What did he remove? I am seriously asking.
Every game played by a team that plays its home games in a dome, even the ones not played in a dome, and yet included every game a team that doesn't play its home games in a dome's games in a dome.
Guy who wrote 'article' - Team X plays 8 games a year in a dome and that has to be the reason they didn't fumble a lot so I am removing them from my data.
Obvious questions - Yeah but what about the games they didn't play in a dome? Did they fumble more, less, the same when not playing in a dome?
Guy who wrote the 'article' - I'm not even going to look...
More obvious questions - What about when non dome teams play in domes? Do they fumble more, less, the same?
Guy who wrote the 'article' - Again, I'm not even going to look...
Even more obvious questions - Team Y plays in a dome and they did fumble a lot. What's up with that?
Guy who wrote the 'article' - Lalala I'm not listening, lalala...
Sure dome teams play 8 games a year in a dome but they also play 8 not in a dome, and yes I am aware a dome team could play a road game in a dome also but you get my point. If we are talking about 5 plus year stretches that is still a significant number of games not played in a dome. You can't throw those out just because you don't fit your narrative.
And my understanding is that while yes teams fumble less in controlled conditions it is not such a significant amount that it is even remotely close to warranting tossing the games out let alone all the games played by teams that only play half their games there. but even if you did want to throw out games in a dome you need to do that for everyone. The fact that a home team's game in a dome doesn't count but the visiting team in the same exact game does is simply absurd.
The data is just flawed. It wasn't I'm going to pull together the data and see what it says. It was I know what I want the data to say now how can I make it do that.