Lately I've been thinking about Doug Melvin's middle-man, crowd-pleasing position: that the Brewers will be willing to sacrifice some offense for some pitching.

I've wrapped my mind into a pretzel over how to determine the value of a run prevented against a run scored (in 2009 NL, the league teams allowed 9 more runs than they scored, which at a basic level means that a run prevented was very slightly more valuable than a run scored (to the tune of 1.01 RA: 1.00 RS).

On the surface, in a consistently depressed offensive environment, it seems like an intuitive way to approach the offseason; anyhow, the Brewers had one of the best offenses in the NL, scoring 67 more runs than the average offense (that's nearly 10% of the runs scored by an average offense), so it looks like the Brewers have a margin to dig into (if they give up, say, even 5% of their 2009 RS, that means around 746 RS for 2010, which could very well be an average to above average offense once again, if recent runs environment trends remain the same in the NL.

However, my thinking on this issue shifts once I think of two things:

(1) The Brewers' available resources

(2) The available free agent pitchers

(3) The likely talent required to obtain proven, average (and above) pitching via trade

(4) The level at which the Brewers' pitching staff needs to improve in order to become an average staff

Here's the deal; the average 2009 NL team allowed 727 runs, and the Brewers' 818 runs allowed requires the team to prevent 81 runs somewhere. The Brewers' bullpen isn't really a place to look for that task, given that the pen allowed 250 runs despite pitching over 540 IP (the average 2009 NL bullpen pitched 501.78 IP and allowed 242.6 runs, compared to 544 IP/250 RA for the Brewers' 2009 bullpen -- a rather exceptional performance when you consider the added requirement of 40 extra innings for the Crew's relievers).

So, the burden really falls with the rotation, which allowed approximately 74 more runs than the average NL rotation, which pitched 942.84 IP over 162 G, allowing 484 R. The Brewers starters' 891 IP and 568 R performance leaves a whole lot to be desired.

Which leads me to ask, how do you make up for that type of performance?

In order to figure this out, I needed to create two potential models for building an average rotation. The first model is simple: how does a true 1-2-3-4-5 rotation perform? The second is a bit more convoluted: how does a rotation of five #3 starters perform? Both models are unrealistic to some extent, because (a) no rotation actually goes 1-2-3-4-5 according to runs performance, and (b) no rotation actually goes 3-3-3-3-3 according to runs performance (and to show you the difference between rotation spots, the Brewers went 1-4-4-4-5 according to runs performance, which should demonstrate (a) the difficulty of using a #1 starter to offset multiple low rotation performances, and (b) the difference between #3 and #4 starters.)

Of course, the other issue is that relatively few rotations even reach 5 regular spots; there really wasn't even a such thing as a true #5 starter in the NL. The two true #5 candidates according to my survey has Manny Parra and Todd Wellemeyer sharing the dubious distinction of being the only two #5 starters according to runs performance (which is really, really awful), and Jeff Karstens lumped in as a swingman (only due to % of GS and IP, really; he was actually about as valuable as a mean #4).

That's my logic, to outline the problem. 

1-2-3-4-5 MODEL: 824.56 IP, 421.57 R

3-3-3-3-3 MODEL: 827.95 IP, 429.61 R

A couple of things should really jump out:

(a) A rotation of full-time, average #3 starters is almost exactly comparable to a 1-2-3-4-5 rotation. As a theoretical point, this should be powerful. I know actual available resources, scarcity of talent, difficulty of developing pitchers, etc., makes it extremely unlikely for a team to acquire five #3 starters, but it should re-open the debate about Doug Melvin's 2008-2009 offseason plan to acquire Braden Looper (a top #3 in 2008 NL) to work with Dave Bush (the top #3 in 2008 NL) and Manny Parra (another #3 in 2008 NL). With the gambles of Yovani Gallardo and the decreasing production of Jeff Suppan fully weighed, that rotation basically looked like ?-3-3-3-4, which isn't all that far off from a 3-3-3-3-3 model if all the starters remain healthy, make all their starts, and don't regress.

And there's your issue. After all, this is only a model, but really, think it through, thoroughly, and allow it to frame your logic of pitching acquisitions for a week. Then revisit it.

(b) Neither a 1-2-3-4-5 or 3-3-3-3-3 rotation completes the league average IP for NL 2009 starters, or fills 162 games. I will repeat it again for effect: Neither a 1-2-3-4-5 or 3-3-3-3-3 rotation completes the league average IP for NL 2009 starters, or fills 162 games.

I am fond of saying that a lot of baseball is just killing time, or biding time, and this seems to reflect that point: a team should expect to use 6 starters in this day and age, and probably have a swingman or two ready on top of that.

Either way, in a 1-2-3-4-5 model, or a 3-3-3-3-3 model, someone has to step up beyond what an average spot produces in a rotation, to the tune of 114-118 IP and a runs average of approximately 4.26 to 4.75.

This is significant for the 2010 Brewers because even theoretically, a 1-4-4-4-5 rotation is so much further from a passable starting pitching performance provided by a 3-3-3-3-3 / 1-2-3-4-5 model before those extra innings filled by a sixth starter are even considered.

1-4-4-4-5 MODEL: 761.09 IP, 432.34 R

This rotation needs 180 IP at a 2.65 runs average just to break even.

Think about that. That's better than a #1. That's like Tim Lincecum good. And that's even giving Gallardo the benefit of the doubt at a #1 (he was, in reality, the best #2 starter in the 2009 NL, but I basically felt like giving this team a benefit of the doubt somewhere, so I'm praising Gallardo). 

THINK ABOUT THAT. Seriously think about it. Imagine that Dave Bush's scar tissue is serious, his elbow issue is more serious than thought, and it affects his ability to bounce back in 2010. Imagine that the Brewers either bring back Looper or need to fill his spot. Suppan is here. Gallardo is here. Parra is somewhere, or else the Brewers need to acquire yet another arm to improve this rotation.

And the Brewers are going to somehow accomplish this through (a) best case scenario projections again, and (b) acquiring two pitchers?

***

Which leads me to point out that

(a) The cost of acquiring the type of pitching that will be needed to improve the club --  not stay the course from 2009 -- will require either (1) giving up more than 5% of their 2009 runs scored equation, or (2) spending more money than is currently available to the Brewers.

(b) The Brewers are arguably in a better position to acquire multiple, reasonable depth options without the goal of reaching an average starting rotation production level, and instead spend the available resources on ADDING offense to the 2010 Brewers.

(c) The logic behind "giving up some offense for pitching" is paradoxical, and therefore false, for the 2010 Brewers. It requires either spending more resources than are available, or depleting the offense beyond an acceptable level. 

My hypothesis is simple:

IF THE BREWERS SURRENDER OFFENSE FOR PITCHING THIS OFFSEASON -- GIVEN THE AVAILABLE RESOURCES -- THEY WILL END UP WITH AN AVERAGE TO BELOW AVERAGE OFFENSE WITHOUT AN ABOVE AVERAGE PITCHING STAFF.

My recommended plan of action is simple:

WORK TO IMPROVE THE ONE NOTABLY ABOVE AVERAGE ASPECT OF THE CLUB (OFFENSE), MAINTAIN THE OTHER ABOVE AVERAGE ASPECT (BULLPEN), WITH THE GOAL OF SACRIFICING THE BELOW AVERAGE ASPECT OF THE CLUB IN ORDER TO ATTEMPT TO HAVE ONE EXCEPTIONAL ASPECT OF THE CLUB.

Food for thought.

radio silence:

Lately I've been thinking about Doug Melvin's middle-man, crowd-pleasing position: that the Brewers will be willing to sacrifice some offense for some pitching.

I've wrapped my mind into a pretzel over how to determine the value of a run prevented against a run scored (in 2009 NL, the league teams allowed 9 more runs than they scored, which at a basic level means that a run prevented was very slightly more valuable than a run scored (to the tune of 1.01 RA: 1.00 RS).

On the surface, in a consistently depressed offensive environment, it seems like an intuitive way to approach the offseason; anyhow, the Brewers had one of the best offenses in the NL, scoring 67 more runs than the average offense (that's nearly 10% of the runs scored by an average offense), so it looks like the Brewers have a margin to dig into (if they give up, say, even 5% of their 2009 RS, that means around 746 RS for 2010, which could very well be an average to above average offense once again, if recent runs environment trends remain the same in the NL.

However, my thinking on this issue shifts once I think of two things:

(1) The Brewers' available resources

(2) The available free agent pitchers

(3) The likely talent required to obtain proven, average (and above) pitching via trade

(4) The level at which the Brewers' pitching staff needs to improve in order to become an average staff

Here's the deal; the average 2009 NL team allowed 727 runs, and the Brewers' 818 runs allowed requires the team to prevent 81 runs somewhere. The Brewers' bullpen isn't really a place to look for that task, given that the pen allowed 250 runs despite pitching over 540 IP (the average 2009 NL bullpen pitched 501.78 IP and allowed 242.6 runs, compared to 544 IP/250 RA for the Brewers' 2009 bullpen -- a rather exceptional performance when you consider the added requirement of 40 extra innings for the Crew's relievers).

So, the burden really falls with the rotation, which allowed approximately 74 more runs than the average NL rotation, which pitched 942.84 IP over 162 G, allowing 484 R. The Brewers starters' 891 IP and 568 R performance leaves a whole lot to be desired.

Which leads me to ask, how do you make up for that type of performance?

In order to figure this out, I needed to create two potential models for building an average rotation. The first model is simple: how does a true 1-2-3-4-5 rotation perform? The second is a bit more convoluted: how does a rotation of five #3 starters perform? Both models are unrealistic to some extent, because (a) no rotation actually goes 1-2-3-4-5 according to runs performance, and (b) no rotation actually goes 3-3-3-3-3 according to runs performance (and to show you the difference between rotation spots, the Brewers went 1-4-4-4-5 according to runs performance, which should demonstrate (a) the difficulty of using a #1 starter to offset multiple low rotation performances, and (b) the difference between #3 and #4 starters.)

Of course, the other issue is that relatively few rotations even reach 5 regular spots; there really wasn't even a such thing as a true #5 starter in the NL. The two true #5 candidates according to my survey has Manny Parra and Todd Wellemeyer sharing the dubious distinction of being the only two #5 starters according to runs performance (which is really, really awful), and Jeff Karstens lumped in as a swingman (only due to % of GS and IP, really; he was actually about as valuable as a mean #4).

That's my logic, to outline the problem. 

1-2-3-4-5 MODEL: 824.56 IP, 421.57 R

3-3-3-3-3 MODEL: 827.95 IP, 429.61 R

A couple of things should really jump out:

(a) A rotation of full-time, average #3 starters is almost exactly comparable to a 1-2-3-4-5 rotation. As a theoretical point, this should be powerful. I know actual available resources, scarcity of talent, difficulty of developing pitchers, etc., makes it extremely unlikely for a team to acquire five #3 starters, but it should re-open the debate about Doug Melvin's 2008-2009 offseason plan to acquire Braden Looper (a top #3 in 2008 NL) to work with Dave Bush (the top #3 in 2008 NL) and Manny Parra (another #3 in 2008 NL). With the gambles of Yovani Gallardo and the decreasing production of Jeff Suppan fully weighed, that rotation basically looked like ?-3-3-3-4, which isn't all that far off from a 3-3-3-3-3 model if all the starters remain healthy, make all their starts, and don't regress.

And there's your issue. After all, this is only a model, but really, think it through, thoroughly, and allow it to frame your logic of pitching acquisitions for a week. Then revisit it.

(b) Neither a 1-2-3-4-5 or 3-3-3-3-3 rotation completes the league average IP for NL 2009 starters, or fills 162 games. I will repeat it again for effect: Neither a 1-2-3-4-5 or 3-3-3-3-3 rotation completes the league average IP for NL 2009 starters, or fills 162 games.

I am fond of saying that a lot of baseball is just killing time, or biding time, and this seems to reflect that point: a team should expect to use 6 starters in this day and age, and probably have a swingman or two ready on top of that.

Either way, in a 1-2-3-4-5 model, or a 3-3-3-3-3 model, someone has to step up beyond what an average spot produces in a rotation, to the tune of 114-118 IP and a runs average of approximately 4.26 to 4.75.

This is significant for the 2010 Brewers because even theoretically, a 1-4-4-4-5 rotation is so much further from a passable starting pitching performance provided by a 3-3-3-3-3 / 1-2-3-4-5 model before those extra innings filled by a sixth starter are even considered.

1-4-4-4-5 MODEL: 761.09 IP, 432.34 R

This rotation needs 180 IP at a 2.65 runs average just to break even.

Think about that. That's better than a #1. That's like Tim Lincecum good. And that's even giving Gallardo the benefit of the doubt at a #1 (he was, in reality, the best #2 starter in the 2009 NL, but I basically felt like giving this team a benefit of the doubt somewhere, so I'm praising Gallardo). 

THINK ABOUT THAT. Seriously think about it. Imagine that Dave Bush's scar tissue is serious, his elbow issue is more serious than thought, and it affects his ability to bounce back in 2010. Imagine that the Brewers either bring back Looper or need to fill his spot. Suppan is here. Gallardo is here. Parra is somewhere, or else the Brewers need to acquire yet another arm to improve this rotation.

And the Brewers are going to somehow accomplish this through (a) best case scenario projections again, and (b) acquiring two pitchers?

***

Which leads me to point out that

(a) The cost of acquiring the type of pitching that will be needed to improve the club --  not stay the course from 2009 -- will require either (1) giving up more than 5% of their 2009 runs scored equation, or (2) spending more money than is currently available to the Brewers.

(b) The Brewers are arguably in a better position to acquire multiple, reasonable depth options without the goal of reaching an average starting rotation production level, and instead spend the available resources on ADDING offense to the 2010 Brewers.

(c) The logic behind "giving up some offense for pitching" is paradoxical, and therefore false, for the 2010 Brewers. It requires either spending more resources than are available, or depleting the offense beyond an acceptable level. 

My hypothesis is simple:

IF THE BREWERS SURRENDER OFFENSE FOR PITCHING THIS OFFSEASON -- GIVEN THE AVAILABLE RESOURCES -- THEY WILL END UP WITH AN AVERAGE TO BELOW AVERAGE OFFENSE WITHOUT AN ABOVE AVERAGE PITCHING STAFF.

My recommended plan of action is simple:

WORK TO IMPROVE THE ONE NOTABLY ABOVE AVERAGE ASPECT OF THE CLUB (OFFENSE), MAINTAIN THE OTHER ABOVE AVERAGE ASPECT (BULLPEN), WITH THE GOAL OF SACRIFICING THE BELOW AVERAGE ASPECT OF THE CLUB IN ORDER TO ATTEMPT TO HAVE ONE EXCEPTIONAL ASPECT OF THE CLUB.

Food for thought.

Jim Breen:

You make a fine point, but this does assume that Bush and Parra will maintain their 2009 production, which I find highly unlikely.

I agree completely Jim, but I'd be interested to know your opinion on the likely chances of improving / regressing given certain conditions, such as (a) Parra pitching more aggressively, (b) Parra gaining more endurance, (c) Bush regaining health, etc.

radio silence:

in 2009 NL, the league teams allowed 9 more runs than they scored, which at a basic level means that a run prevented was very slightly more valuable than a run scored (to the tune of 1.01 RA: 1.00 RS).

Wrong.  This is, in fact, impossible except for one thing.  That the NL scored 9 less runs that they allowed in interleague play.  That means the opposite is true of the American League.  Frankly, that has nothing to do with the value of a runs scored vs. a run allowed.  It simply means that the AL is a little bit better than the NL this year.

brewhawk:

radio silence:

in 2009 NL, the league teams allowed 9 more runs than they scored, which at a basic level means that a run prevented was very slightly more valuable than a run scored (to the tune of 1.01 RA: 1.00 RS).

Wrong.  This is, in fact, impossible except for one thing.  That the NL scored 9 less runs that they allowed in interleague play.  That means the opposite is true of the American League.  Frankly, that has nothing to do with the value of a runs scored vs. a run allowed.  It simply means that the AL is a little bit better than the NL this year.

Sure, and that also explains the slightly depressed runs environment which was the 2009 NL.

It's one of many factors that explains the value of runs in the NL.

radio silence:

Here's the deal; the average 2009 NL team allowed 727 runs, and the Brewers' 818 runs allowed requires the team to prevent 81 runs somewhere.

Or 91, if you are trying to take the difference of those two numbers.

radio silence:

brewhawk:

radio silence:

in 2009 NL, the league teams allowed 9 more runs than they scored, which at a basic level means that a run prevented was very slightly more valuable than a run scored (to the tune of 1.01 RA: 1.00 RS).

Wrong.  This is, in fact, impossible except for one thing.  That the NL scored 9 less runs that they allowed in interleague play.  That means the opposite is true of the American League.  Frankly, that has nothing to do with the value of a runs scored vs. a run allowed.  It simply means that the AL is a little bit better than the NL this year.

 

Sure, and that also explains the slightly depressed runs environment which was the 2009 NL.

It's one of many factors that explains the value of runs in the NL.

No.  There is not a depressed runs environment in the NL. 

There is, or rather was, possibly a depressed runs environment in interleague play for national league teams in 2009.  But it is not predictive of anything for 2010.

brewhawk:

radio silence:

Here's the deal; the average 2009 NL team allowed 727 runs, and the Brewers' 818 runs allowed requires the team to prevent 81 runs somewhere.

Or 91, if you are trying to take the difference of those two numbers.

Good catch.

brewhawk:

No.  There is not a depressed runs environment in the NL. 

There is, or rather was, possibly a depressed runs environment in interleague play for national league teams in 2009.  But it is not predictive of anything for 2010.

Yes, the NL is a depressed runs environment, compared to (1) the AL, and (2) the last 15 NL seasons.

The overall trend is a correction of peak mid-1990s runs environments to earlier 1980s/1990s environments. 

 For reference:

AL: 781 RS, 771 RA

NL: 718 RS, 727 RA

radio silence:

brewhawk:

No.  There is not a depressed runs environment in the NL. 

There is, or rather was, possibly a depressed runs environment in interleague play for national league teams in 2009.  But it is not predictive of anything for 2010.

 

Yes, the NL is a depressed runs environment, compared to (1) the AL, and (2) the last 15 NL seasons.

The overall trend is a correction of peak mid-1990s runs environments to earlier 1980s/1990s environments. 

1.  Interleague play is the depressed runs environment, not NL play.  And in that interleague environment, there is no way of knowing whether the run differential is the result of a lack of scoring runs, or a lack of preventing them, because the rules are different. 

2.  Last season, compared to the last 15 seasons, is not predictive of the runs environment in 2010, which is SOLELY dependent on the results of interleague play.  Interleague play favors the AL, in my opinion, because the AL rosters are easily adaptable to NL rules because a DH makes a very effective PH, whereas National League rosters lack the typically above average hitting DH roster spot to play by those rules. 

It truly has nothing to do with the value of a run scored vs. a run prevented.  And since the runs environment is created solely by the existence of interleague play, to extrapolate that result to all regular NL play would yield a false conclusion. 

 But it does matter for considering personnel moves in trades and free agency, which is the whole point of considering the runs environment.

radio silence:

 For reference:

 

AL: 781 RS, 771 RA

NL: 718 RS, 727 RA

Someone is going to have to explain to me how in all of baseball, there has been one more run scored than there has been allowed.

brewhawk:

radio silence:

 For reference:

 

AL: 781 RS, 771 RA

NL: 718 RS, 727 RA

Someone is going to have to explain to me how in all of baseball, there has been one more run scored than there has been allowed.

Those are taken from B-R's estimates. They round without decimals.

radio silence:

 But it does matter for considering personnel moves in trades and free agency, which is the whole point of considering the runs environment.

Yeah.  The NL doesn't go around looking for DH's.