So Bill James isn’t good enough for you? But seriously, why does your projection vary so much. Is it having more importance on the last season, or does it not account for aging as aggressively.

One thing that needs to be considered is how aging should effect projections on a position bases. I wonder if these projections do that. Would it even make a difference?

Another thing could be to have another list of projections by fantasy sites and such. ESPN/Yahoo/CBS/RotoAuthority/etc and other mainstream sites all have projections. I bet these would be totally different and be way more optimistic and not as accurate.

So Bill James isn�t good enough for you?

The James projections are not available in one handly downloadable format without paying for them.

But seriously, why does your projection vary so much. Is it having more importance on the last season, or does it not account for aging as aggressively.

The weights are similar to the other systems.  My guess is I am not regressing towards the same mean and probably a little conservative on aging.

One thing that needs to be considered is how aging should effect projections on a position bases.

I know Pecota and ZiPS use position and similar players in their aging/projecting.  Marcel does not, not sure about Chone, and I know Cairo doesn’t yet, although I will consider it for 2009.

Another thing could be to have another list of projections by fantasy sites and such.

I don’t see the sense in paying for this:

I bet these would be totally different and be way more optimistic and not as accurate.

One thing I’ve always wondered about Posada was what ever happened to that issue with a tendon (or something?) in his knee. In ‘06, he had a special pad on his uniform pants for it, and the word was that he was going to get it take care of in the offseason.  Then he comes back for an 07 and is miserable behind the plate but has a great offensive season (although his super high BABIP explains his high BA.)

SG: Just wanted to check in and say that I’ve been enjoying the various and sundry offseason features. I haven’t been commenting much, but I have been reading and appreciating. Honest.

Looking forward to the end of March…..

Have there been any reports about changes in Posada’s swing or approach last year that might help explain his great season?  That might speak to the likelihood of a change in his baseline wrt the standard age curve.  (Is there any evidence for such effects more generally?  There’s an article in the NYT about Delgado that’s relevant.)

SG, very good work, as usual. Just wondering if CAIRO might have a bit of Yankee bias in the projections? It’s troubling that you’re out of line with the other systems on Jorge - or maybe Jorge is an exception or maybe you just have different aging factors. Either way, something to think about.

Posada himself claimed that he was swinging more level last year and that it helped him hit more line drives.

Here’s what Fan Graphs says:

<u>Career</u>
LD%:20.3%
GB%:42.5%
FB%:37.2%

<u>2007</u>
LD%:22.1%
GB%:40.4%
FB%:37.5%

Posada hit 91 line drives last year, compared to 75 in 2006.  He’d hit 69 in 2005, 62 in 2004, 80 in 2003, and 84 in 2002.  I’m being lazy and not adjusting for balls in play right now but a weighted average would have predicted around 72 last year, so hitting 91 would seem to be statistically significant assuming a Std Dev of around 8 in either direction.

Still, a line drive percentage of .221 should support an expected BABIP of around .341.  Posada’s was .389.  Career is .323.  So he may have made a tangible change that helped, but he was likely a bit fortunate as well.

Just wondering if CAIRO might have a bit of Yankee bias in the projections?

Posada’s the big outlier, I don’t think I have an inherent Yankee bias, although I do think I am less harsh on aging than the other systems which could lead to what appears to be a Yankee bias.  Bill James projects .281/.386/.469.  I’m probably about 10 pts high on the avg/obp/slg for Posada.

Do you think that Posada will feel the effects of age. In general age regression should be done in terms of a model based on probability and such. Looking at history players tend to do worse as age becomes a factor. But when theres a player that doesn’t fit the general average, should age be considered. Arod is at the point where age regresses his pefermence on a projection level. But should it? Usually players at his caliber tend to age better than the average player. Maybe take age out of the equation on players like him, or factor it in more for weaker players. This is one area where Prospectus does a great job in projections.

SG,

Do you recall where Posada’s comment about a more level swing appeared?  I have the opposite memory: that of Po saying he’d made no changes at all, and attributing the bounce in his numbers to luck.  I ran a google search and found an article quoting him to that effect.  I certainly didn’t *see* any change…

It would be great if even some of the improvement was potentially sustainable!

Do you think that Posada will feel the effects of age. In general age regression should be done in terms of a model based on probability and such. Looking at history players tend to do worse as age becomes a factor. But when theres a player that doesn’t fit the general average, should age be considered.

I think we should assume everyone will feel the effects of age.  For a a 36 year old catcher we should expect it more.

In general, under 27 and we should expect players to get better, after 27 they should get worse.  Of course, there will always be outliers but it’s tough to find those in the data.  That’s where we’d expect a system like PECOTA or ZiPS which uses similar players as part of their projections to do better than the others.

Arod is at the point where age regresses his pefermence on a projection level. But should it?

Sure, it should.  But since he’s at a such a high-level, even the expected decline will have him as a very good player.  And how he does in 2008 will adjust his projections going forward accordingly.  2007 bumped up our expectations for 2008-2010 for example.

Maybe take age out of the equation on players like him, or factor it in more for weaker players.

Any age factor is based on adjusting what a player has already done, so the better players are not penalized as much when we add the aging component to their projection.  The most important part of any projection should always be what a player has actually done, with the most recent performance weighted the most heavily.

Do you recall where Posada’s comment about a more level swing appeared?

I’m not sure Rob, I thought he mentioned it an interview last year but I can’t find any corroborating evidence.

Part of the problem is that predictors depend in part on comparable players, and there are only two players in the history of baseball—Elston Howard and Carlton Fisk—who are remotely comparable…....

I don’t understand why you bumped the $4.4mm figure up to $4.5mm. The Yankees are in the sweet spot where just one extra win can mean the difference between missing the playoffs and making the playoffs, which obviously has immense revenue implications. The Yanks also have a stake in the YES network and a larger revenue base than average (by a lot). Using the $4.5mm figure implies that a marginal win is worth 2% more to the Yankees than to an average team - that is clearly absurd. Someone like Vince Gennaro would know more about this than I do, but I would *guess* a marginal win would have to be worth about 50% more to the Yankees, or $6.6mm per win.

What complicates the matter even further is the luxury tax. Marginal spending is more costly for the Yankees than it is for other teams, so this might offset the additional revenue returns from marginal wins. It’s something that is worth looking into.

And I was just joking about the Yankee bias.

Check out MGL’s post at The Book: http://www.insidethebook.com/ee/index.php/site/comments/sabermetric_moves_of_the_2008_pre_season/

To sum it up, the $4.4mm number is not how much a marginal win is worth, but the average price of a marginal win on the free agent market.

To add to that, the Yankees are definitely willing to pay more for a marginal win than an average team, for a couple reasons. First, the revenue returns on a marginal win are higher than average for the Yankees. Second, the Yankees stated goal is winning, not making a profit (this comes from George). It seems like Hal is trying to reverse this, but who knows if he will succeed.

The luxury tax still complicates the analysis. I think leaving the number at $4.4mm makes more sense, when trying to compare the Yanks’ return on investments relative to the market rate.

I don’t understand why you bumped the $4.4mm figure up to $4.5mm.

I pulled it out of my ass honestly.

Someone like Vince Gennaro would know more about this than I do, but I would *guess* a marginal win would have to be worth about 50% more to the Yankees, or $6.6mm per win.

According to Gennaro in this article about the cost of getting Johan Santana:

If we consider the star-studded, A-Rod-anchored Yankees to be a 93-win team before adding Santana, then we can say that his presence projects the club to 98 wins. Based on an analysis of recent history in the highly competitive American League East, those five wins would raise the Yankees’ chance of reaching the postseason from an already strong 62 percent to a comfortable 91 percent. The expected dollar value of this increase would be modest over the next two years while fans flock to see old Yankee Stadium for one last time in 2008 and sample the new stadium in 2009, regardless of the team’s won-loss record. But by 2010, when the new stadium is passé, I estimate those five wins will be worth a cool $24 million per year in incremental revenue, driven by how Yankee fans historically respond to a playoff-bound ballclub, along with the impact of the new stadium’s higher ticket pricing and additional luxury suites.

So that’s $4.8 million per marginal win.

Luxury tax and revenue sharing do complicate the analysis, as well as revenue from YES.  I guess it’s reasonable to just hold the Yanks to the same standard as the average team for these purposes, so I’ll keep it at $4.4 million for the rest of the players.

The $.1mm is pretty much negligible, I wasn’t trying to be critical. I just think it’s a very interesting question and is something definitely worth looking into. Gennaro’s work is an excellent start but is far from definitive.