[Experimental] What will be the consensus crowd forecast on October 1, 2021, on whether COVID-19 originated in a lab in Wuhan, China?

Started Jun 14, 2021 11:00AM UTC
Closed Oct 01, 2021 05:59PM UTC
Topics
Seasons

Note on this question. This question is part of a pair of experimental questions. Its sister question asks about forecasters' genuine beliefs about the origin of COVID-19. That question will not be scored. Its purpose is to serve as a ground truth to be forecasted in this question. The sister question will be scored. Its purpose is to elicit forecasts about the crowd consensus for the sister question on October 1, 2021.

We recognize that the pair of questions can be gamed by providing disingenuous forecasts on the sister question to improve one's score on this question. We kindly ask that forecasters not do that. 

Related question. Related questions are on Metaculus and Hypermind.

Context. The leading hypotheses for the origin of COVID-19 are (i) animal-to-human transmission, and (ii) a leak from the Wuhan Institute of Virology ("lab-leak theory"). Two Chinese researchers stated in a paper in February 2020 that the virus likely originated in a lab in Wuhan, China. The lab-leak theory also had champions within conservative circles in the United States. The World Health Organization dismissed the lab-leak theory, however, and -- as reported by the New York Times -- the apparent politicization of the theory in the U.S. created a "fake bubble of consensus" among scientists and the media against the theory (quoting Matthew Yglesias).

The lab-leak theory is now receiving renewed attention. On May 14, 2021,18 scientists stated in a letter to the Journal of Science that the lab-leak theory is "viable." Other scientists have recently come out as less dismissive of the lab-leak theory. And on May 27, President Biden asked the U.S. intelligence community to investigate the lab-leak theory and report back to him in 90 days. (See also Washington Post timeline on how the lab-leak theory became credible.)

Resolution details. This question will resolve based on the consensus crowd forecast of its sister question as of closing, at 11:59 ET on October 1, 2021. The consensus crowd forecast is always shown under "Crowd Forecast" button. It's the mean of the 68% most-recent forecasts. The bins are up to but not including the upper boundary, meaning, for example, in the unlikely event the consensus forecast is exactly 30%, the 30%-40% bin will be correct. Foretell staff reserve the right to omit forecasts when calculating the crowd consensus if we have reason to believe they were made in bad faith. In such a case, we will communicate to forecasters that we are omitting forecasts, and the effect on the consensus forecast, as soon as we are aware of the problem.

Why are we asking these questions? If we will never know the ground truth for a question, we cannot score it in the conventional manner. Many questions of this form are of great interest, however. A new method for scoring unresolvable questions -- called "reciprocal scoring" -- treats the crowd consensus as the ground truth and scores forecasters based on how close their forecasts are to the crowd consensus. See Karger, Monrad, Mellers, Tetlock, "Accelerating the Evaluation of Policy Options in Crises: The COVID-19 Pandemic" [Forthcoming]. For a similar method, see Hypermind's Drip Rewards.

For this pair of questions, we're treating the crowd consensus on October 1, 2021 as the ground truth. This question is effectively asking two questions: (i) what new information about the origins of COVID-19 do you expect to be available by October 1, 2021, and (ii) what impact do you expect it will have on the consensus view of the crowd.

***

To suggest a change or clarification to this question, please select Request Clarification from the green gear-shaped dropdown button to the right of the question.

 

Possible Answer Correct? Final Crowd Forecast
0-10% 4%
10-20% 5%
20-30% 11%
30-40% 45%
40-50% 15%
50-60% 8%
60-70% 4%
70-80% 4%
80-90% 2%
90-100% 1%

Crowd Forecast Profile

Participation Level
Number of Forecasters 99
Average for questions older than 6 months: 58
Number of Forecasts 342
Average for questions older than 6 months: 204
Accuracy
Participants in this question vs. all forecasters better than average

Most Accurate

Relative Brier Score

1.
-0.102
3.
-0.092
4.
-0.088
5.
-0.088

Consensus Trend

Files
Tip: Mention someone by typing @username