|
Peter Bancel wrote on Fri, 13 Jul 2001 12:55:13 +0100:
"The site for the 2008 Olympics will be announced at noon Princeton
time. This would be a good GCP prediction. There will be
a billion Chinese following this intently."
This was accepted as a formal prediction, but required more
specification. It would be a point event, for which the GCP's standard
treatment predicts deviation in the data from a period surrounding
the event. As specified in the registry:
"The formal prediction is for a deviation during a half-hour period
surrounding the moment of the formal announcement, using
12:00 noon as the point unless there is a significant change in the
time for the event as determined from news coverage. This
corresponds to the period 15:45 to 16:15, UTC. We will use seconds
resolution for the analysis and specify medium expectation."
In fact, the decision was made and announced about two hours earlier,
according to news sources.
The CNN story announcing that Beijing had won was posted at
15:38 GMT, and the BBC story at 14:22 GMT, which
implies that the announcement itself was made probably at about
14:00 give or take a few minutes. Thus the prediction period was
moved back by two hours, setting the period from 13:45 to 14:15 UTC
for definition of the formal dataset.
The graph shows the data from 31 reporting eggs for this period of
time, with 14:00 marked to identify the
likely time of the announcement. The deviation is positive, with a
fairly steady trend and a final Chisquare of 1870.5 on 1800 degrees of
freedom, for a probability of 0.121.
|