The blog Bayes Ball cites my post among several theories attempting to explain the recent trend of perfect games. He then conducts a careful analysis using the Poisson distribution, which is perfect for studying events that are extremely rare (in this case, perfect games) but have many, many chances of occurring (there are thousands of regular season games each year). His analysis shows that a season with two perfect games is well within reason.
After reading the post, I had a few protests. For instance, if Jim Joyce gets the call right, we'd have three perfect games this year, and the Poisson analysis goes out the window.
Even if we let Joyce's call stand, I speculated that a season with three games in which the pitcher gets the first 26 batters out would still be exceedingly rare by historical standards, and we'd have to start questioning the assumption that the dominance of pitching relative to hitting hasn't changed much in the past few decades.
In my previous post, I argued that the string of perfect games in fact did demonstrate a change in favor of pitching. I even suggested that baseball should change its rules (like adjusting the height of the mound or the dimensions of the strike zone) to restore the historical rates of success and failure between batters and pitchers.
But after a while, I was forced to wonder: What if I had spotted a trend where there wasn't one? What if I had advocated that we fix a problem that doesn't exist?
There is a connection to macroeconomic policy. To keep the economy on track between the undesirable extremes of too much unemployment or too much inflation, the federal government often modifies monetary policy (such as changes to the interest rates) or fiscal policy (such as changes to taxation and government spending).
The chief advocate of such tools was John Maynard Keynes, who is regarded as a hero among most economists but whose theories are generally dismissed among the George Mason economics faculty and students.
One has to wonder how often we are tempted to implement macroeconomic policy changes in response to economic news that might just be the result of statistical fluctuation. Perhaps action is occasionally warranted to rebalance both baseball and the economy, but we should think twice before moving in the fences and allowing each batter five strikes, or introducing stimulus programs and holding down interest rates.
Guns, Suicides and Natural Experiments
1 hour ago