Quote:
Originally Posted by cardsagain74
I am completely aware of the very elementary statistical principles that you just described. But none of that has anything to do with my point, which is sometimes coming to a biased or subjective conclusion from your data (and using just the "good contact pitcher" study as an example.
As a statistician, if McCracken didn't already have a tendency to lean toward certain preferred results, he never would've used those quotes by Maddux and Pedro to supposedly "prove" his point some (and actually would have likely mocked any mention of them doing so).
More often than not, I'm with you. Am the first guy to look for that progressive royal video poker machine that's gone to 100.75% in EV with perfect play and has a high enough hourly rate to be worth playing, etc etc. Anyone who doesn't trust the math in those completely quantifiable spots is ignoring undeniable reality.
But when it comes to spots where human elements are involved, my point here (and many others' point) is that there are too many intangibles that may or may not apply in some spots to come to such sound conclusions, even when some individual smaller pieces of that particular puzzle have been proven statistically.
It reminds me of how the EMH in the financial markets is seen by statisticians. When I actually read the "proof" that the market is supposedly 100% random, I was stunned by how elementary the research was. Especially given how some elite full-time traders have been highly successful over the course of many, many thousands of trades. Results that would be impossible by chance, no less.
You see someone like Bill Russell as highly overrated and extremely lucky. I see someone who likely didn't win 15 titles in the 16 years he was the core of his team from high school on by chance (and yes everyone, I know there were two years he didn't win a ring the pros...he got hurt early in their losing playoff series during one of those two seasons.) And see someone like Chris Archer as a guy who's just gotten extremely unlikely, rather than a pitcher who give up the big hit to decide a game much more often than the norm (over a period of many years and over 200 starts).
Sometimes there might be more to these things than just being that occasional outlier on the bell curve.
Dynamic events are just a totally different ballgame to datamine proof from than anything set in mathematical stone, imo. And I know you won't agree. Though I can't prove it  |
John,
Excellent post, great points, and I'm one of those people that keeps bringing up the variables, like you. Also like you, I understand the principles and such behind the math, and their use, especially in the example you gave with regard to the video poker machine. Great example.
What is funny is how statisticians seem to really despise the word "opinion" when used in reference to what it is they do. But you know right away in dealing with statistics, such as provided by baseball, that you're going to have recognized deviations and varying levels of confidence in the stats of a player as to what his "true" statistical measures (stats) should be. And since by a statistician's own understanding there is never going to be a 100% certainty as to a player's stats being their "true" stats then, the use of that player's stats for some comparison is never going to be 100% accurate, and therefore it is NOT an undeniable fact. And if a statistician uses something that is not an undeniable fact as even part of the basis for determing some answer, then that answer is their opinion, and can never be fact. They should look up the definition of the word "opinion". Putting it a simpler way, if someone says water is wet, that is a fact. But if a statistician says they are 95% confident water is wet, therefore I'm going to say it is wet, then that isn't a fact, that is their opinion. Simplistic, maybe, but I hope it drives my point home.
And using current stats to compare players all playing MLB at the same time is one thing, and obviously hard enough to do based on a lot of what has been discussed already. But to expand such comparisons to now include several players over a number of different years and dramatically different eras of the game, and to then think that statistics alone could ever provide a definitive answer to such comparisons, is sheer lunacy. And making and using a blanket statement that pitchers like Grove and Spahn played back in the day when players were weaker and nowhere near as good as today's ballplayers, and therefore Grove and Spahn are nowhere near as good as today's pitchers, as part of such a statistical analysis to compare pitchers, is not just wrong, it is downright insulting to Grove, Spahn, and everyone playing MLB back when those two were pitching. Just taking a pitcher today and throwing him back in Grove's day, completely ignores the context of the different times, different game, and on and on.
Often when people talk about financial matters from different periods or eras, they may at least try to reflect or account for those different eras by adjusting numbers for inflation so they can say something that cost $XXX in 1921, would cost or be worth $YYY in 2021, as "adjusted for inflation". It likely is not a perfect way to account for the different eras, but at least it lends for the information being more relevant and comparable to people today. What is needed to even begin to make blanket statements about how good (or bad) players from a certain era were, is some type of "adjusted for inflation" equation or formula to at least try to account for ALL the differences and variables faced by players from different times and eras. I doubt a formula/equation to perform such an "adjusted for inflation" calculation even exists. But I wouldn't be surprised if you asked 10 different statisticians to independently of each other work up such a formula/equation, assuming you could find 10 to even try, that they'd all (or at least a majority of them) come up with something at least slightly different. In other words, they'd each probably have their own, differing "opinion" as to what this "adjusted for inflation" calculation should.
And the very last line you typed highlights maybe the biggest problem of all. In the case of a subjective question like who's the greatesty lefty pitcher of all time, the statitisticians can't definitively prove they are right. But then unfortunately, we can't definitely prove them wrong either.