Geoff Baker covers the Mariners for The Seattle Times. He provides daily coverage of the team throughout spring training, and during the season.
February 11, 2008 3:40 PM
Posted by Geoff Baker
Just to be clear, I am very in-tune with what Graham from our comments thread was trying to say in his Lookout Landing post I refered to earlier today. The point is, all that anyone trying to predict the M's fortunes this season through numbers has to work off of is probability. You can take "X'' as the general public consensus of what makes a good team and work off the probability of that. Or, you can add to that with some sabermetric numbers and try to make it more educated than just a random "X'' guess. The latter would seem to be the way to go. Fair enough? Great. I understand it. I assume the vast majority of you do too. We don't need an "Intro to Stats 101" class to explain the whole point behind the coin-flip analogy either.
If you flip a coin with a 50/50 chance of getting "heads" each time, but you wind up with "heads" nine times in 10, it doesn't make you smarter or "right" for defying the odds of probability. Here's my point, once again, for those not yet clear on the concept. When you try to push this probability thing to a baseball context, the point gets blurred. How do we define probability? Whose probability numbers do we believe? Which numbers, when added to the 'X'' quotient, are making it a better guess?
If we're examining Bill Bavasi and his logic at building a team in regards to the Pythagorean record the M's had last year, which probabilities do we look at?
Do we argue that the M's had only a 78-win Pythagorean record last year, ergo they can't possibily contend this season even with Erik Bedard?
Or, do we look at numbers showing that teams outperforming their Pythag one year have a better chance at improving the following season than teams that underperformed their Pythag. Since the M's fall into the former category, it reasons that they stand a chance at improving upon their 88-win season to...what exactly? 89 or 90 wins? 94 wins? In my book, anything close to that makes them a contender. Especially with the addition of a potential No. 1 starter in Bedard.
Two theories, two very different win totals. So, which probabilities are the "right" ones for Bavasi to play in regards to the Pythagorean expectation?
Not so cut-and-dried, is it? That's kind of how we end up with simulations like this one, performed by the U.S.S. Mariner using ZiPs projections, which have the M's at roughly 77 wins, while others like this one done on Detect-O-Vision, using some of Ron Sandler's projection system, see Seattle as an 88-to-98-win team. That's quite a win total spread by both sides. Who's right? Both use some impressive numbers to make their points. But who to believe? Which of the two is trying to defy the probability odds?
So, how do we judge Bavasi to be right or wrong in playing the odds, if the numerical projections can be that far off? If Bavasi wins a division title this year, can any of the anti-trade advocates truly say he was "lucky' the way any folks coming up "heads" nine times in 10 while flipping a coin would be? Or, if Bavasi finishes third in the AL West, will those who advocated against trading for Bedard be the "lucky" ones -- assuming the trade was the "right" move to make as per Detect-O-Vision?
I'm waiting to hear why those advocating against the M's making their Bedard move right now are so steadfastly certain that their numbers are the best place for Bavasi to start his building plan. I've already told you why I believe the team made the right move. It's got nothing to do with my media access. I'm just not buying the counterargument. And if my point of view leads to a winning, playoff club, can that really be dismissed as "lucky" the way coming up heads often in a coin flip would be?
Sorry, but that sounds awful smug. Way too certain for folks dealing with so much uncertainty.