Originally Posted by kaldaniels
The raw data ladies and gentlemen. Any dispute or claim to the contrary will include some sort of adjustment or qualifer. They were more common.
Again you are using a chart that is not adjusted for park and league factors. It is a basic statistical fact that there will be greater variation when numbers are higher. Since home runs were higher in the 90s due to non-steroid factors (strength training, small parks, juiced balls, dense bats, diluted pitching, small strike zone etc.) you would expect higher statistical variation due to basic probability. Once you adjust for the normal expected statistical variation it turns out that power spikes were actually LESS common than in prior eras.
Here is the correct chart:
As an example of the added variation you would expect when numbers are larger, one would expect to see more games decided by one run when the league is in a low scoring environment than in an era of higher scoring. In a year where the average team scores 3 runs per game you would expect to see more close games than you would when the average team scores 5 runs per game. That is not caused by a mysterious "spike" in lopsided games but rather just the expected normal statistical variation predicted by the mathematics of probability.