|
In an email from an Israeli journalist named Gadi Guy,
doing an interview for a Hebrew web magazine called "Weekend", he
asked:
5. On January 5, 2003, at approximately 17:30pm GMT,
a bomb exploded in a crowded market in Tel Aviv,
killing more than 20 people and injuring over 70,
causing widespread fear and anger all over the country.
Has this event registered on any of your sensors,
especially the one in Israel?
I said I would put this in the list of predictions, with him as the
source, and do an
analysis. I described the process of making a formal prediction, and
indicated that we would specify an interval of time around
the event similar to those
specified for previous, similar events. I said that I would explore
more broadly, but the formal prediction would be for three hours
beginning at 17:00, half an hour before the explosion.
In addition to the formal analysis using data from all eggs, I indicated
that I would look at the egg in Israel, hosted in a town not far from Tel Aviv.
The formal analysis, using data from all 52 eggs showed no strong trend,
with a Chisquare of 10782 on 10800 degrees of freedom, and a p-value of
0.547, essentially a null result.
The single Israeli egg, number 2225, was unfortunately off at the time
of the bombing, but came on one hour later. The graph of the data from
egg 2225 for the rest of the day
showed something far more interesting than the full network,
as you can see in the second
graph below. The strong downward trend that begins at the time of the
bombing has a probability (if this were a formal analysis) of 0.977
(equivalent odds of about 2 in 100).
This suggests an effect driving the data to small deviations.
My aesthetic, speculative impression is of dismay. No longer
surprising, these bombings are devastating to the soul.
|