Posted by cbriese on 4/23/2012 8:46:00 AM (view original):Some folks apparently still think the pitcher with the most wins should get the Cy Young, and the hitter with the most RBIs should win the MVP. That appears to be the kind of logic used in determining your world rankings.
Look, we simply said that the analysis could be improved via a few changes:
Some folks apparently still think the pitcher with the most wins should get the Cy Young, and the hitter with the most RBIs should win the MVP. That appears to be the kind of logic used in determining your world rankings.
- Use standard deviation rather than gross averages
- Look at at three-year period rather than a singe season
- Incorporate runs rather than wins
Is this, by the way, part of the manifesto?
Not at all. I only went through one year initially for a couple of reasons. 1. If the results made no logical sense--or if I could make improvements to grab better data, I didnt want to have to go through 3 years worth of data twice. As I posted throughout the thread, I will incorporate multiple seasons.
I get your point on wins--and admittedly looking at 100 wins/100 loss teams was a bit simplistic. But initially to put these out it was easy to grab the data. I will be grabbing top 4 wins/loss teams. I think for the parity aspect run differential will be helpful--I expect it will just reinforce what I will get looking at the top and bottom w/l. Although I am sure there are a few worlds where that data will tell a different story.
Cbriese, as far as I recall, you mentioned nothing about Standard Deviations until this post. As I mentioned before, the way I set it up initially didnt lend itself to grabbing all team data and making calculations based on that data.
Im not opposed to it in theory--it may result in slightly different rankings and better score the worlds based on era and +/- where some monster teams are driving up the average, but I also think that by weighing the parity score more heavily achieves pretty close to the same result.
I get people defending their worlds and trying to find some criteria that they think will better judge worlds. But what I dont get is people like Robusk coming in and just being douchey (I actually do kinda understand it since its coming from Robusk). If I were a paid employee of WIS and this were my job--it would make sense. I just put this together over about a day because I had some time and thought it would be interesting.
And you know and I know if I went through the laborious process of calculating Standard deviations and Variance to the mean (or whatever the heck else he suggested) for 160+ worlds for 3 seasons each across multiple categories, Robusk would still find a reason to **** on the rankings.
Any ranking system will be inherently subjective..and people will disagree with the results--no amount of statistical complexity will change that.