On the day I read about the challenge, I immediately thought of a simple split of the prize money. I registered a website, and started to put comments on every news site that I could find. The next day I continued to make comments, while I let the domain forward to the official DARPA site. I started to flesh out the basic operation of the site, and had no idea of the kind of traffic to expect on the 5th. I read reports of people planning to set up fake balloons, and thought that it would be too much work.
When the registration was opened up a few days early, I panicked. I had not set up a webserver yet. I worked frantically to get an email server working, and that took me nearly 2 days to get it set up correctly with a catch-all address. At this point I was expecting to need to set up many accounts to give me the best chance to submit a correct balloon list if I were overwhelmed with bogus reports. I registered 100 email addresses at red40k.com. Later, a clarification of the submission details and the fact that registration never seemed to close meant that I was never in any danger of running out of chances to submit lists.
As the challenge day approached, I created my submission form and finalized the submission rules. I decided to make the first part as rigid and well defined as I could initially, and then as free form as possible for the second part. The first part consisted of select boxes that would only allow the location to be submitted in the correct format. The second part was open ended questions with text boxes that allowed submitter to fill in as much detail as they wanted. The second form could only be accessed using a code given after the submission of the first form.
As the day began, I was surprised to find the volume of submissions was quite low. I got 1 obviously false initial submission at approximately 5am before I woke up. The coordinates were not changed from their initial values, and only an email address was entered.
I received my first valid balloon report about a half hour after the competition started. the submitter filled out the form, and completed the second part of the form with plausible data. I called the submitter, and talked to him, verifying the report.
All the other submissions were also very easy to sort. If the second part of the form was filled out completely with plausible data that included a phone number to contact the submitter, then the balloon turned out to have been real. If, on the other hand, the data from the second part was short or incomplete, then the balloon turned out to be false.
A grand total of 9 form submissions were recorded.
1 was
obviously false.
2 were short and incomplete.
4 were correct
and complete
1 was an exact duplicate entry.
1 was an
independent duplicate of an already submitted balloon.
Balloons
correctly identified by submitters were #2,#3,#5,and #9.
The
biggest surprise for me was the low volume, and the high quality of
submissions I received throughout the day.