requires a computer to calculate, but comes up with almost the same results if there are no slates, and if there are slates, generally reduces the slate choices to one.

What they did is assign a single point to each nomination form, divided among the choices on the nomination form. So for the slate votes, every pie got 1/3 of the point and the slate nominations were each worth 3 1/3 points, except for Granny's, which got an extra vote and was worth 3 2/3 points.

For non-slate votes it was the same. Elayne's pie, which got 8 nominations, was worth 2 2/3 points, Mark's pie was worth 2 points and Nevaeh's pie was worth 1 2/3 points. All the rest of the pie nominations had less points because they'd been nominated fewer times.

Then the Pi-lovers started with the nominees that had the fewest points (those that got 1/3 a point) and eliminated them based on how many nomination forms they showed up on. When there was a tie, both nominees were eliminated. Every time a nominee got eliminated, the points were redistributed on the nomination form, so if something on a form is eliminated, the remaining two are now worth half a point each instead of 1/3 a point.

Because most voters had at least one unique nominee, most non-slate nominees jumped in value from 1/3 a point per nomination to 1/2 a point as the lowest nominees were removed, raising the total point value of those nominations. In the meantime, the slate's choices continued to be valued at 3 1/3 points (except for Granny's, worth 3 2/3).

Elayne's pie, which started with 2 2/3 points, got more points as other nominations on the 8 nomination forms that included it were eliminated. After the first round of elimination, which effectively removed all the unique nominations, Elayne's pie was the only nomination left on two of the eight ballots (2 points), one of two on five more ballots (2 1/2 more points) and one of three on a single ballot (1/3 points). So Elayne's pie was now at 4 5/6 points, higher than the slates.

A similar thing happened to Mark's pie, which had 6 nominations and started with 2 points. As the unique nominations were removed, Mark's pie was the only one left on three ballots (3 points) and one of two nominations on the other three (1 1/2 points) giving it 4 1/2 points, also higher than the slates.

Several more rounds removed more nominations. Those worth 2/3 points went first, then ones worth 5/6 points. And so on. Each time a nomination was removed, the remaining choices on the nomination forms got bumped up in points - EXCEPT for the slates, since none of them had been eliminated. Some nomination forms got discarded as all their nominations were eliminated, as well.

Eventually the elimination rounds reached the slates because the slate nominees now had the lowest point values. Dan's and Susie's pies both had the same point value, 3 1/3, and both had the same number of nominations. Therefore, both were eliminated. Granny's points instantly jumped from 3 2/3 up to 10 1/3, which guaranteed her pie a spot on the ballot.

In the end, Elayne's pie and Mark's pie got the other two spots due to the highest point values, leading to a much more fair ballot. The pie-lovers across the land lamented that the system wasn't already in place, but nobody thought anyone would be so rude and obnoxious as to game the system.

The new method of selecting the final nominees for the ballot from the nomination phase appeared to work, so it was put in place for the next year of the fair.

Now, scale that up by an order of magnitude, add more pie categories (savory pies? breakfast pies?) and increase the number of nominees per ballot to five. There's the E Pluribus Hugo method of preventing slates from dominating the ballot.

Yes, it's math, but voting is already generally tallied by computer. The system was tested, as well as it could be tested, against data from 1984 and data from last year. The neat thing about it is that Hugo nominators don't have to change their behavior at all. It won't be better to only nominate a couple or a single item, because your works get more points as the less popular stuff drops off anyway. The trick with it will be convincing people that the math does indeed work, which is why I suggest reading the PDF instead of reading about my pies (mmmm... pies).

The second biggest problem with it is that it will take at least two years before it goes into effect. If it's proposed this year, it will not go into effect until after next year's Hugo nominations. That gives Curville another year to nominate yucky and mediocre pies. While I originally thought enough other pie-lovers joining the fight would swamp the Curville slates, I was presented with some disturbing mathematical calculations that pretty much put that notion to bed. While more fans participating is a great thing, it's doubtful that enough can join to really stop slate nominating.