It is easy for someone looking at these projections to just look at the bottom line and not look beyond that. But it is clear from everything I read from you, Smith and Silver about these projections that what is most important is the deeper information. Silver, for instance, says that every projection has to bee seen as a range of possibilities, as opposed to a specific projection of a .335 OBP or 28 HR’s.

Thanks again for the info and your insight. Do you have a schedule in mind for the roll out of the team ZiPS, and do you have any idea when the Royals projections will be out?

If a outfield of Teahen – Dejesus – Guillen (more O, less D)

will be better then one of Dejesus – Crisp – Guillen (less O, more D).

I’ve read many of these statistic geeks’ (meant in a complimentary way) debates and I gather that these projections use sort of a 3 year weighted average to “project” next year, or something like that. So, trend should matter within that context. Although, I see what you are saying if the recent year is weighted more than the earlier years.

But, if that were true then Teahen’s constantly declining production still should play into that and show worse projection.

who just happens to be playing in the outfield. If anything, he was moved to COF because he’s an inferior athlete than Gordon.

projections and advanced statistical models are really cool and often paint a clearer picture of players actual abilities.

but in this case, i really really really have to go with my (gut) feeling (eghhh). totally subjective, gut feeling tells me that gordon hits 270/375/460.

One thing, for example, Alex Gordon had 66 BB’s in 493 AB’s. Why would he have 7 less in 27 more AB’s? Is that figure based on “similar” players career paths?

writing a long email to someone else… we need to get rid of OPS and stuff… Gordon is already in the “not the problem” category, even if he isn’t A-Rod, David Wright, or Longoria. He was above-average overall this year, and might even have been around 3 WAR, if you add BP’s baserunning stat (one of the best things over there other than PECOTA and EqA) to Fangraph’s… Hell, I had him at 2.84 without baserunning.

A 3 WAR player won’t carry a team (especially with a “supporting cast” of Guillens, Olivos, Teahens, and Jacobses), but they can be a great value anyway.

But, as you note, the problems are with lefties. In 2007, he couldn’t make contact but actually had decent ISO (.217/.266/.420). In 2008, he improved the contact but lost the power (.234/.312/..317).

Here’s some bullshit I stirred up, but it might help us feel better (other than looking at his great second-half in a small sample size): I think it’s fair to say that his 2008 vs. RHP was legit: .273/.370/.491. Now, against lefties, take his BA and OBP from 2008 (not great shakes at .234/.312) but add the ISO from 2007 for a .437 SLG, for a.234/.312/.437 line. No great shakes, but a definite improvement. Combine those line with his RHP numbers, and you get something like .260/.350/.473 line.

 Now, an .823 OPS (whatever that means) may not impress people too much, but, although conversion to a real runs-proportional stat like wOBA isn’t possible without all the components, but in 2008’s run-environment, that would be worth as much as 8 runs above average more over this year’s underrated performance by Gordon.

That said, as I acknowledged, that’s a silly way to do things, but does make one wonder about what sort of studies have been done on the development of platoon splits among left-handed hitters, given that the character of those splits are different than those of RH hitters.

They are very valuable, very meaningful and have a record of being pretty darn accurate.

Valuable? To an extent. Meaningful? To an extent. Accurate? Not accurate enough.

Player X is most likely to perform somewhere in the range between his 25th and 75th percentile projection.

The problem is that position players tend to be far more consistent than pitchers in their production. One can give ‘general terms’ on how a player will perform and that prediction will be deemed “accurate”, even though it’s a large enough spectrum that the accuracy – or precision in their ‘projection’, isn’t that impressive.

This is meaningful information, not just some random number crunching.

I’m not sure if you’re responding to me or responding to other people who have made similar arguments in my post. If you’re responding to me with the bolded text, you need to read my comment again.

Does that mean he can’t breakout or collapse in 2009? Certainly not. But it gives us, arguably, the most likely trajectory of his short-term future.

It’s likely based on too limited information. This ‘definitive accuracy’ that sites like B.P. post is much hoopla and promotion, in my opinion. It’s a meaningful effort, and a necessary one, which is why I buy their book every year, but it’s hardly definitive, as they will have you believe. I feel like I could predict much of this information, based on: MiLB info., comps., health history, etc. etc. and fall into an only slightly inferior accuracy than projection systems. By the way, check out some of the projections for the 2008 Royals. Many of them were dead. freaking. wrong. Gload at .300/.350/.450? That’s a laugher in and of itself.

Just ran the projection and ZiPS has Butler at 283/346/433, almost identical to Marcel. The projection is similar to Gordon’s, but with a higher variance – ZiPS sees Butler as more likely than either Gordon or Teahen to be an elite hitter in 2009 but also more likely than either to be a terrible hitter.

yeah, I like that one better

I’m sure Aviles’ defense is just as sustainable as his offense, so I look forward to the next few years of Jeter-with-power offense, and Adam Everett-in-his-prime defense (bUZR had him at +29/150 in 2008). Lock him up now, Dayton! Right after you get done with JoGui’s extension!

Dan, thanks for coming here, even if the news wasn’t quite as good as I’d hoped. I can’t remember if I asked this earlier, but how much (if anything) does ZiPS do with platoon splits? I ask because Gordon’s are quite pronounced. He improved markedly against RHP from 2007 to 2008, but while his contact and walk rates improved against LHP, his ISO got cut in half. Have you simply found that stuff like this is insignificant? Just curious.

P.S. — I know we can figure it out ourselves, but any chance you’ll starting accopanying your projections with wOBA (along with league woBA and linear weights above average)? Just a request.

Thanks for kicking ass with a publicly available projection system.

is exactly the same as this past seasons, assuming that the “standard” wOBA weights somehow all equal out to the custom ones from 2008. Which they won’t. His offense isn’t what we want or expected, at least not yet, but the defense really hurt him last year. Not to mention the injury. Maybe the two are connected?

And this doesn’t consider his baserunning. BP had him at about +3.4 runs. Of course, the Amazing Avilanche was at about +4.5 (!). Holy shit. that puts him over 5 WAR in 109 or whatever for sure…

Unfortunately, Mark “Baserunning 4 Life” Teahen is at -1.8 (did anything go right for him in 2008?), even worse that JoGui’s -1.7. Teahen and DDJ were both around 3 in 2007.

Too bad they don’t go back before that — I wonder why? Surely all this can be reconstructed with retrosheet…