I'm fairly certain they have the broad strokes of this right, at least to the degree allowed by available data. But on the finer points I'm not as convinced. D

I'm fairly certain they have the broad strokes of this right, at least to the degree allowed by available data. But on the finer points I'm not as convinced. Dating a sample to within a year or two based on minute differences in the script seems quite unreliable. While there are certainly trends that develop over time, individual authors are by no means tied to them. The trends followed by one author could be decades away from implementation by another, or could be ignored entirely. Even betting on internal consistency is risky, and vulnerable to influence from things like the author's mood, diet, sleep patterns, and the phase of the moon.