I'm one of them. I asked this same question in my team a few weeks ago: I got very similar sampling. The team is composed of 7 hackers (native code, embedded speech-related software). In other words, I tend to believe the results we're getting so far.