From the commencement of the 51st class of neophytes, given by most holy Dean Rebecca in the 60th once-autumn since the establishment of Kepler's microbe colonies—

What does it tell us if we have one failure in a group of three? Is that statistically significant?

Says Aria: Of course it is statistically significant—a 33.3% failure rate is far too large to ignore. And says Dafydd: Of course it is by no way statistically significant—there is a clear majority even at this small scale, and to ignore that is folly.

Both are wrong, of course.

The true answer is that we have no way to tell. At three trials, one failure could be either. We cannot begin to see true statistical significance until the dataset has expanded, until we know if that one failure is truly 33.3 percent, or if it is only one among a hundred, and we found that single lapse early.

And even then, I would caution you. Do not write off that single failure. You know, of course, of penicillin

One culture contaminated with mold. One. A single outlier, out of so many that we do not bother to count them. And from that, the greatest miracle of medicine until the Ares-1 team returned from Mars.

There is our paradox. One single data point is never enough. That contaminated culture was repeated, again and again, until certainty was achieved. Three data points is never enough. Even if all are in agreement, three cannot encompass the full breadth of experience. And yet, one data point, among a thousand others, can still be enough to say to us: There is something here.

And so, I say to you all: There is something here! Listen for the call, and face it boldly when you hear it, for you will know the truth of that which you study. Someday, perhaps, that call will be from our beloved Anomaly itself, and to that call, I urge you: Look, and listen, and answer.