In a “Hugos/Sad Puppies Guest Post by Jameson Quinn”, “the guy who came up with the basic idea for the E Pluribus Hugo proposal”, does his best to unravel the coalition that passed it on its first reading at Sasquan.
In order to understand this, it’s important to see that the Sads actually did have the germ of a valid grievance: in past years, many Hugo nominators have been from a pretty small and insular group of authors, editors, and hardcore fans, who often know each other personally and whose vote is probably influenced to some extent by factors extraneous to the work itself. Writings from authors who are personally well-liked, or whose overall body of work is stronger than the individual writing, probably have had a bit of an unfair advantage in getting nominated.
Of course, that’s not to endorse the Sad Puppy point of view. Of their three complaints — that the Hugos have been too artsy-fartsy, that they have been too political, and that they have involved logrolling — the first two are sour grapes, the second two are hypocritical, and the relationship between the three exists only in their heads. Only the third could be even slightly legitimate as cause for organized action; but certainly not for the action they took, which was basically to vandalize the awards as a whole, without any hope of actually accomplishing their objectives.
The efficiency of Quinn’s self-sabotage is impressive when you consider that half of the post is wasted concern-trolling the Republican primaries.
Discover more from File 770
Subscribe to get the latest posts to your email.
Bill Ruhsam –
Not sure how advising someone to be careful in their generalizations of past Hugo nominators when they’re trying to pass a new voting measurement system through those same people is vitriol (or petty). Just looks like common sense.
There is something a bit ironic about the fear that a slight accusation of past voting according to personal attachments and cliques rather than critical judgement will lead to future voting according to personal offense rather than critical judgement.
@Mike Glyer
To be fair, I don’t know how long or well Bill Ruhsam knows you, but given your general demeanor in the past few months I’ve been here, that was about as close to vitriolic as I think I’ve seen you 😉 . I would have said, maybe, adamant? I was wondering, too, what the problem was with that article. I think I get it now. I hope the coalition for EPH isn’t too unsteady – given the voting results, I’d assumed it was almost certain to pass next year.
Bill Rusham: “My take is that you feel that by voicing an opinion about EPH and linking in the SPs and the GOP, JQ is somehow undermining its case.” That’s basically it.
I’ve been in a lot of close votes in fandom over the years, any number of times on the losing side, and one of the lessons of that has been how many different individual reasons people have for casting their vote. Now, this being the internet, somebody is immediately going to formulate that as, “Oh, Mike says this is about close votes, and EPH won by a mile, so I’ll just file this advice in the old memory hole.” No, the point is that you often can’t tell what has persuaded a person to vote in a particular way, so it’s silly to cross the streams (as they say in Ghostbusters) of your primary issue with irrelvant or secondary controversies.
Like it or not, Quinn is a leader of EPH — the guy taking credit for thinking of the idea, who crowdfunded a trip to Sasquan to argue for EPH, and who spent hours in corridors convincing people to wear a pro-EPH ribbon. Yes, he is one of the faces of EPH.
I don’t think this essay will have any effect on the EPH voting whatsoever. Remember, the only opinions that count are the people who’ll be in the MAC2 Business Meeting, and anyone who gets up early and sits through all that isn’t someone easily swayed by one article a year beforehand. Hakuna matata.
EPH breaks any and all cabals, which is what everyone wants, right?
But the Republican digression was pointless. That should have been left out, or a past election brought in instead, or a past Hugo award, or… nah, just left out.
I voted No Award on at least one of the categories which had non-Puppy entries this year. I didn’t like any of ’em.
MattY: the paragraph you excerpted and italicized was certainly true this year. Cuts both ways.
Mike: thanks for all your hard work throughout this tsuris.
Well, if it were merely advice, I would have expected to see it as a comment on the original post. What makes this reaction stand out to me is it is posted here, as a standalone post, while most separately posted items have recently been things like memorials, announcements, updates on previously posted high-interest stories, things like that. Because it isn’t fitting into the pattern I’ve observed, it piques my curiosity and my desire to understand what makes this thing apparently different.
I get your point. Without going into a whole digression about the earmarks of vitriolic writing (which would draw on past masters like Harlan Ellison and Adam-Troy Castro), my intent this morning is something on the order of “What were you thinking?”
Now if somebody said Larry Correia Channels His Inner Dogberry was vitriolic, I wouldn’t argue the case…
Well…
The irony that someone so expert (no sarcasm there) on the mechanics of politics can be so blind to the personal aspects is, well, blinding.
There was a great article on VOX about this a few days ago. Nerds and politics tends to be a bit volatile and confused.
I don’t think this is going to kill ratification of EPH, but I think a lot of EPH supporters who are suspicious of the mechanics are going to have second thoughts about trusting it.
I think it’s probably helpful to understand that the point of mentioning the Republican candidates is not concern-trolling — it’s to show how a system which is similar to EPH or the Hugo voting process, IRV, would cause the candidate with the highest preferral rate to rise to the top.
If 30% prefer Trump, but the other 70% can’t stand him and prefer, in varying orders, candidates A, B, and C, those candidates would rise above Trump in that voting system — instead of the small 30% minority candidates getting Trump to the top.
The Hugo administrators are under no obligation *whatsoever* to release this year’s results as if run with EPH, let alone to hand out anonymised data to anyone who has asked for it (and I am one of those who has asked). Whining that they haven’t done it yet is unlikely to accelerate the process.
It’s mean of you to characterize either of those two mentions of the 2015 nom data as “whining.”
Nicholas Whyte: Sasquan’s Glenn Glazer said they would do it.
It’s true I argued they shouldn’t.
But Sasquan itself set our expectations for what they are going to do. How long should it take? I don’t know, but in addition to the inevitable post-Worldcon letdown they have also been forced to pay attention to the e-mails published by Meg Frank and the antics of Lou Antonelli.
What is your sense of how long it would be appropriate to be patient?
In an ideal world, that would be brilliant and it’s exactly how they should feel – but then in that ideal world, they’d do the math and check it and find that it actually does what it says on the tin. That old “trust but verify” thing.
BTW, on a tangent – do people actually think that thinking Trump is a small hamster riding a loud bald guy is the same as ragging on the entire GOP when Trump is standing up during televised debates and openly calling every other GOP candidate a paid (well, bribed) shill for the Koch brothers? I mean… that seems (at least from across the atlantic) like a rather odd way of looking at it.
Um, aren’t they under an admittedly self-imposed obligation since they stated at the Business Meeting that they would do so?
That said, for anyone who’s not done it, anonymising data isn’t trivial (seriously, it’s an actual bona fide research topic in medicine and other fields), and if you’re not computer-savvy you wind up verifying it or even doing it by hand and with what, nearly 6000 human-entered nominations to go through, that might take a while. And presumably the admins have one or two small minor things to handle after Worldcon like paying bills, closing accounts, attending secret meetings with Mrs. Knipperdowling, writing up reports, handing over to the midamericon II team anything that needs handing over and so on. I’m looking forward to seeing the data too, but I wasn’t expecting it this month…
Treating EPH like a done deal because it passed once is foolish.
Don’t take its second passage for granted. A lot of the yes votes at Sasquan might not have been solid yes votes. They could have been provisional ones from people whose real position is “I’m not sure yet, but I want the chance to consider its enaction in a year.”
You should avoid making EPH sound like something that can achieve partisan goals (such as keeping a bad candidate from winning the Republican primary). That’s a misrepresentation of what you had a hand in creating. EPH is as politically neutral as instant runoff voting. It’s just a counting method.
Mark Dennehy: That said, for anyone who’s not done it, anonymising data isn’t trivial (seriously, it’s an actual bona fide research topic in medicine and other fields), and if you’re not computer-savvy you wind up verifying it or even doing it by hand and with what, nearly 6000 human-entered nominations to go through, that might take a while.
This is true of medical data, which contains name, birthdate, address, city, state, zip code, phone — and any of those items might also appear in patient history text fields as well as in the designated field for that item.
With the Hugo data, the only identifying info is the membership number. Remove that, and the ballot has been anonymized.
Mike, I feel you are really off base here. Having read the article I feel it is a very reasonable explanation of the decision making process this person went through prior to the proposal.
Those two paragraphs that you quoted , taken out of context, with the entire piece throw the article into a very slanted light. So Mr. Quinn believes that the Hugos may well be chosen from a somewhat biased group to begin with – if we can’t let the man have his own personal beliefs about something then we are no better than the Puppies and as reactionary as that contingent.
I ALSO do not see where Mr. Quinn trolled Republican primaries for half the article. Instead what I saw is Mr. Quinn making a quite reasonable analogy of Trump supporters to the Puppy contingent in terms of how small groups can influence elections. What Mr. Quinn said did not diminish the value of the EPH proposal nor the group supporting it.
Mike, I am a newcomer to this site and have been extremely impressed by your coverage so far but I truly feel that you are light years off base with this commentary.
Saying they would do it at the business meeting qualifies as an obligation, though I think most people who are hoping to get the data know that they might change their minds.
How do you know nobody accidentally typed personal data into the wrong field on the web form? Or that there’s not been some other accidental snafu?
I’m not sure what the US equivalent is to the Irish Data Protection Act, but if you were doing this task over here, you’d read every line of the data you were going to release twice, then give it to another hugo admin to double-check it again before releasing it because of the legal penalties you could otherwise incur from such a small accident.
UncannyValley: Mike, I feel you are really off base here. Having read the article I feel it is a very reasonable explanation of the decision making process this person went through prior to the proposal.
Those two paragraphs, taken out of context, with the entire piece throw the article into a very slanted light. So Mr. Quinn believes that the Hugos may well be chosen from a somewhat biased group to begin with – if we can’t let the man have his own personal beliefs about something then we are no better than and as reactionary as the Puppy contingents.
I ALSO do not see where Mr. Quinn trolled Republican primaries for half the article. Instead what I saw is Mr. Quinn making a quite reasonable analogy of Trump supporters to the Puppy contingent.
Mike, I am a newcomer to this site and have been extremely impressed by your coverage so far but I truly feel that you are light years off base with this commentary.
I went to the full post, expecting something horrible based on Mike’s commentary and the fact that he had chosen to make this piece a post all on its own.
I ended up mystified.
I have to agree — completely — with UncannyValley. I just don’t see what you’re seeing, Mike. Not any of it.
Mark Dennehy: How do you know nobody accidentally typed personal data into the wrong field on the web form? Or that there’s not been some other accidental snafu?
This would (should) have been caught in the initial data-cleaning process. Yes, that’s a valid point — but although IANAL, I don’t think that in the U.S., someone would have legal grounds for suing if, for instance, the data got released with a name in the Title field.
If they put their name in the Title field along with the Title or Name of nominee, their name would have been removed by the admins. If they put only their name in the Title field, then no one knows what Title/Name they intended and no harm done (especially since their name is on the Sasquan Member List, unless they specifically requested it not be).
Having done the anonymising task myself on similar sized datasets once or twice, and given that this is done in the spare time of people with fulltime jobs and families and who have a rather long tasklist competing for that time — and who probably want to take a short break to decompress after what must have been the most stressful convention this year by a wide margin — I’d hope to see it sometime in October or November and I wouldn’t start bitching about it until somewhere around christmas (and that mainly because that’ll be when Kilo next has a block of vacation time to go play with the data if his semester times are anything like what they are in Ireland).
No, it’s not nearly that simple.
You also need to eliminate any nominations that are unique to one or a handful of people, as otherwise those nominations could be used to identify people. But then those ballots aren’t actually representative for the purpose of testing the algorithm. So you need to actually replace those with other nominations, that happen not to perturb the algorithm in any way.
For example, see the lawsuit over the Netflix Prize, in which Netflix released anonymous ratings data in order to have a competition for who could make the best prediction of people’s ratings based on past ratings. However, it was possible to correlate that data with people’s IMDb ratings to de-anonymize it; you could actually pick out unique people based on public ratings they had made on IMDb, and then find information on ratings that they had thought would be private when submitted to Netflix.
Likewise, it may be possible to do the same with these nominations and Amazon or Goodreads ratings, so you have to be very careful before releasing “anonymized” data that just strips off the membership number.
IAANALOAS* but over here, they would have. You could cite honest accident and point to your due dilligence – but you’d do so as a defence, so you’d have to get into the case first.
*I am also not a lawyer or a solicitor
IAANALOAS* but over here, they would have
European privacy laws are FAR more stringent and protective of the consumer than here in the U.S. As a result we in the U.S. have near weekly spills of private information from major companies who have not implemented proper security because the consequences of not doing so are trivial whereas the cost is not.
Brian C: No, it’s not nearly that simple.
Actually, it is. If the data is released as one line item per nomination per category (which is how it must be arranged to run it through the calculation software), you get:
Membership Number
Category
Nominee Title or Name
Author or Publisher or Source
possibly date submitted
If you strip off the Membership Number (and if necessary, date submitted), you have no way of correlating to each other the nominations in various categories made by one person.
I also want to add that, as a Mensan mind you, I am in all likelihood one of the votiest voters currently out there in fandom and I think we all can relate to the immortal words of Gandalf on the bridge standing fast against the Balrog when I quote, “Some shit just ain’t worth getting this upset about!”
Mike is right. I supported EPH from the start because it would allow one work per category to be nominated by puppies each year in the hopes that over time authors who tell *stories* would get nominations.
Having the author of EPH describe Sad Puppies as “half-hearted trolls” and charaterize our arguments as “sour grapes” and “hypocritical” is not the way to win 2nd passage of EPH.
After all, many people may have voted for it not because they are for it today, but to preserve it as an option or to give more time for supporters to convince them. In other words, the margin of passage may be deceptive.
Also, characterizing 30% of republicans as supporting “outright hate” and “racist rhetoric” in a post about EPH is so stupid that all I can say is it takes a special kind of genius to be that dumb.
Nah, I see it a bit.
That suggests an interest in changing more than a vote counting scheme, and is going to undermine Quinn’s reputation as an impartial election technician. Business meeting voters are some of the most invested fans, and are going to be a bit insulted by this.
This is also a bit problematic. There are a lot of fans who might be just a bit more politically conservative than the average person attending their first Worldcon might guess, because they’re not culture-warrior conservatives, and they’re not outspoken about their personal politics. They just might be more positive on the state of the Republican Primary candidates than I am.
And (from what I hear, I was stuck at a table instead of at the meeting) there were actual sad puppy supporters talking and voting in the meeting in support of EPH, because they (like me) believe that the rabids were wagging the dog. They might find the whole thing a bit insulting too.
We’re talking a tiny number of people who can make or break the ratification. Turning any one of them off is a problem.
That won’t work for testing EPH. You need to group all of a single person’s nominations per category together, so that the score can be divided between them. Then you would be able to see correlations within a particular category. Sure, you could still gain some anonymity by disaggregating people’s nominations between different categories, but you would still need to make sure you removed entries that are unique or rare enough within categories to be used for fingerprinting, or possibly replace them with opaque identifiers, so they can still factor into the algorithm without being usable for correlating against other sources.
Anonymity is hard. Given the smaller population, and ability to break things down per category, it’s probably possible to do a better job than Netflix could and not be subject to a similar lawsuit; just a list of nominations conveys less information than ratings do, and with most people only submitting a handful per category it’s probably hard to get very much in the way of interesting information out. But I do think that there should probably be at least a minimal effort to hide unique and rare nominees (which, as pointed out, may also be personally identifying information accidentally entered as nominations), in order to make an effort to prevent fingerprinting.
OK–let me try that again. I didn’t realize that the five-minute clock kept ticking even when I re-opened a comment to add more…
And that is the problem that our Hugo system admin folks have been running into. When one of them generated a draft of anonymized nominating data, it didn’t take the other very long to determine who some of the voters were, simply from the voting patterns.
I haven’t heard an update from them recently, but we all have a lot immediate post-con concerns that need to be taken care of (says the man with stacks of boxed Hugo trophies in his dining room, waiting to be shipped to winners who could not attend Sasquan).
There are two things I want to respond to:
1) The Hugo Administrators (John Lorentz and Ruth Sachter) did not promise to release the nominating data. That was Glenn Glazer (Sasquan vice-chair), who immediately stated this would happen before Ruth and I could even respond to the request. (We had expected that we would have been part of that conversation.)
2) No one (not even Glenn) promised that this data would be released before the Sunday Business Meeting. It simply wasn’t possible to do this, with all the other work we all needed to take care of. (Ruth and I were basically on-duty non-stop from 9AM Saturday through 1AM Sunday.)
I have no idea when the raw data will be released. Only the two Hugo system admin people have access to that data (all the work we did as Hugo Administrators was performed using front-end tools that they developed).
All I know is that our immediate concern is to get those Hugo trophies out and then we are done–never to take on this task again. (Out of the four times that I have been a Hugo Administrator, this has been by far the worst. Even LoneStarCon 2, when I took over at-con reg three weeks before the convention, was easier than this.)
@Brian C
I could see going with a substitution algorithm where “sad puppy XXX” or “rabid puppy XXX” or “sad/rabid XXX” are used to replace slate candidates based on which slates they appear, and some other randomized mapping replacing the other candidates.
I could also see breaking out by category, so you can’t figure out what the person who nominated “The Drink Tank” in fanzine (and might be identifiable because of that) did in any other categories.
@JJ to test EPH you need to know all of the blanks for a category for each nominator. I believe getting 5 disconnected nominees isn’t going to make it possible to identify slates statistically, and isn’t going to allow the leveling to be tested.
@ Happy Puppy:
Who are you referring to by “our arguments” ?? If you mean the sad puppies then yes, Mr. Quinn did characterize their arguments that way – as we have all pretty much stated on this forum. But if you are trying to imply that “our” refers to we here in file 770 or the bulk of fandom then you are wrong – that was never said.
If indeed you are a sad puppy supporter yourself then it would be in your best interests to vote against something like EPH and I can understand your condemnation of Mr. Quin in that context alone.
Don’t ask me how I know, but JQ has nothing on me when it comes to eating flaming kittens alive. I am, in all modesty, while not needing to repeat those immortal words, one of the finest flaming-kitten eaters currently alive.
“Hey, people who just voted for EPH and are likely to voting on it again next year, here’s who I think you are!”
pretty small and insular group of authors, editors, and hardcore fans, who often know each other personally and whose vote is probably influenced to some extent by factors extraneous to the work itself.
Have a nice day!
This would (should) have been caught in the initial data-cleaning process
Well, that’s nice. “Should” and $5 will get me a latte.
This is the kind of thing that gets social scientists in trouble. Just because something “should” be done in a certain way, and thus we can do this next thing, doesn’t mean it actually was done that way.
Andrew Trembley: That Vox article is great– thanks for posting the link!
Everyone else: Read it! The later parts can also serve as a great primer on US politics for anyone outside the US.
Okay, I see where the issue is coming in.
When EPH was tested with the 1984 data, the membership number was replaced with a random ID number. This enables correlation of all the votes in a given category for a given person, without personally identifying the person.
This is only a problem if the same ID number is used for the same person across all categories. If a different ID number is used for different categories, the only real-life correlation which can be determined is by, say Susan Blogger, who says “I voted in this order for this category”.
But she’s already told everyone that publicly, so there is no violation of confidence. If her votes in other categories all have different ID numbers, then you still have no idea how she voted in other categories.
I think that we may be misunderstanding each other.
To run the EPH algorithm, you need to know the full contents of one ballot in each category. If I have nominated 7 works in a category, then you group all seven of those into a single ballot. That then gets one point, divided evenly among all seven works. Once one is eliminated, the points are distributed to the remaining 6 works on that ballot.
How can you do that with individual nominations separated from each other?
Also, I’m unfamiliar with the 1984 data. Is it public, or was it just provided to a select few? I don’t see how you could apply EPH to it if you didn’t know which nominations came from the same ballot (within a category, I agree that you don’t need to group nominations together between categories), unless you grouped them in some artificial manner, similar to how artificial data was generated for the 2013 nominations.
@ Total :
Except that was not what was said and is a complete mischaracterization of what Mr. Quinn actual said which started with ” … in past years, many Hugo nominators have been from a pretty small and …”. Which refers to some Hugo nominations in an unspecified past – NOT what you stated about EPH voters.
Frankly if one is so thin skinned as to throw the EPH baby out with the bathwater just because the originator of the proposal has some critiques about the nomination process and voter pool then you are not helping.
Mr. Quinn should be hailed as a hero in proposing EPH and his supporting mathematical arguments in the business meeting. Attempts to vilify him for some opinions which may run contrary to one’s own are, in my opinion, petty, unwarranted and overall a distraction to the actual proposal which only serves the Puppy interests in the end. Real fandom need to be better than this.
As far as the release of the 2015 data and the 2014 data (which was also promised), I am as anxious as anyone to have a look at it.
But Worldcon is barely over, there is a hell of a lot of work yet for the committee to do tying things up, and we all need to be patient.
We have almost a whole year to run tests and discuss the results. I don’t think it’s helpful right now to be critical of the people who will be supplying the data.
@UncannyValley – Don’t forget that there are many Sad Puppies who honestly believe that there were conspiracies running secret slates in prior years. A number of them have said they like EPH. They aren’t all in love with slates.
It’s the culture warrior Rabids and people like BrianZ who are seriously against EPH.
@UncannyValley:
It’s perfectly clear from context that he did.
Wrong. Wrong wrong wrong. Endlessly, bottomlessly, proudly and arrogantly and stupidly wrong. Puppies are quite right to support EPH, and quite right to ignore fools who tell them otherwise.
@JJ
One of the other things I believe people want to do (besides test EPH) is review the nominating ballots for lock-step voting on sad and rabid slates, to gauge how people actually followed them (or if they did).
A substitution algorithm (on normalized ballots) would make it easy to determine if people followed slates for their whole ballots or only specific categories. But it would be more work than just breaking the ballots apart by category and removing names.
Hi, Andrew Tremblay and Brian C.:
Please see my revised comment above.
@Uncanny you’re failing at the “personal” aspect of politics, and expecting logic to win.
I’m not going to say that never happens, but it’s at a strong disadvantage.
@Maximillian – perhaps true. It’s true that I tend to group the two groups together and am skeptical to there being any real difference based upon the rantings of the assumed leaders of the Sad contingent.
@JJ I think we’re on the same page re the “break apart the individual ballots by category” scheme to make the data less identifiable being a viable option.
Andrew Trembley: One of the other things I believe people want to do (besides test EPH) is review the nominating ballots for lock-step voting on sad and rabid slates, to gauge how people actually followed them (or if they did).
People may want to review that — but it’s not required for testing EPH, and it’s not the purpose for which the data is being released.
I firmly believe that the anonymized data should not allow correlation across categories, and I hope that the data is anonymized so that it will not be possible.
No, even within a category, if someone nominated two unique works in that category, and I can find their profile on Goodreads that way because they’re the only person who rated both of those works on Goodreads, then I have (with high probability), de-anonymized them, and been able to find out perhaps jucier information on other works that they nominated. Or maybe I find that there was exactly one person who nominated the editor for one of those works, so there’s a good chance that those two ballots came from the same person. And so on.
John Lorentz points out that they did a first pass of trying to release anonymized data, and could figure out at a glance who some ballots correspond to. If it were the 1984 data, that may be less of a problem; it’s so far in the past, there aren’t any currently open wounds about it. With as politicized and nasty as this years nomination and voting have been, it could cause serious problems to release data and then have people comb through it and figure out things like who puppies were (beyond the ones who have voiced public support), or maybe the puppies comb through and come up with conjectural cabals linked to the people they want to demonize.
Yep, saw it, but timed out on editing my own reply, so have now done so separately.