Again, I'm using standard player models for Pedro Florimon, which is all we can use for players who haven't shown any great strides improving their play on the baseball field whereas you are hoping that Florimon gets better without any real evidence to suggest that he will do so. Which analysis sounds more likely to be accurate?
1. Which model are you using? Can I see it. This isn't me debating this.
But please SHOW me the model, versus telling me what the model says. I was a research Scientist for 5 years in instrumental chemistry, I then taught AP Calculus (show me your proof!), and now I'm in analytics. I care very little for conclusions without statistical models to verify. It's no offense meant, but don't 'tell' me, 'show' me and be prepared for 30 "why" questions. And 'have you considered...' comments.
2. Also, what are you defining as 'success'... for Florimon, if you are at all defining success for him.
3. How accurate have these models you've been using correct? Do you go back to the Zips, CAIRO, Bill James, etc or whatever you're using to see if they're successful? To what degree? Maybe run a confidence interval on them for standard error? Some run higher on the projections, some run lower. Is your projection system (or the one you use), historically high or low? I know it's not exact - or you'd be making money left and right and wouldn't be here. Forget Nate Silver...but which is it? Have you taken account for this into the projections? I guess the one you may be referencing is an age one...is it changing at all year to year? Humans are maturing at different ages...body sizes are different, e.g. WW2 average American male at age 18 was 5'7'' 145 lbs. Now it's close to 5'10'' and 178 lbs.
It's fine to live and die on something, but say what it says, but don't argue for it as Bible truth without acknowledging all projection systems are flawed at best.
The projections say...." "
If you have a confidence of 95% (2 standard deviations) on that projection, well then it's something you got a 19/20 shot of happening. I'll take it, even with that 1 occurrence landing outside the expected value.
One guy was getting attacked by personal agendas in 2 different forums. And he wasn't wrong in either case. A lot of this is grey, not black and white.
Will you put a year's worth of beer consumption on your projection model for Florimon for next year? And I get any instance where he doesn't exactly meet that? what about 2 SD's within that number? Btw, I can drink a keg-ton of beer. I have more alcohol dehydrogenase in my system than mitochondria...and it's not even close.
I like your posts, you're [Brock] one of my favorite contributors. You seem to have a good sense of numbers and baseball.
But there are late bloomers. Michael Cuddyer was 25 before he really became a regular in MLB. He won a batting title at age 34, hitting .331.
Who could have predicted that? Not me. Not in a million years. Which projection system had Cuddyer hitting even .300 in his age 34 season?
He hit .311 on the road btw. In 2006, he hit .302 at home (the Dome) but otherwise, he'd never hit .300 at any home/away split. Pretty crazy. The stars aligned I guess?
Age projections are about 60% correct? That is, about 3 in 5 guys fall in line with those curves. Others are more linear or more exponential.
What do those age projections do to Stephen Drew? For his next 3 years?
What's the WAR/$ ratio of ROI for Drew vs. Florimon?
Edited by twinsfan34, 15 January 2014 - 12:33 PM.