The idea that we can use distributions to describe political systems really stuck with me. Sure, it’s not the only one way to do so, but it seems to offer a particularly interesting and unbiased approach to measuring the survival of political systems. Unfortunately, when I discussed this fact with an example of the Roman Empire, I got it wrong. Not the general conclusion. Just the detail. The way I did it was wrong… which is embarrassing. But there you go, now there’s going to be a record of how I did it wrong and how you do it right, which was kind of what I wanted to do with this blog in the first place.
The rest of this post is organised under the following headings:
- How did I figure I was wrong?
- I counted duration in approximate (integer) years not in days
- I made a mistake calculating the Kolmogorov-Smirnov Test
- Getting accurate data
- My Final Say – Data, Procedure and Results
If you are just interested on how it gets done, scroll through to the end, where I’ve actually posted the excel file with the correct solution to the problem. I hope this helps and somewhat redeems my mistake.
Continue reading Correcting the Probability of Roman Emperors – Data, Empirical CDF, Exponentional Functions, Kolmogorov-Smirnov Test, Confidence Levels and Critical Values