- Geocachers - This was our not so secret sauce for the competition. Our community of GPS enthusiasts already have the travel bug, so it made sense that it would help in a spotting competition
- Instant Cache Notifications - We sparingly used the Instant Cache Notification feature of the site to enlist geocachers on the ground. Our ability to inform geocachers of potential balloon sites helped us verify or disprove locations.
- Groundspeak Lackeys - I was impressed with the strong showing of Groundspeak Lackeys (employees) willing to come in on Saturday at such short notice to work on the challenge. We had over 10 balloonies at one point in the day.
- Social Engineering - At one point we started calling nearby buildings to verify some balloon sightings. This worked exceptionally well as long as you said the right things. We weren't the only ones doing this since some people we called said they were called by others about the challenge.
- Microsoft Excel - We were showing tweets on the big screen which turned out to be pretty useless. Drew switched to his computer and put up an Excel Spreadsheet listing all of the information we knew and what we were working on. It turned out to be invaluable.
- Fat Tire and Banana Muffins - Beer (Thanks Sara!), enjoyed responsibly, can make a Saturday challenge much more enjoyable. Doubly so with Banana muffins (thanks Sandy!)
What Didn't Work
- Last Minute Planning - We found out about the challenge late Tuesday and didn't decide to participate until Thursday. Since a large part of our strategy was to enlist the aid of geocachers, it didn't give us much time to rally the troops. Many other teams had a month (or more) to prepare.
- Altruism - In the end, MIT's profit sharing strategy likely made the difference between us winning and losing. We did, however, receive information from Craig due to the charity we selected, so it is questionable whether we can't win with a completely altruistic goal.
- Lack of Good Tools - I felt a bit like John Henry during the challenge. For the most part we were relying less on creating an ad-hoc network than using our community. MIT's steam powered hammer, machine learning, didn't beat us handily but it likely made the difference.