Johanna suggested that we go around to some of the summer camps that are here, because the kids are likely to be interested in new technology. Tom has a contact in a math and science program, who set us up with 6 high school boys who were our best testers yet. They actually followed our instructions, listened to what we said, had a lot of good questions, and were happy to take the time. We then started putting all of the test results into a spreadsheet for analysis, which is where we ran into problems. Stephen gave Christen some logs from old tests he did, but the format was different. It didn't indicate when a test timed out or failed, and there were random numbers of tests. For example, the program might ask the user to hold up 5 fingers 7 times, but never ask her to hold up 3 fingers. This made the analysis more difficult. Then, we noticed that the new logs from the tests we performed also had quite a few discrepancies. There were supposed to be 2 requests for each number of fingers or number on the keyboard, but there weren't. The averages were figured wrong, so that if only 1 request for 3 fingers was made or if 1 of the requests timed out, the average would be reported as half of the actual value. So we had to do a lot of work to put all of the data into a spreadsheet. We managed to find the average and standard deviation for each user for each type of request (type 3 on the keyboard, hold up 4 fingers, etc.), for the Finger Counter and the keyboard overall, and then found the breakdown of both of those for all of the test subjects. We also helped Diane with some validation of her Bat Tracker.