Details of judgment
All "wikipedia pages state X" claims require the change to still be present 30 days after initially made, and for edit wars to have settled. So in reality a change might be made, then reverted, then ignored, then after 29 days there's an edit war judgment that goes on for 6 months. I'm sorry for your mana but it's better to just wait. Truth will out
Dates: "Mid X" means midnight, June 30, that year, California time.
LLM judgment: I'll try to use the best public LLM / ai assistant, with public data, hopefully preferring gpt-4+ from OpenAI or successor orgs, available for $100/month or less, and conforming to "mainstream" conventional US reality (no "super based" home-running only, "illegal in EU" models, even if they do see more "true" sometimes)
LLM conversation usually requires some variation of the prompt and repeating of tests so you are sure it's answering what you are asking. If these are contentious I'll do them on stream/publicly/at set times/with set seeds or whatever is needed. I'm trying not to bias the market too much here, but some human interaction is needed
Also, if we switch models, we have to re-validate prior numbesr. i.e. if today's GPT-4 says the current page is 85-95% confident, and we're judging against that page, we need to reevaluate the page with whatever future GPT we use, and also revalidate the claims the entire premise of the question was made upon. I'll do what seems logical here; for example in B, it seems the question is whether "the toughening against lab leak in 2023, since 2022, reverts back to even weaker levels", we'd need to confirm future LLMs see this same toughening of views over the time period.
All claims are judged from the moment of creation; if they did it before, that one doesn't count, but if they do it again, it would count.
"Coming out for lab leak" means saying something like "we have to take lab leaks very seriously" or "it was a lab leak". As long as they say/write/post it, it counts, even if they are forced to withdraw later.
Big picture: I want to resolve things based on what seems "actually meaningful and useful" at the time of writing and resolving. So if there is a cheap trick corner case which ruins a market in a non-useful way, I'll hope not to do that. "Hard cases make bad law". I'm trying to do something sensible and meaningful. if you think you've found a corner case and then bet a ton on it, then try to change judgement at the last minute, please reconsider this strategy and talk to me (privately or publicly) first to let me clarify. It's just not interesting when long-lasting markets resolve meaninglessly.
General warnings:
for claims related to wikipedia, it's a huge pain to monitor the state daily. In cases like that, we'll just come back a week later and see of something has been settled.
My views
Unfortunately, I'm already 90%+ believing that it was a lab leak. The allegedly "unconnected" location of this "random zoonosis, in no way related to WIV", just so happening to be.... 10km from WIV, a lab closely related to COVID research, holding COVID-19 precursor samples, deeply connected with Fauci, Dazack, Shi Zhengli - is just absurdly near by.
If zoonosis happened, it means WIV had nothing to do with it. So why did zoonosis allegedly happen right next to WIV? If it's unconnected, out of all the rest of china, including... hundreds of wet markets closer to the caves of origin... why did it "pick" the one wet market right next to the [completely unrelated] WIV?
(I know the virus did not "pick" the location, I'm speaking informationally. In a world where zoonosis happens, the place the zoonosis outbreak occurs is the product of some kind of distribution function which contains info on population travel and density, cultural practices, distance from the caves, language, etc. Zoonosis proponents claim that this random function "just so happened to pick a spot right next to WIV". But I claim that it "knew", that is, it was NOT a random function, and in fact the function was actually on the leak rate from an insecure lab, not a function of bat-collection/etc. And since it's a leak rate of individual travel from a lab, among workers, the distance was small since they all lived in Wuhan, which had big and fast public trans, so going 10km was not a big deal at all.)
I'll try not to let this interfere with my judgment but speaking honestly it's nearly impossible to imagine how my view could change here. That is a weakness in my logic, I realize, and would like to be corrected on it. My main probabilistic argument could produce this outcome, after all, under zoonosis.
I'm creating this because I'm hoping that in the long term, as relevant people pass on to other worlds and other jobs, that more information will leak out and at least it'll be widely accepted that it was a lab leak. In general many of the claims are about public perception, anyway, rather than scientific reality.
Timing: I'm trying to make the timelines short, but there's a risk everything just NOs due to Biden / Trump still being in charge. Still, once Fauci goes things may change, and post 2028 I expect a shift. If claims expire but still seem possible, I will recreate them again with shorter or longer timelines if asked.
Submissions: if you're interested, please submit any ideas and I'll do my best to add them. I'm really trying to defeat my somewhat irrational certainty here, so any edge to cut into it would be nice, particularly in looking for positive evidence against lab leak. (rather than seeking confirming evidence).
Running the market
This is a big interest of mine, so I plan to add more markets to this one, and resolve others as they come due. I wanted to create this to lump a bunch of markets together, to share rules, definitions, and judgment criteria, and have an easy place to dump one-off or interesting ideas we could bet on related to COVID. I expect MC software support to continue to improve, which will make this even better (sort by traders please!, and more columns in general!)
I will not bet except on items that are extremely easy to verify (Fauci passing away, for example; I won't bet on claims like "X supports lab leak" since they are possibly contentious.
Links and Data
A: https://en.wikipedia.org/wiki/SARS-CoV-2#Reservoir_and_origin
B: https://en.wikipedia.org/wiki/COVID-19_lab_leak_theory
![](https://firebasestorage.googleapis.com/v0/b/mantic-markets.appspot.com/o/user-images%2Fdefault%2FelKuF9wDYn.png?alt=media&token=07731bd4-f9b3-485c-9162-c441960c63f5)
S:
![](https://firebasestorage.googleapis.com/v0/b/mantic-markets.appspot.com/o/user-images%2Fdefault%2Fz9nzMBDxpq.png?alt=media&token=b28ba0ed-5cee-4354-9a67-b826e6f073c6)
T: "Do you think covid-19 came from a zoonotic origin? Give percentage estimates based on your opinion for each of the top 5 theories: "zoonosis", "accidental lab leak of modified virus", "accidental lab leak of unmodified, collected virus", "intentional leak or other complex false flag operation", or "other" and list what you include in other. For each, specify your estimate of probability in general, 0-100%, along with a short list of reasons. Go."
Answers as of 2023/12/24, five trials
Zoonotic Origin 72% (70, 70-80, 60-70, 70-80, 70-80):
Accidental Lab Leak of Modified Virus 7% (5, 5-10, 5-10, 5-10, 5-10 ):
Accidental Lab Leak of Unmodified, Collected Virus 12.5% (10, 10-15, 10-20, 10-15, 10-15):
Intentional Leak or Other Complex False Flag Operation 2.4% (2, <5, <5, <5, <5)
Other 7.2% (13, 1-5, 5-15, <5, 5-10):
Averages treat ranges as the center, and treat "<5%" as 2.5.
Totals do not add up to 100% because GPT
https://chat.openai.com/share/3d0f23c4-b227-46b7-99bc-003b528706d8
My other markets about COVID which someday I'd like to incorporate here!