Our browse page has 2x more words than it needs to have. We'd like to move to using more concise, declarative question titles.
Example transformations:
"Will a ceasefire between Russia and Ukraine be declared within 90 days of Trump entering the office of the presidency?" => "Russia-Ukraine ceasefire in Trump's first 90 days"
"Will the first AI model that saturates Humanity's Last Exam be employable as a software engineer?" => "1st AI model to saturate Humanity's Last Exam employable as a SWE"
"In early 2028, will an AI be able to generate a full high-quality movie to a prompt?" => "Full, high-quality AI movie generator by early 2028"
"Will there be a US recession by EOY2025?" => "US recession by EOY2025"
"Will I go to the gym this week?" => "I go to the gym this week"
"Will I get a girlfriend this year?" => "I get a girlfriend this year"
'When will Manifold implement a charity program?' => "Manifold charity program by date"
"US doctors negatively affected financially by AI by EOY2025" => "" (no change - already concise and declarative)
"Substack publications at 5M+ paid subscribers by 2025" => "" (no change - already concise and declarative)
"'A Minecraft Movie' (2025) Rotten Tomatoes critics score" => "" (no change - already concise and declarative)
@Eliza yeah, to me it seems like the "declarative title" is more useful in e.g. browse tab where the title is amongst a bunch of other titles where "Who will Y" and "Will X verb Y?" are both common, especially since Y is usually the most distinctive part of the title. But then on its own a "declarative title" reads much worse since that's not how English questions are usually written. So having both a "marketing title" and a "full title" to use the right form in the right context could be the right solution.
I could see it causing a real headache with resolutions though when one is more or less ambiguous than the other (or even worse if they contradict), so would need to be clear that e.g. the "marketing title" is never to be referenced for resolution. Maybe if it were always AI generated and not editable (except by mods) that would make that clear?
Brevity is desirable, but often headline-style phrasing is difficult to parse. Beware of making the site less usable for the sake of appearing more clean.
Most of the examples are not "declarative", they are just fragments. I think this is fine in simple cases but goes off the rails quickly as one tries to fit in more complexity.
Most of the examples seem OK. These seem slightly worse to me:
"Will the first AI model that saturates Humanity's Last Exam be employable as a software engineer?" => "1st AI model to saturate Humanity's Last Exam employable as a SWE"
"In early 2028, will an AI be able to generate a full high-quality movie to a prompt?" => "Full, high-quality AI movie generator by early 2028"
"AI can generate full, high-quality movies in early 2028" would be nearly as concise and truer to the original meaning.
As a (snarky) example of why headlinespeak sucks,
'A Minecraft Movie' (2025) Rotten Tomatoes critics score
could mean
What score will 'A Minecraft Movie' (2025) receive from Rotten Tomatoes critics?
or it could mean
Will Rotten Tomatoes critics who review 'A Minecraft Movie' (2025) get laid?
Hopefully such ambiguities are clarified by the market type. But in general English has too friggin' many words that are both nouns and verbs, so headline-style fragments are rife with this kind of thing.
@jcb fwiw i don't think this is a persuasive example because by the same token, even with a full sentence it's still theoretically possible to stretch the meaning of words until you find a misinterpretation. there are of course actual examples where a declarative title has meaningful ambiguity that is resolved by a full sentence
I agree that this isn't a compelling example because the intended reading is the most natural. But it's not much of a stretch either (once I noticed I couldn't unsee it 🤦). And I think having one occurrence (even though it's a bit of a stretch) in 10 selected examples suggests a higher ambiguity rate than question titles have today.
I’m against it, although my concern is about the general homogenization of market titles if this becomes a universal enforced standard, and less that I particularly care if an individual front-page market of mine is retitled (I don't mind).
I don’t think there’s a “one-size fits all” solution, & while many creators pick bad titles, I’d still rather let people customize (& see what rises to the top) rather than enforce a specific standard.
Declarative titles usually work if the market is incredibly simple (e.g. “2025 NCAA Tournament Winner”). They also can be good when it’s actually a messy concept that’s impossible to properly capture via a title, so you shorthand with something simple & pithy which pushes people to read the description to understand it (e.g. “Trump Releases Epstein List Today?”).
I tend to dislike them when it’s something in between—sometimes adding a few more words to a title can more precisely explain the market, avoiding future misunderstandings.
But mostly I just think there are lots of small stylistic reasons that a creator might prefer a wordy title, & that's hard to codify. E.g. I recently made this market :
Will the Chinese film "Ne Zha 2" outgross "Avatar: The Way of Water" (globally) within 6 months? (>$2.3B)”
Super wordy title. Totally understand if others think it's cluttered/hard to parse (“Ne Zha 2 grosses >$2.3B” would be roughly equivalent), but it was an intentional choice that reflects my own preferences. I add "Chinese film" because people might not be familiar with the context here. I could just state ">$2.3B", but these big numbers don't mean much to most people, and I like having a reference point they can intuitively compare it to ("Wait, this movie I haven't heard anything about might outgross Avatar 2?"). The "globally" could be in the description but people ask that constantly so putting it in the title can save time. Again I don't know if this is a "good" title but it was intentional, & I generally think it's good to let creators express themselves as they choose (even if for specific cases, sure it's find to step in & change the title for some front-page market, when necessary).
(Also, while the short declarative titles Polymarket prefers have their advantages, I’ll note that they aren’t super strict about it either! E.g. at the same time as they run “2025 NFL Draft: 1st Pick”, another market”, they're also running “Who Will Make The First Pick Of The 2025 NFL Draft?”—even Polymarket doesn’t seem to think it’s worth it to force everything into a terse declarative style.)
I mirror my Metaculus questions here, and I'm already annoyed that the title length on Manifold is more constrained, since some questions can't have the same titles across platforms.
I prefer for the title and resolution criteria to correspond as closely as possible, with the resolution criteria merely providing the details that don't fit in the title.
I don't like the declarative format. Compared to the phrasing of "Will this thing happen?", the phrasing of "Thing happens" makes my brain hop through an extra hoop of "what the heck are you talking about," rather than automatically answering the question (and then following through on that impulse).
This is all culturally constructed, anyway.
Drive the change you want to see, and people will follow.
FWIW, Kalshi is way more declarative.
It causes me to read descriptions and fineprints more, since titles are insufficient to understand the resolution criteria.
It makes it more challenging to fire a quick bet, though.
You might see decreased volume.
More thoughts: I intentionally titled a few markets with more clickbaity/short titles, including https://manifold.markets/jgyou/will-there-be-a-largescale-bird-flu-810463cb11ad
This drove engagement. Some people got mad, though 😅 And moderators ended up adding more details to the title.