My analysis of the Burnett acquisition: http://fantasybullpen.com/free-agency/burnett-signing-worth-it-for-yankees/

I really hope they get one more pitcher, Lowe or Sheets. Hughes could be great some day, but we need to make sure he’s ready. He’ll probably get his shot this year with Burnett and possibly Sheets in the rotation.

As far as comparisons, the advantage to calculating it about replacement level is we can see what the fair market value is in terms of salary. Without doing so, we get Lowe at +35 runs next year and Pettitte at +30, which is a difference of 5 runs, or 0.5 wins.

We cannot definitively say Lowe is a half-win better than Pettitte but based on their true talent levels, this is the case.

The reason Lowe might not get Burnett money is the fact that he is 37 years old. Burnett is younger. I think I’ll turn this into a full post and we can talk about it there.

Conceptually, I understand the points you are making and wouldn’t quibble too much with what you’ve said. I think my only beef here would be that if you are rounding numbers and estimating what Lowe’s FIP would be in a harder league, I guess I have a hard time expecting that the numbers include the type of precision where one can definitively say Lowe is a half a win better than Pettitte. As I say, I’m just a neophyte trying to learn. Maybe the projection numbers are reliable enough and precise enough that a projected half win difference is a sizeable difference. I realize that trying to find out what relative value a player has based on their peripherals is an instructive exercise and can have important implications. I can understand that you can confidently say someone is 1 or 2 wins better than someone else. But at some point I think you reach point where parsing things in to ever more minute differences in projected value with any certainty or precision becomes difficult. Maybe it’s a quarter of a win. Maybe it’s an eighth of a win. I don’t know. But to me, it seems difficult for me to say with any sort of p=95% that a half a win based on projections portends any real, meaningful decision making tool. I suppose I could be convinced, otherwise.

Is Lowe really only going to wind up at $14 million for 4 years? I’d be surprised if he wound up that low given the rumors of Burnett’s alleged 5 year, $85 million deal.

One other point I would make, too, is that your average/fair market value analysis makes sense, but only if you do so in a vaccum. Sure, if they were both on one year deals and Pettitte was making $16 million and Lowe were making $14 million, Petittie would be overpayed. But if Pettitte makes $16 million next year and Lowe makes $14 million a year for 4 years, I’m buying of value something for that extra $2 million next year – namely the payroll flexibility of not being tied up for the next 3 years. For other teams, maybe they need to maximize value per dollar, but if I am Brian Cashman and running the Yankees, I would be willing to overpay Pettitte by $2 million just so that I don’t have extra years of contracts choking me. Collecting unmovable contact after unmovable contract put the Yankees in position over the last few years where turning over the roster was a near impossibility. So if I were Cashman, I would be willing to pay an extra $2 million for a comparable pitcher like Pettitte if it meant that he could maintain a more flexible payroll/roster.

This, obviously, is tough argument to have because we’re missing 2009′s data. We don’t know if Lowe will tail off at all or if Hughes will be able to stay healthy. That said, I’m not ready to tie Hughes/Kennedy to a tree a shoot them – a lot of players struggle in their first time on the merry-go-round. Given that, as well as other intriguing arms in the Yankees’ farm system and other free agent pitchers that may come available (Lackey, Harden, etc) I don’t know if it is crazy to think that in years 2010 and 2011 the Yankees could find an arm to give them what Lowe could. So why tie up money in Lowe if you a rough comparable in 2009 and an expectation that in 2011 and 2012 they’ll have a chance to find another arm to give them what Lowe can give them or more?

When we’re talking about the runs above replacement for pitchers, we are comparing the pitcher in question to 150 IP at 5.50 FIP for a replacement starter (a replacement starter won’t log 206 innings), and the difference goes to a 4.50 FIP reliever.

So, Lowe is projected for a 3.67 Marcel FIP, but I adjusted that to 3.80 given that the AL is a superior offensive league. 3.80 FIP at 185 IP is worth 78 runs.

A replacement starter with a 5.50 FIP in 150 IP is 92 runs. The remaining 35 IP go to a 4.5 FIP reliever, for 18 runs. The difference here is 110 for replacement, 78 for Lowe, which is +32. Then we add on 3 runs because he’s pitching these innings, not the relievers, so he’s a +3.5 win player.

Pettitte, at 182 innings and a 4.00 FIP (round it up to make it easier) would be at 81 runs. For him the replacement would be 92 runs (150 IP at 5.5 FIP) and 16 runs from a 4.5 FIP reliever. That is 108 vs. 81, for a Pettitte advantage of +27 runs. Add on three runs and Pettitte is a +3.0 win guy.

Lowe is about a half-win better than Pettitte in this regard for 2009. Therefore, via fair market value, Lowe should make about $2.75 mil more than Pettitte. Andy wants at least a $16 mil deal. Lowe is going to get a 3-4 yr deal for something like $14 mil/yr. It makes sense to sign Lowe over Pettitte if Andy isn’t taking a pay cut.

If Andy accepts the $10 mil contract, then it is a good deal, because he is only a half win worse than Lowe, and has a FMV worth more than $10 mil.

Eric:
Thanks for the info about the various run environments; I’ve never seen that data.

I would say, though, that when noted “I know FIP is an isolated statistic, and that Lowe was about a half a run better, but over the course of 200 innings, what is that, 10 runs difference?” I was citing both Pettitte and Lowe’s numbers for 2008, trying to compare similar data sets because I imagine that someone would come along and cite Lowe’s FIP and Pettitte’s FIP.

But as I mention to Samg above, I think Lowe is coming off a career year, and I am not entirely sure that Lowe will outperform Pettitte. If you take a look at their 2009 Bill James and Marcel Projections on their fangraphs pages and calculate runs:

James projects:
Lowe: 206 Innnings, 3.80 FIP = 87 runs
Pettitte: 192 innnings, 3.66 FIP = 78 runs + (14 relief innings @4.5 FIP) = 85 runs, which is fairly comparable if ever so slightly in favor or Pettitte

Marcel projects:
Lowe: 185 innings, 3.67 FIP = 75 runs
Pettitte: 182 innings, 3.98 FIP = 80 runs + (3 relief innings @ 4.5 FIP) = 81.5 runs, which means a 6 and a half run difference in favor of Lowe.

(I am something of neophyte, so my math here may be completely off; any corrections in my calculations or things I am ignoring would be extremely helpful). So I guess my point here is that I think it is at least up in the air about who might be a better bet to produce next year. And, to me, if it is up in the air, why not go with the guy willing to sign a shorter term deal in order to preserve payroll flexibility.

For what it’s worth, I haven’t seen a whole lot suggesting that Pettitte is looking for a raise over last year’s $16 million, just that he doesn’t want to take a pay cut, nor have I seen anything suggesting that Lowe will sign a 3 year deal. I thought he was set on 4?

Here is the reasoning behind why 10 runs = 1 win, from TangoTiger himself:

10 runs is a win: why?

The basic idea is that if you look at all teams in baseball history that have scored 1 more run than they allowed, per game (+/- 0.1, to increase the sample size), you will find that they have a .600 win%. And similarly if they allowed one more run than they score, they will have a .400 win%. That means each additional run leads to 0.100 additional wins, above the .500 mark. And 1 divided by .1 is 10.

The lower the run environment, the more impact each run has. And the higher the run environment, the less impact each run has. So, the 10/1 ratio is not fixed, but dependent on the run environment.

Here is a chart that shows the various win%, for various run environments, at various run differentials:
http://www.tangotiger.net/winactuals.html