An edit-off is an interesting concept, but it's a pretty big turn-off if the editors can't write their own entries. That's really going to thin down the pool of prospective editors.

Re: the edited event idea, maybe we could make editing someone else's story part of the event. Like, for instance, you'd put in your submission and it'd get looked at by a randomly chosen fellow contestant while you also take a swing at another guy's work. Or, to simplify things, everyone gets paired up and we swap stories (with a possible three way swap in case of odd entries). That way, you'd be able to participate in editing without losing your ability to participate.

3916576
What if people entered as teams, with both the writer and editor getting credit for the win? (either sharing it or some division that seems fair.) That would attract some people to edit as a way to still get points and credit while not having to dedicate the time to write, while other people can write knowing their editor of choice will be backing them up on making their story suck less.

3924951
It is mostly just an idea; I'm not sure that it is really a good one for the reasons you and others have outlined here. Right now, we sort of have a whole bunch of people basically contributing in this manner already via reviews.

It seems like a neat idea, but I'm not so sure if it really is as neat as it sounds off the bat. A lot of reviewing is iterative anyway, which probably works against the idea anyway.

Ah well. Fun to contemplate, at least.

EDIT: Also, next weekend is another writeoff. I'm excited! Been a while since we had a minific writeoff.

I assume the next event will post on the site tonight if we are sticking to the Jan 3 minific contest?

Will it be a single prompt submission again or have you implemented that graduated scale idea where you can post a few that start truncating the longer the submission list gets, pairing down to your personal #1?

3933013
I think I vaguely recall the idea being floated. It's not that necessary since there are more than enough prompts with everyone submitting only 1.

Since we seem to be experiencing an influx of participants, I wonder if it's time to start using the prelim round again.

Currently, the way it works is it eliminates approximately half of the entries. This doesn't exactly scale well because it's still O(n). If instead the prelim round cuts off at a fixed amount of entries, no matter how many participants we have, reading through all of the entries that passed the prelim round should be doable.

3943006
Wow, I don't even remember that being a thing. D:

3943006 Why not halve the number of stories in the prelim rounds until it lies within some acceptable range? (Having never been in a write-off with a prelim round, I may be missing something here.)

3943757
That would take too long.

3943826 Computationally or voting-wise? Because I'm still conceiving of a single elimination round, but with the size cutoff variable so that the voting can be partitioned as evenly as possible.

3944105
I'm not sure I understand what you're saying.

The prelim round gives a score to each entry. The list of entries can then be cut in any way desired. The current implementation is to cut off at score >= 0. Since the scores are zero-sum, this has approximately half of the entries make the cut.

I find it would be more in line with the goal of scaling work load for readers to cut off to a fixed number of entries (or, more specifically, fixed approximated work load for a voter).

3944155 OK, I'm confused too, over a bunch of little stuff, so I think I would have to go through a writeoff with a prelim round to grok what's being discussed.

3943006
If I understand right:

Basically the idea here is that each person is displayed some number of entries (say, 20) which they are expected to vote on, assigned at random to each participant?

You can vote on more than the stories that you have assigned to you, but you must vote on all the stories you were assigned to have your vote counted, yes?

And then, after some number of days, the story list is curbed down to the top X stories, which then get graded by everyone?

3953289
Not quite.

You are assigned S (the number of stories you submitted) groups of N (currently 6) stories and have to sort them from best to worst. You must perform this sorting for your stories to pass the prelim round. You may optionally request additional groups of N stories to sort.

The sorting gives stories a score of 1 + N - 2i, where i is the position of the item. (For example, the story you rank 2nd will get 1 + 6 - 2*2 = 3 points.) This score is used to determine which stories pass the prelim round. Currently, stories pass with score >= 0. I will likely change this to have a fixed number of stories pass instead (or more specifically, a fixed estimated voter work load).

After that, things proceed as normal with the reduced number of entries.

3953319
Makes sense.

How long does the preliminary round last?

3953331
However long I set it to. In the past it's been 7 days.

The next round is going to have preliminaries. For those of you familiar with the prelim rounds, it's being changed slightly. An explanation of the intended implementation is below.

The purpose of the preliminary round is to scale the write-off for any number of participants. This is done by enlisting the help of submitters to pare the list of submissions down to a reasonable size.

Since we're concerned about the time it takes for readers to read and consider the entries, there should be some approximation of this for the average reader. I'm going with `work = 10 + words/200` as a starting point, meaning it takes about 10 minutes, plus a minute every 200 words, to read and review a story. This definition is up for debate (i.e., I'd like to hear opinions on whether this is a good approximation).

The prelim round will seek to reduce the list of entries to a certain work threshold. I'll define this threshold with consideration to past write-offs that people have said were too much work. The two that come to mind are Cutting Ties (which resulted in the initial use of the prelim round) and All In which have a total work of 1046 and 1199 minutes (work for past events). With this under consideration, I'll say that anything over 900 minutes is too much (induces reviewer fatigue), so events with over 900 minutes work will have a prelim round in them.

Having a prelim round only to pare down to 900 minutes work would be fairly pointless in most cases, though. (Spending a week to eliminate three or four entries is a good waste of time.) Instead, we'll say that we want the prelim round to eliminate at least a third of the work. In other words, the prelim round will pare the entries down to at most 600 minutes work.

The final thing to consider is how much work to assign to to each prelim record. When the prelim round was initially devised, I set the number at 6 stories. This was with consideration to Cutting Ties data, which averaged at about 30,000 words per 6 stories. Using the above work definition, this totals about 210 minutes work, which I think is a good number. In other words, each submitter will need to do about 210 minutes work (per submission) in the prelim round for their story to be able to pass the prelim round.

3981159
My only concern is we'll find out that the hugeity of All In scared off a bunch of people, and that combined with short story entries forcing fewer submissions per person will mean that we'll have significantly less than 600 hours of work, as you put it, and leave the prelim round moot. Do you have a way to set it up dynamically, or do you feel it's more important that we plan to have them?

3981815
If the work for the event is less than 900, then the prelim round will just not happen and things will proceed as usual.

3981904
Oh. I guess that is thoroughly reasonable. :B I think 6 stories is about right, then, and everything else in your proposal is 2math2+2me.

3981159
3981948
Okay so I did some mathies (I'm a bit rusty), and while I initially really liked the work equation it looks like it has some bad side effects. Here's my Excel sheet on it.

The way I saw the equation as breaking down (I may be wrong here), was that the "Word Count/200" is the estimated time it'll take someone to read a story. 200 is a reading speed: 200 words/min. So a two-thousand word story would take about ten minutes at this speed. The equation then adds ten minutes to this number for what I assume is reviewing, giving the total estimated time spent on both reviewing and reading. Total Work Time for a story.

But the added ten minutes for reviewing is a constant here. This means, then, that we only ever have ten minutes to write a review, whether the story is 600 words or 6000. So as a story's length increases, our allotted time to think and write about it shortens, and does so exponentially (see the excel sheet for a graph). This is obviously not good, since the longer a story is, the more time we ought to have to think and write about it. More content means more to write about.

Setting up the preliminary rounds this way means that for a 2000 word fic, 50% of our time is expected to be spent on consideration and reviewing; for an 8000 word story, we're expected to spend only 20%.

I'm not exactly sure how to fix this, but I imagine that turning the added ten minutes into a variable also based on word count would help. Perhaps for every five 500 we get another ten minutes? For every 1000? So something like "Work=(word count/1000)*10 + ...." which just equals "(word count/100)". Thus, a one thousand word story gives ten minutes for reviewing (perhaps added to the already existing ten minutes?).

Or maybe we should set it up so that we always get 50% of a our total work time reviewing.

Or maybe I'm thinking too hard about this.

The original equation works out fine for minific rounds, though. There, assuming a 750 word count, we get a little under 14 minutes total, with reviewing taking up about 70% of that time, which seems fair to me.

The other thing I'm worried about (but haven't tried fiddling with the numbers on yet) is whether the Work Time Reduction (so Taking 900 minutes of work and reducing to 600, the whole point of the preliminary round, which is definitely necessary) will unfairly favor shorter stories over longer ones, or vice versa. Will a longer story get dropped from the preliminary round if it pushes the Total Work Time over 600 by only ten minutes or so because it was 7000 words instead of 2500, no matter its quality? A writeoff composed of longer entries will have fewer stories pass the prelim round than a writeoff composed mostly of shorter ones, all based on wordcount. But maybe that's not such a bad thing?

I dunno. Maybe we can employ the Work Time equation when it comes to considering whether we should have a prelimary round or not (I think it serves this purpose very well), but we don't govern the actual prelim round with it, instead determining what passes through by some fraction of the total number of entries. So in other words, we say "if the Work Time for a writeoff exceeds 900 minutes, we're having a prelim round" and for this we use the equation; if the equation tells us the Work Time exceeds the limit, we then say "The Preliminary round will cut out 50% of the stories (or 40%, or whatever). This way, each story is represented equally, as one whole instead of as sticks of different lengths, where we can only allow in a certain total amount of footage.

I dunno does that make sense?

3983590
What part of 2math2+2me did I not make clear? >:B

3983594
I tried to explain things! D:

3983590

Without touching on the math, I'm not sure Roger's assumption is an unreasonable one. In my experience, writing a review usually takes longer for long stories than for short stories... but not a lot longer, and in any case it's not the thing that correlates most strongly to how long a review will take.

I think I'm okay with an assumed flat ten, personally, but if other people are concerned about reviewing time, a better workaround might be different flat lengths for events (e.g. 10 minutes for minifics, 20 minutes for short stories. Or whatever--numbers used for example purposes only). If ten minutes is what I need to review a 750 word minific, it definitely does NOT follow that I need more than an hour and a half to review a 7,500 word short story!

3983715

in any case it's not the thing that correlates most strongly to how long a review will take.

That's certainly true. Story quality has a lot to do with it too, along with how thought provoking it may be or clear vs ambiguous. I'll probably say more about a bad story than I will about one that really impressed me, simply because for a story you think does everything right, there probably isn't much else is to say but "i loved it for such and such reasons". Though really, I suppose it's most affected by how in depth you wish to go in your reviews. Some people write very short ones, so the proposed system would work just fine for them.

Personally, ten minutes is barely long enough for me to review a minific, but that's just me; I'm a slow reader and a slower writer. I'm very happy to find some average everyone else if fine with, if everyone thinks it's better to just stick with a constant review time as opposed to one that increases with story length. ^.^

I'm actually kinda more worried about the system somehow biasing the results of the preliminary round in one direction or the other, due to it ultimately being based on word count; to be honest I think figuring that out is more important than hitting the bulls-eye on exactly how much work a given story will take to read and review--we can mess up on that and take a few tries to get it right, but cutting people's stories out unfairly, not so much. But perhaps my fears are unfounded. *shrugs*

3983590

200 is a reading speed: 200 words/min.

More-or-less. I noticed that while All In had considerably less total word count than most short story contests, it induced more reviewer fatigue. It follows that people spend a greater proportion of their time reviewing than reading for shorter entries, and vice versa.

However, 200 is not just reading speed. To get people's average reading speed I did a quick search and found:

The average adult reads prose text at 250 to 300 words per minute. While proofreading materials, people are able to read at 200 wpm on paper, and 180 wpm on a monitor.[0]

Slowing the reading speed down from 250–300 to 200 essentially allots some[1] of your reading time to proof reading (i.e., reviewing). So you do get slightly more time to review entries the longer they get. If you wanted to allot more time to reviewing the longer a story gets, then simply making the factor less than 200 would do that.

I came to the number mainly by considering the amount of reviewer fatigue expressed across events. I tried different values and checked it against the events to see which gave total work loads that lined up most with reviewer fatigue. The best values were in the range 180–200 (coincidentally in line with the above source's approximation of proof reading speed).

[0]: 1998 study, see http://en.wikipedia.org/wiki/Words_per_minute
[1]: If we say the average reader reads at 275 wpm, then a factor of 200 allots 1 - 200/275 = 27% of your reading time to reviewing. This doesn't include the base 10 minutes allotted.

3984107

If we say the average reader reads at 275 wpm, then a factor of 200 allots 1 - 200/275 = 27% of your reading time to reviewing. This doesn't include the base 10 minutes allotted.

Yeah, I totally agree that our reading speeds are lower while reading entries because we're also proofreading and reviewing in our heads, and this causes us to read slower. This also means that some of the reviewing "work" goes on while we read, as you just said. But I simply considered most of it to happen afterwards (though maybe I'm wrong; you did say that with longer stories most work seems to be spent on reading and not reviewing, which makes sense). Anyway, I worried that for longer stories, ten minutes might not be an accurate representation of our true workload (more thinking after reading + writing), and thus the equation would be underestimating it. Though by how much is difficult to say, if I'm even right in my thinking here.

However, I certainly trust the 200 wpm; and I also trust your conclusions that the equation seems to fit previous writeoffs where exhaustion occurred. Even if it didn't accurately predict our true work load, if it predicted which writeoffs would even cause exhaustion and which didn't, it would be useful enough.

What do you think about possible bias during the prelim round towards longer or shorter stories--think my fears may be unfounded? I'm personally not sure. Haven't tried any math on it yet, if I even could.

3984400

What do you think about possible bias during the prelim round towards longer or shorter stories?

This is an appropriate concern. The bias would occur regarding the last qualifying entry. Example:

We've taken 8 stories to total 570 minutes work. The next story has a calculated 35 minutes work. It won't qualify, because that would have the qualified entries exceed 600 minutes work. If it had less than 30 minutes work (i.e., if it were shorter), it would qualify.

To avoid this bias, instead of taking stories up to a maximum of 600 minutes work, we can take stories until the total work of qualifying entries exceeds 600 minutes.

3984437

instead of taking stories up to a maximum of 600 minutes work, we can take stories until the total work of qualifying entries exceeds 600 minutes.

Sorry, but maybe because it's late for me I'm having difficulty understanding the difference here. Could you clarify?

3984452
Say that the thing is presently at 570 minutes of work after, say, 19 stories have been added in.

If it capped at 600, then if the 20th best story took 30 minutes or less to do (i.e. was 4,000 words or less), it would be included, otherwise it would be excluded.

If it caps at the first story to exceed 600, then it doesn't matter how long the 20th best story is in, so even if the 20th story was 8,000 words long (estimated time to read would be 50 minutes) it would still get in.

I think that this is a much better way of doing things, as it doesn't penalize stories for being longer.

3984461
Ah, okay, thanks TD. ^.^

Anyway Roger that seems like a fine solution to me.

I think your numbers look pretty good, though I do generally endeavor to read ALL THE STORIES.

I didn't actually time how long it took me to do it last time, though. Maybe I should, just for reference.

Ah well.

Having never been in an event with a prelim round, I'm actually excited here.

I think the work calculations above seem fair and balanced. Would dynamic allocation of stories based on the work only matter for the cutoff for the finalist list, or would that also factor into the prelim reading list count as well?

I plan on reading all the stories, like always, even if we get a ton of them. Will this be possible through the prelim system, or will I have to wait for the full list to populate during/after the finals?

Has there been any testing with the prelim code to determine if it could undermine the author ambiguity like last round? Has the issue that surfaced in the last round been corrected?

And just to sneak this in here to be lazy, would it be possible to implement a select box/drop down on the submission page to show fimfic-like categories? I'm thinking 3 boxes with all possible categories listed in each, as well as a blank, so that the author can tag the story type(s). Hopefully these items could be displayed at the top of the entry under the title, and also be checked to make sure they don't display the author in any download form until after the voting is done. I think it would go a long way to set the mood for the story seeing as I tend to get a jarring experience bouncing between comedy, romance, and tragic stories as I read through the list. It hurts the stories based on their placement when it takes me half the words just to get in the right mindset for what I'm reading. Also, it could help Bad Horse determine which ones he's going to give a -1 to without having to think about it too much.

3999279

Will this be possible through the prelim system, or will I have to wait for the full list to populate during/after the finals?

The fic gallery opens when the prelim round starts and looks as it usually does. After the prelim round is finished, the public candidates are separated from the other entries (example),

If you intended to read all the entries, you can request additional prelim records to fill out after you've done your assigned ones. You can keep doing this until you've rated all the stories that aren't yours.

Has there been any testing with the prelim code to determine if it could undermine the author ambiguity like last round? Has the issue that surfaced in the last round been corrected?

The nature of bugs leads me unable to make any guarantees, but the previous bug has been fixed.

And just to sneak this in here to be lazy, would it be possible to implement a select box/drop down on the submission page to show fimfic-like categories?

The absence of tags or any kind of categorisation metadata is intended, for similar reasons that anonymity is important.

3999378

If you intended to read all the entries, you can request additional prelim records to fill out after you've done your assigned ones. You can keep doing this until you've rated all the stories that aren't yours.

Curiosity again -- if I keep requesting more batches, will they be of equal numbers, even for a final batch? Will they randomly sort without precedent so that I might see duplicates in successive lists? Do the prelim scores correlate at all to the final voting round or carry over in any way? (I would assume not on that last item)

Is all of this in a FAQ somewhere that I'm just too lazy to find or read?

3999675
You'll get equal numbers until the last lot which will be the remaining amount. You won't see any duplicates.

The FAQ needs updating.

3981159
Pre-post disclaimer: I am an outlier, because I write crazy long reviews, but.

This all sounds good, but I'd like to advocate being more conservative with the numbers. I remember feeling some burn as early as "There Is Magic In Everything," and the 51 entries of "Famous Last Words" had me scrambling for an alternative review style. The 93 entries of "All In" had me in full emergency mode, slashing my reviews to the bone just so I could give everyone something; that was ludicrously high rather than just uncomfortably high. Injecting preliminaries at the current threshold you propose won't pull it back into my comfort zone, it'll just stop the bleeding.

Using your numbers (10+(wc/200)) and the list you linked, I'd say my starting-to-burn threshold is closer to 600 work-minutes. But I think it's more accurate to say that my $base_time_per_story is closer to 20 than to 10; that would have put "Famous Last Words" at around 1200 horizon-work-minutes and "There Is Magic In Everything" at around 900 hwm, and "All In" in the 2000+ range if I'd done my traditional reviews. Using the revised formula of (20+(wc/200)) to calculate hwm, I'm comfortable saying that 900 hwm still feels like a reasonable burn threshold, and getting the finalists down to 600 hwm (plus the work of the preliminary round) feels like a much more comfortable amount of volunteer time.

600 hwm is approximately 25 minifics or 15 short stories, which intuitively seems like a reasonable number of finalists.

600 roger-work-minutes (original formula) would give us ~45 minifics or 20 short stories in the finals, which still seems high, especially for the minifics. (I know one of my silent lurker friends who had judged the writeoffs back in the days before the prize rounds gave it a shot in the arm noped out of here when it first hit the 40-fic threshold.)

4003131
Ultimately it's a balancing act. We want to allow as much as we can handle.

To get to the right numbers, we have to start at what we want. You end there, with:

25 minifics or 15 short stories

Let M be the desired number of qualified minifics
Let m be the average wc of minifics
Let S be the desired number of qualified short stories
Let s be the average wc of short stories
Let b be base time to review a story
Let r be the time to read a story per word

M * (m / r + b) = S * (s / r + b)

m and s are known constants, approximately 670 and 4570 respectively. This gives a proportionality between r and b:

I think 200 is the correct number for r, so with M = 25 and S = 15, b = 26.

I'm not convinced 25 is the right number for M though. I think 35 qualified minifics is a better amount. With S = 15, M = 35, and r = 200, b = 11.3, which is pretty close to my proposed 10.

4003576

I'm not convinced 25 is the right number for M though. I think 35 qualified minifics is a better amount. With S = 15, M = 35, and r = 200, b = 11.3, which is pretty close to my proposed 10.

That's reasonable, but: M is the winnowing number, and NOT the upper limit of fics we think reviewers can handle. M is the number of fics which results in w=600. Preliminaries don't happen until w=900, which means that under your proposal, a given round can have over 60 minifics* before the preliminaries kick in to correct it to M=35.

Keep in mind that this also only measures the public finals. Entrants will have an additional workload.

I recognize that I am only one data point, and I really hope some other reviewers chip in, but I've already stated (in the All In thread) that I will no longer be providing my usual reviews for minific rounds where the number of stories I have to review is greater than 40. M=35 plus 6 in my preliminary round (plus 6 more for each additional fic I submit) already breaks me, not to mention the case where w is calculated between 600 and 900 and thus we have "overflow" past 35 entries without a preliminary round kicking in. That case concerns me.

--
* "Famous Last Words": 35,138 words, 51 entries. Even spotting you b=11.3 instead of b=10, this is w=(175.7+576.3)=752. Additional stories add (670/200+11.3)=14.7 each, so an additional 10 stories gives you w=899 at 61 entries.

Non-math tl;dr of 4003949 for 3983594:

"4003576 is arguing for balancing the numbers in such a way that a minific competition's preliminaries would reduce the number of finalists to approximately 35 45.* I'm arguing that minific preliminaries should reduce the number of finalists to approximately 25, and pointing out that we might end up with more than our desired number if there are more stories than our target but not enough stories to make the preliminary round happen. We both agree that ~15 finalists is a good number if a short story competition has a preliminary round. I said that ~15 finalists is a good number if a short story competition has a preliminary round, and he didn't seem to disagree."

--
* Note to 4003576: This is what b=10 gives you. Even b=11.3 results in 41 finalists, not 35, so I think your math might be a little off.
w = M * (670/200 + b). For w=600 and b=10, M = 600/13.35 = 45.

4004003
If you are looking to limit the finals in a minific to 25, which I believe it reasonable, then the prelims should kick in somewhere between half of that number in overage to account for having a prelim round mean anything. This will indeed make the possibility that the count could end up being 35-40 minifics in a final without a prelim, but I don't see that happening in the current mix. Perhaps January was a fluke. 93 fics is quite a lot, but we had 59 different authors. Even if everyone submitted only one story that would far exceed the threshold for a prelim.

Since the average length of short stories is about 6-8x longer than a minific, it would seem to need to account for a much smaller pool in final voting. That's only considering the word count. Since the bulk of time we are measuring seems to sway toward the reviewing portion, this relation doesn't hold true. I would wager that having 20 finalists in a short story group might be on par with the 25 minific count. If even half of the previous minific round's authors contributed a story, we would be kicking in the prelim more around 30. This is the same number of stories from Behind Closed Doors. Title Drop (polished event) had all of 21 entries over 10 days. Most everyone seemed comfortable with that level of reviewing. Given than some portion of your prelim reviews should end up in the finals, I would say you should see around 25 stories total if you weren't out to read/review everything.

I feel that the maths work out for both points, statistically and realistically. Finding that happy medium might take a little flexing. My middle of the road numbers look like this:

Minific = 25 finalists, start prelim at 40
Short Story = 20 finalists, start prelim at 30.

Another thing to note is how long we want to give for prelim reading/voting? Should this be an entire week? Reading six stories or so, even of the short story variety, shouldn't take more than 2-3 days. Granted people can't always devote time to this right away in some cases. Also, nothing stops those who want to read ahead from digging into another batch. For those that only read their batch, does it make sense to speed up the prelim round? Should the entire voting block be segmented into a front and back half of the same week (3 prelim/4 finals)?

I'm sort of partial to the idea of having the prelim go three days and the rest of the voting go another 5 or six. That pushes us out into the 8-9 day range on the voting portion. For those that only put in the minimum read/vote count, the work remains pretty even. For others that read everything, the time fluctuates more with participation levels. For those that review it all, hooray! I hope we don't break you.

4004003
The cut off value for w is arbitrary and likely not 600. We calculate w from this, e.g., w = 35 * (670/200 + 11.275) = 15 * (4570/200 + 11.275) = 511.875. If you assume w = 600 and calculate b from that, M * (m/r + b) != S * (s/r + b), which was the point.

We can choose M and S to mean the number of entries such that a prelim round occurs (and cuts off to 2/3s of these values) or number of desired qualified entries. It makes no difference. (Rather, the difference is exactly a constant factor of 2/3.) For the sake of argument, I'll stick with the original definition I used: the desired number of qualified entries, given a prelim round. A prelim round will occur at approx 150% this number of entries (since a prelim round seeks to eliminate at least 1/3 of entries).

The question is, what are desired values of M and S? I think we can agree on r = 200. I think M = 35 and S = 15 are good choices. 53 minifics or 23 short stories is not particularly egregious, and they are worst case scenarios.

4005104

I look at the S value and think that seems a bit low. S could go up to 20 for me and I still feel good with the numbers. Over 30 kicks in a prelim for a SS round.

M looks good with a prelim kicking in around 50. It could be a little lower, but I'm okay with it sitting at that level.

4005104
Oh, okay, I had assumed w was fixed, but your way does make sense, since w was only ever a value to store assumptions in, and talking about the number of finalists is a much more concrete way of discussing the idea.

After sleeping on it a bit, I'm willing to give 35/15 a shot, and see how it plays out in the next minific round (for me, that'll be a better stress test of it than the short story round, because my time spent skews toward reviewing rather than reading).

I think that 35/15 sounds entirely reasonable and managable, with the worst case scenarios being something along the lines of Famous Last Words or Title Drop, which were not onerous.

Will the prompt submissions be going live tonight?

Also, has the length of the prelim round been decided on?

4014894
It's up now.

7 days.

4015022
Thank you sir!

I /salute your dedication!

4015022
Since it came up in the main thread:

Does the extra 7 days for the preliminary round push back the start of the next competition by an additional 7 days? (I kinda hope so.) In other words, are you trying to maintain the 28-day cycle, except the writeoff is happening for more of those days, or does this push it to ~14 days out of 35?