SCFZ

we haven't done our annual 'scfz's top 100 films ever' list yet in 2019. we had a discussion in stop the lists recently about perhaps running the poll differently this year. the scfz top 100 lists have always been somewhat eclectic (except one year where we were flooded by letterboxd voters who i'll exclude this year no matter what), but our more identifiably scfz favourites have always appeared alongside films like The Godfather and 2001, which are found in any 'canon' list.

the question arose whether we should introduce some restrictions, or some vote-weighting, either to exclude canonical films like The Godfather, or to elevate less-seen films (that might rank more highly were they more widely seen). this poll is only to find out if people want an open poll with no restrictions, or if people are more interested in a poll with some kind of restrictions or weighting -- if that option wins, we'll work out what those restrictions will be, democratically (and if we can't agree on anything, we'll go back to the default).

i don't want this thread to reopen a debate about whether restrictions are good or bad, but in case anyone didn't read the stop the lists discussion, i'll summarize a few of the arguments on both sides for people to consider. if i've missed a compelling argument for or against restricted polling, please post it below.

arguments on both sides, but mostly on the 'in favour of restrictions' side, because that's probably less immediately appealing and requires more justification:

- if our top 100 is going to truly reflect our community's favourite films, we need to be free to vote for any film we want

- everyone already knows about 2001 and The Godfather, so if people are using our top 100 list to discover new films, it's not useful if they appear on the list

- we rarely discuss some of the most canonical american films on scfz, so if we want our top 100 list to serve as an advertisement for scfz, as a way of saying "we're a community that likes talking about these films", those canonical films probably don't deserve a prominent ranking

- the most widely seen films rank well in part because everyone has seen them, so they're eligible to get votes from everyone. there may be films only five people here have seen that would rank near the top of our top 100 list if everyone saw them. so, to overcome the popularity bias that favours famous films, we should give some added weight to less-seen films in our rankings

- we've done four annual polls already with no restrictions, so maybe it would be interesting to try something new

i hope all of that makes sense, i didn't get much sleep

i see the votes have tightened up. i'm good either way (voted for some restrictions tbh), but the point i like the most is:

flip wrote: ↑Mon May 13, 2019 3:35 pmwe've done four annual polls already with no restrictions, so maybe it would be interesting to try something new

not in favour of restrictions, but i am definitely in favour of weighting

as in, my order is:

1. weighting
2. nothing
3. restrictions

how should i vote?

i agree with bure, but with minimal weighting.

yep.

minimal weighting
nothing
restrictions,

in that order.

i don't think i said anything in the stop the lists thread, but...i think the idea of doctoring the rules of a poll so the results match perceived self image is genuinely bizarre. i do think weighting helps, maybe.

nrh wrote: ↑Tue May 14, 2019 1:14 pm doctoring the rules of a poll so the results match perceived self image is genuinely bizarre.

bure420 wrote: ↑Tue May 14, 2019 5:07 am not in favour of restrictions, but i am definitely in favour of weighting

as in, my order is:

1. weighting
2. nothing
3. restrictions

how should i vote?

vote yes! this poll is just to find out if i'd be wasting my time compiling all of the suggestions in stop the lists, and also to get all of the potential justifications for/against a new system in one place, for anyone who missed the stop the lists conversation.

if 'yes' wins, then i'll create another poll where people can vote on each system, yes/no, vote yes to as many things as you like, and we'll use whatever rules a majority approves of.

nrh wrote: ↑Tue May 14, 2019 1:14 pmi think the idea of doctoring the rules of a poll so the results match perceived self image is genuinely bizarre.

but as soon as you create any rules for a poll, you're 'doctoring' the rules somehow. there's no objectively correct way to run a poll like this, and any set of rules creates systemic biases. for example, why do we ask for a top 50 ballot from each person? there's nothing objectively correct about asking for ballots of a fixed length. with top 50 ballots, the preferences and viewing choices of people who have seen the fewest films overall have the most weight in the poll, since those people can include all of their favourites, while the person who has seen 10,000 films needs to exclude many of theirs. we could instead ask people to post a ballot of the top 1% of the films they've ever seen, or ask people to rate films dtc-style. but those methods have their own biases too.

as i see it, any weightings/restrictions we might use are to compensate for the biases of the imperfect system we've been using.

i also don't think we're trying to match 'perceived self image', or at least i'm not. there are films and filmmakers we talk about, and others we don't, so i think our 'self-image' can be evaluated a bit objectively. i thought i'd try to demonstrate that by taking a high-ranking filmmaker from our recent top 100, and searching the forum for his name to show that we don't talk much about him, and this is what the forum tells me:

The following words in your search query were ignored because they are too common words: fellini.

so that hilariously failed to prove my point.

Yeah, it's intrinsically biased. If a enough people vote for a movie they love to make the list, but an equal amount of the rest of us hate that movie does it represent the feelings of the site? Weighting means movies fewer have seen, and thus may or may not like make the list, where no weighting favors commonality at the possible expense of higher regard. That isn't even including the purposeful distortion that comes in when people select a list to represent some idea, whether of balance, "true" liking, some idea of merit beyond "liking" alone, or to stand up for underappreciated movies in place of the famous stuff. No way to smooth out all those issues no matter what we do.

but i still think the point itself is probably valid. anyone can judge - i found it hard to track down all four of our past polls, so i'll link to all of them here in case anyone wants to see the results our 'standard' rules have produced:

https://letterboxd.com/m_penalosa/list/ ... nema-list/
https://letterboxd.com/fliptrotsky/list ... sary-poll/
https://letterboxd.com/pabz/list/scfz-f ... 7-edition/
https://letterboxd.com/fliptrotsky/list ... sary-poll/

poll #2 was a bit of a weird one, because a few people encouraged letterboxd folks to stop by to vote (which might have been great if it had encouraged people to stick around, though i don't think that ended up happening) so that poll least reflects our core community

Poll two doesn't look radically different than poll four really. I mean they're all "champish" for being eclectic with some continuity, but none seem all that much more decisively "true" to the site than the others off hand.

As a matter of logic I don't see how weighting or any form of restriction will make the poll reflect this community better. In fact, it would reflect it worse since honest choices are being left out. We can always tinker with the results afterwards.

greg x wrote: ↑Tue May 14, 2019 2:47 pm Poll two doesn't look radically different than poll four really. I mean they're all "champish" for being eclectic with some continuity, but none seem all that much more decisively "true" to the site than the others off hand.

i think if you look specifically at the top ten from each, and ask 'how many of these films would make some familiar canonical top 250 list from another source', poll #2 and poll #4 are very different. poll #2 still has some eclectic choices, but they're further down the list mostly.

Yeah, that's probably fair, but ye-balling it poll four overall seems the slightly less adventurous with more "famous to cinephile" and recent titles jumping out, to whatever difference that might make.

greg is talking funny in this thread and it has me worried

I don't like restrictions, or vote weighing -- too Electoral College for me. Let's just vote for what we like. At most a suggestion that we re-consider voting for the usual suspects.

Why not just create two polls? One without restrictions and one 'obscure favourites' poll.

if someone wants to hire me to run scfz polls, i'll run as many polls as people want. but i'm not tallying two of the top 100 polls in one year, it's a ton of work just doing one of them.

i suppose i misunderstood the weighting idea. i was thinking more weighting of individual ballots by voters, not weighting based on obscurity or whatever other criterion we use. i think in the past we could give more support to a top 10, then list the other 40 that were also counted but not weighted as heavily. i liked that system.

thinking on it, we could do something subtle. in the past three polls at least, we've always had tiered ballots - that's useful, or else we have a situation like in poll #1, where 45 films are tied for fifth place. in poll #2, we had just two tiers, ballots were in 10-40 split. in polls #3 and #4, we had optionally five tiers, awarding 5-4-3-2-1 points, but voters could merge tiers if they didn't want to be bothered ranking so granularly.

we'll do tiers again, but one thing we could do is introduce a rule whereby voters can put a small number of films in a top tier that awards the most points, but only if those films don't have a ton of letterboxd views. that compensates a bit for the fact that some films will miss out on votes they'd get were they more seen, without actually excluding anything from the voting.

greg is talking funny in this thread and it has me worried

Whaddya mean? Ye balling is a perfectly normal expression where I come from.

Honestly, believe it or not, I've been busy reflecting on Doris Day's career and what it sort of suggests about movie history, so I probably wasn't paying close attention to what I wrote. What can I say? I'm a weirdo.

Since it is an annual vote, I think there should be something new every year... even when my own favorites are still the same. Let's say something like this:
Year 1: No restrictions.
Year 2: Top 100 of last year must be excluded.
Year 3: Top 100 of year 1 and top 100 of year 2 must be excluded.
And so on.

heh i like angel's idea

What is the list for, what is it intended to measure? I don't see any point in generating... for example... a list of "the 100 best movies except for the ones that we liked best last year."

Maybe if The Godfather lands alongside The Eve Of Ivan Kupalo in our top ten, that doesn't mean "Godfather is best, just like Roger Ebert and all the fanboys have told us all along," but rather "Eve Of Ivan K is pretty nearly as deserving of attention as the Godfather is." Instead of excoriating our lameness in putting Mr. G. first, celebrate how cool we are for elevating Ivan K. to the same rank.

Oh you know Angel, always wanting to double or triple the canon. I guess the main pleasure with that format would mostly just be whatever pleasure there is in listing new movies each time. Which seems sorta silly on the one hand, but then again the idea of choosing 50 or 100 films as somehow better than the next group or group after that is kinda silly too as there really isn't all that much difference between the best films like that for most people I'd think. I mean I'm sure some make a concerted effort to claim some movies as representative of "bestness" they want to hold as emblematic, but that seems more a purposeful stance about movies than a meaningful claim of merit at some point.

If you ask me to give a list of "best" or "favorite" movies, for example, I'd make a representative choice, picking a selection that encompassed a wide range of eras, countries, and styles while avoiding much duplication of directors. It's artificial but pretending to pare down any other way would be just as fake since I don't really have 50 movies that I think deserve more notice than a few hundred or so others beyond a "statement" like perspective on it all.

Yeah, I would do that... the first time (and I think I did). But we are talking about an annual survey, there should be something new every year. Otherwise I would resubmit exactly the same ballot again.

I go for something more playful, that's all. Anyway, only after 10 or 20 years this would start to be really interesting.

oh snap 50/50 results shit just got reeeal

i'll be on the road the next few days, so i'll work out what to do with this in a week's time (or a bit longer). still time to cast a vote in the poll! if 'no' wins, that makes it easy i guess, we'll just run a standard poll, but if 'yes' wins, i'll start up another poll to see what kinds of "restrictions" people like and don't.