Wednesday, July 11, 2007

A pointless exercise to determine trivial facts that don't matter to anyone and is stupid anyway because it probably won't work and's for stupid jerks

Also known as baseball stat blogging!

So cnn has each team's total runs scored and total runs given up alongside their record, which means you can, if you can work excel, plot correllations. Or, if you're me, and lazy, you can just do the following half-assed exercise to determine, once and for never, whether pitching really is key to the pennant.

Okay, so I added the total runs scored, for or against each team, and also noted the total number of games over or below five hundred for each team. Then I gave each team a rank from 1 to 16 (I only did three divisions, ALC, NLC, and NLW because, again, I am lazy, and the exercise is pointless and stupid) based on how many games above five hundred they were, where 1 was the most above five hundred, and 16 was the most below. Then I assigned ranks from 1 to 16 based on total runs involved in the games, where 1 was the most runs involved in all games the team has played and 16 was the least runs involved in all games in which the team has played. Then I determined the absolute value of the difference between each team's ranks in these two statistics, i.e. if a team were ranked 1 in games above five hundred and 1 in total runs involved in their games (like the Tigers) the result would be an absolute value of zero (1-1=0). But if a team were first in games above five hundred but last in total runs involved in their games, the absolute value produced would 15 (1-16=-15, the absolute value of which is 15). To try to see whether teams that tended to play in high scoring games tended to win more or less than teams that tended to play in low scoring games, I then calculated an average absolute value and looked to see whether it was closer to the average absolute value that would occur if rank in games over five hundred were perfectly and positively correllated with rank in runs involved in the team's games or closer instead to the average absolute value that would occur if rank in games over five hundred were perfectly and negatively correlated with runs involved in the team's games. If teams that tended to play in high scoring games always finished better than teams that tended to play in low scoring games, you would get an average absolute value of zero (team ranked 1 in the standings would be ranked 1 in total runs scored by and against, producing an absolute value of zero, team ranked 2 in the standings would be ranked 2 in total runs scored by and against, etc...). On the other hand if every team that tended to play in low scoring games wound up better in the standings than every team that that tended to play in high scoring games, I think the average absolute value would 8.
It turns out the average absolute value was 4.875, which, if this exercise had any validity, and weren't fucked up six hundred ways to Sunday, would suggest a small and almost certainly statistically insignificant tendency for low scoring teams to do better in the standings. Note howver that the sample includes both AL and NL teams, including the Tigers and Cleveland, who have the best two records in the group and play in the highest scoring games of any teams in the group, which you would think would bias the stats in favor of hitting. The AL is just a better league this and, sigh, every year, so you would think that the interleague play would depress the NL teams rankings, thereby tending to show, perhaps speciously, that teams playing high scoring games do better. So the fact that the pitching still appeared to exert the stronger influence says something. I mean, says something?

Pitching it is then!

The careful observer will note that this data clearly shows that the Astros deserve the unrelenting loyalty of every human being that has ever or will ever exist since the dawn of time.

1 comment:

D. Chang said...

funny. i was moved.

are you at astrosdaily.com's forums? you should be. :)

(I'm scrybbler over there.)