We've run out of options, new market ?

JD is sitting on the chair behind the Trump looking like the little silly billy:

What about a phrase like "Making America great again?"

Come and get it Crypto bros

@Mana I've been filled. Hopefully I don't regret this!

@Mana Oh, there's still a tiny amount left. Basically filled though.

@Mana Completely filled now. Hmm, maybe I'll add a bit more to it.

@Mana On second thought, nah. 10k is enough.

Filtering by some term seems to break "show other answers" on the mobile app for me in Android

Is there an easy way for anyone (creator, trader, mods) to check the calibration of a market (and highly traded options in particular, maybe) at a given time (before the speech)? Should be interesting for one with so many options. I'd like to know if I'm correct to think that options tend to be overestimated on average in this kind of market because of availibility heuristics (or something)... my result will tell me a little bit about that (traded no on a dozen options, YES only on Elon / Musk) but would like to know about calibration results

@ADings what precisely would you want to measure?

(my best guess, but lmk if not)—i'll assume it's the "calibration of the market prices (perhaps 1h before start time) for the ~150 options" (i.e. bucket by market probability, check the % within each bucket that resolve YES, with ~150 data points). in that case:

• it is doable (albeit not trivially) for any user to get the market probabilities via the API, if anyone wanted to put in a bit of effort.

• my guess is the sample size would be a bit too small—calibration is super noisy due to bucketing.

• note that "market probability" is not necessarily so meaningful for this market. there's very little liquidity per option, so the less popular options (without stable limit orders) bounce around a ton (although you could do an avg over time if there's lots of activity).

• while there are a couple dozen actively traded options, the ~majority of options have very little activity (<10 traders). so whether they are well calibrated doesn't tell you much. it'd probably be more interesting to focus on the most p

• if you wanted to expand the sample size, i've run a a bunch of these markets so you actually would have a ton of options to work with.

I'd like to know if I'm correct to think that options tend to be overestimated on average in this kind of market because of availibility heuristics

fwiw, whether or not you quantify it, this is 100% true (well you can decide what heuristic leads people to overestimate it, but the probabilities tend to err on the side of being too high). the last one of these markets had some quite obvious examples of such fallacies (comment thread discussing it)—e.g.:

["Martin Luther King" or "MLK"] was at 33% while ["King" or "Prince"] was at 7% (both resolved YES when he said "Today is Martin Luther King Day"). hopefully anyone looking at that seriously would notice the error, but when skimming options it's easy to miss!

@Ziddletwix If what you want to test is “are these optioned on average overvalued”, you could

1) take the market implied probabilities

2) compute the distribution of number of “yes” answers implied by this (might need to assume independence between outcomes)

3) see where the actual number of “yes” answers falls in this distribution

I did this for the first 10 answers in this market (MAGA, greenland, canada, tariff, panama, zelensky, elon or musk, drill baby drill, AI, losers or suckers or morons, and billions and billions) and got this:

so here if actually only 4 or fewer of these answers resolved yes, that would be evidence for your thesis that the market is overvalued at p=0.017 (ie if the market probabilities were right, only 1.7% of outcomes had this many or fewer yes answers)

@brod yup i think in terms of a point estimate, taking the EV is totally fair (& requires way less sample size than trying to get some sense of calibration, while still answering the original question). but i'd stick narrowly with the point estimate—the distribution implied by that p-value requires independence, and the options in this market are absolutely not independent (not just from overlap/similar topics but also the length of the speech & etc)

(unfortunately that does makes it harder to translate the over/under-estimate into a level of confidence, you'd need to make some strong-ish assumptions to get a stderr)

@Ziddletwix yeah the real distribution might have fatter tails bc they all load on speech length, so extreme outcomes more likely

otoh maybe time spent talking about x trades off with other topics, so perhaps there’s some negative correlation to counteract it?

Thanks everyone for these highly insightful answers. Seems like there is no easy way to get a really conclusive answer. The dependentness of the amswer seems like a difficult to solve problem especially. The MLK fallacy is amazing (although anecdotal)!

.

@Ziddletwix Is this being closed before the speech? Timer says no, which is what I would prefer, but would like clarification.

@Kraalnaxx i believe sg changed the close time so it could be used during the livestream. which is totally fine with me (& i will not overrule mr ceo), i'm just hoping people see my msg & make sure to set their limit orders to time out (i'll try and TAL this afternoon if i remember & msg ppl)

i might be around to watch !

@traders looks like sg is going to use this for the manifold TV livestream & pushed the close time to be after the speech so it can be used for livetrading. i am guessing submitters don't mind much bc options were set to be so cheap, but make sure to NOT leave any limit orders up tonight!!

@Ziddletwix You are the good man, allow us to come share and add. Keep open for the trading during inauguration.

Thank you @SG !!!

Limit order at 20%, can add more

@Robincvgr I want him to do it