In December of 2026, will Wikipedia ascribe higher credence to the lab leak theory than it did in December of 2023?
36
249
710
2027
32%
chance

https://en.wikipedia.org/wiki/COVID-19_lab_leak_theory

https://en.wikipedia.org/wiki/Origin_of_COVID-19

https://en.wikipedia.org/wiki/COVID-19_misinformation

https://en.wikipedia.org/wiki/COVID-19

https://en.wikipedia.org/wiki/Wuhan_Institute_of_Virology

https://en.wikipedia.org/wiki/SARS-CoV-2

As of December 2023, these articles are relatively consistent in describing the lab leak theory as misinformation and low probability, but plausible according to a minority of scientists, and not yet definitely disproven.

In December of 2026, I'll ask Manifold users to compare the 2023 versions of those articles to the 2026 versions, and vote on whether it seems Wikipedia treats the theory as significantly more credible. Regular users get 1 vote each, moderators get 5. Suspected bot/spam accounts are ignored.

If at least 50% of the vote believes Wikipedia has started treating it as significantly more credible, this resolves YES. Otherwise it resolves NO.

Note that both "Wikipedia treats it as less credible, labeling it as completely debunked" and "no significant change" resolve to NO. This only resolves YES if there's an easily noticeable change in favor of lab leak. For example if Wikipedia in 2026 treats the subject as 50/50, giving the lab leak theory equal credence to natural origin and using neutral language to describe both, this will resolve YES.

Get Ṁ200 play money
Sort by:
predicts YES

it's wild that the FBI prefers this theory to zoonosis (although many other large parts of the usgov prefer zoonosis), but all admit that they are pretty uncertain and want more information, yet wikipedia has lab leak (even relatively innocuous "accidental lab leak of unmodified cave-collected samples") down as misinformation, and loves to lump it next to "intentional bioweapon" and "anti-chinese racism" etc.

This is a classic example of "when debate gets complex, people simply get away with directly cheating and being illogical". The same person who wouldn't deny "1+1=2" WILL deny that wikipedia has this bias, cause the argument is large enough that most of us can't remember it all in one moment. And try as you might, they'll just stubbornly refuse to "see" what you mean.

bought Ṁ10 of YES

Related market here: https://manifold.markets/Ernie/covid-origins-megamarket-lots-of-un?r=RXJuaWU where I'm attempting to figure out any angles we can use to appreciate a change

Interestingly for this market, I found that according to an LLM (and to my reading) in the year prior to Dec 2023, the "COVID lab leak" page has been getting much more confident in zoonosis and more dismissive of lab leak.

predicts YES

@Ernie especially this one:

The claim before refs 14,19 seems like a pretty wilful misunderstanding. First, those refs are to papers, neither of which in by scanning mentions other cities with laboratories studying coronaviruses. Also, this is needlessly complex. We know that wiv staff repeatedly went to the caves where the nearest ancestor to coronavirus lived and we know there were samples there, and I do not believe we know that about any other lab. So they're over general in the scope of "places a lab leak could happen" which weakens this specific claim, and very deceptively weak on the fact that lab leaks of covid-like things were much much more likely here than these unnamed other "normal" coronavirus labs.

And finally they have no discussion of how insane the claim is that

  • Zoonosis happened

  • And first showed up 10km from the only lab in China publicly known to have samples the progenitor of the virus

  • And the only lab publishing tons of gain of function research on this exact virus

  • That's just a crazy claim to ignore.

Root claim goes into lots of detail here. https://www.rootclaim.com/analysis/What-is-the-source-of-COVID-19-SARS-CoV-2

Asking Manifold is perhaps a flaw in this method. Manifold users are unusually fond of the proposal.

bought Ṁ10 of YES

@Najawin Manifold is unusually well calibrated on some topics.

bought Ṁ3 of NO

@Najawin If a big chunk of Manifold users are irrationally fond of the lab leak hypothesis, then the rest of us should be able to make good money by betting against them!

(I personally don't believe the lab leak hypothesis (though I'm very uninformed and low confidence) but I agree with Isaac that the current state of Wikipedia is really quite negative about it - more than I would have expected. It seems plausible that both:

  • The lab leak hypothesis is false

  • Wikipedia will be more sympathetic to the lab leak hypothesis in three years' time)

@Najawin I think Manifold users tend to be smart enough to understand the difference between their own beliefs and the espoused beliefs of an external platform.

@NicoDelon If by "well calibrated" you mean "have oricular powers that transcend listening to the experts and agree with my preferred conclusion", I can see why one might believe this.

@IsaacKing I've noticed precisely the reverse. People are just as likely to engage in motivated reasoning about what other sources say for the purposes of the gamification here as they are on any other topic.

predicts YES

@Najawin Oh ‘the experts’, how have they been recently? Hadn’t heard of them in a while.

predicts YES

@NicoDelon Good calibration doesn't preclude bias. I could ascribe 0.01% probability to the lab leak hypothesis and still show up as well calibrated as long as I make another 9,999 bets at that probability on more certain things.

predicts YES

@IsaacKing Sure. We just don't know that about the lab leak yet. The market hasn't been externally validated yet.

@NicoDelon You've not heard of biologists looking into the issue at hand? Well that would certainly explain why you're suggesting that this site is well calibrated on this issue.

Look, I'm not certain how this is even controversial here. It's oft quoted that markets can remain irrational than you can remain solvent. I agree that in some idealized circumstances these markets would trend towards truth. But we just don't have those idealized circumstances, and Manifold in particular has some very specific slants in terms of views.

predicts YES

@Najawin Biology? Never heard of it.