Reflections on four months of blogging

I started this blog four months ago because I wanted to contribute to the public debate over the unfolding drama in Ukraine. I had given a number lectures and interviews on the crisis, and had written two opinion pieces, but events were unfolding very fast and I wanted a way to contribute quickly and frequently, so I decided to try my hand at a blog.

I have very much enjoyed the experience, despite the depressing subject matter. It has forced me to follow the events as carefully as I can, albeit from a distance, and it has led me to explore certain topics, such as the Transnistria situation, that I had not written about in the past.

My goal has been to weigh the evidence and try to predict what is going to happen to the best of my ability. Initially I thought I should put my money where my mouth was and spell out what I thought the odds were of certain outcomes (e.g., “There is a 35 percent chance the Russians will invade in the next two months”). After trying this for a while, however, I decided the practice suggested a precision that I did not intend (and sounded odd), so I changed to nominal scaling (e.g., likely, unlikely, highly unlikely, etc.).

That said, as someone who spent far too much time in college playing backgammon for more money than was wise, I think social scientists should be encouraged to predict/forecast – it is humbling to be regularly reminded that you can be wrong. Making predictions frequently, and getting some of them wrong, is also a good way to improve forecasting skills. One has to be careful, however, to distinguish between bad forecasting and being wrong (a distinction that escapes many people). If I say the odds of rolling double sixes is 1/36, and that it is therefore very unlikely that you will roll double sixes on your next play, but you then promptly roll them, I would be foolish to conclude that my forecasting had been poor. (It would be nice if people remembered this distinction when a very improbable event happens, and intelligence agencies are immediately blamed for an “intelligence failure” – improbable stuff happens all the time, as anyone in the prediction business understands.)

At any rate, reflecting on my posts to date, there are at least two that I wish I had written differently. The first is my post of June 20 (“What’s behind Russia’s ‘humanitarian intervention’ in eastern Ukraine”). The language in the post itself was appropriately hedged (“If a Russian ‘humanitarian intervention’ is indeed under way…”), but the title of the post was not, which was a mistake. I decided not to change it after the fact, though, because I thought it would look like I was trying to cover up an error – and it is true that I thought, incorrectly, that the evidence suggested that an overt Russian intervention, and not just a surge in men and materials crossing the border, had begun.

My second regret is a careless insertion of the word “more” in my July 28 post (“How will Russia react to Ukraine’s battlefield successes?”). The original sentence, with the offending word highlighted, was: “And we may well witness air-to-air combat between Ukrainian and Russian fighter jets, more Russian SAM strikes against Ukrainian aviation from Russian territory, Ukrainian SAM missile strikes against Russian aviation coming across the border, and artillery and rocket exchanges across the border.” The “more” was inappropriate because, as someone pointed out to me, there was no evidence that SAMs had been fired from Russian territory at Ukrainian aviation. It was also unnecessary, since I was referring to what might happen in the future. Because the word did not reflect what I actually thought at the time, I removed it once the error was pointed out. Testimony, I suppose, to the disadvantage of writing quickly without vetting by peers and a professional editor.