I Left My Heart with Francisco Liriano

The Minnesota Twins Season Preview

What made the difference between the playoffs and also-ran status for the Minnesota Twins in 2006? The short answer is “Francisco Liriano.” Sure, one could attribute some credit to the piranha offense, to Joe Mauer’s continuing breakout, an all-world bullpen that got better as the year went on, or the MVP (ha!) performance of Justin Morneau. Nonetheless, nobody surprised or provided more value than Liriano, who literally pulled the team out of the second division, even without finishing the season in good health. More than 40 runs of VORP in limited innings is more than just a good second starter, it’s the type of kid-gloves season that the Mets dreamed of when the signed Pedro to a four year deal.

Yes, Liriano was brilliant. But that sort of hero-worship may paint an unnecessarily bleak picture of the 2007 season, when other young pitchers will have to stand in for Liriano and approximate his immense value, albeit with more durability. On the other hand, not just anyone can stand in for Johan Jr., and the rotation has an awful lot of value to replace with Radke and Liriano following different paths toward arm recovery. All in all, I find it hard to believe that 96 wins is in the team’s sites once again, unless somebody comes out of the blue to pull off another borderline-MVP act, ask Liriano did last year.

Generally, the line on the Twins going into 2007 is that the offense will have to step up and start outscoring the opposition a little bit more. On both sides of the runs scored/runs prevented ledger, the Twins have evolved into much more of a stars-and-scrubs lineup than the balanced one that carried them to three division titles from 2002-2004. In fact, starting with their first winning season in a decade in 2001, the Twins had different team leaders in OPS+ and ERA+ in each of the four seasons from 2001-2004. Since then, it has been the Joe Mauer and Johan Santana show in their respective categories with no intervention. Even perceptually, many regarded Torii Hunter as the team’s biggest star in the early part of the decade, and while he has not declined much, he is now probably less valuable than Mauer, Morneau, Santana, Nathan, and arguably Cuddyer. On the other hand, there are far less players within shouting distance of Hunter’s value than there were in 2002 or 2003, when Koskie, Mientkiewicz, Pierzynski, Guzman, Jones, Guardado, Radke, Mays, Reed, Ford, and Stewart took turns being well above average. Now, the Twins hardly have an above-average player at any position, except for Luis Castillo; nearly everyone else is either a superstar or filler.

Practically, this roster construction means that Mauer, Morneau, Hunter, and Cuddyer have to produce enough runs to compensate for a suddenly pedestrian pitching staff. Luckily, all four of them project to very strong performance through the PECOTA system, which has a very strong track record for offensive projection. Below is a table of positional player value from last year and their accompanying projected value for this year. Rather than comparing a player to himself or his direct replacement, I have roughly aggregated positional data to view, for instance, how many runs they stand to gain by playing Jason Bartlett for a full season at short rather than giving him 400 plate appearances, and letting Juan Castro suck away the rest. The playing time distribution is imperfect, since Rondell White did not get all of his plate appearances at DH last year, and will not have all of them in left field this year, but I think it gives a fair representation of playing time and potential value.

Defensively, I took Chris Dial’s defensive numbers from last year, which put them team 17 runs above average last year, then compared it to the change in Baseball Prospectus’s Davenport Translation from last year to this year, which tend to regress heavily to the mean, moderately projecting improvement or decline instead of big changes. As a result, the only major change comes from Cuddyer’s improvement in his second full season in left field- about three runs over the course of the season-, although the possibility exists that Hunter’s cascading injuries could further erode his defensive value.

At the plate, Mauer projects to remain one of the elite talents in all of the game; only injury should prevent him from becoming a perennial all-star at this point, and nearly every projection system agrees. Castillo, Punto, and the conglomeration of left fielders projects to a similar value from last year to this one, and modest changes at shortstop- full time for Bartlett- and CF- decline from Hunter- offset one another to a large degree. The potential for decline rests with the players who overachieved last year. Justin Morneau broke out dramatically, though PECOTA cannot overlook his difficulties the last two seasons. The positives arise from an odd source- not necessarily improvement, but from building value while regressing to the mean. That is to say that it would be exceedingly difficult for White or Ford to be nearly as bad as they were last year. Just reaching replacement level could be a 2-3 win improvement for the team, and PECOTA puts them very near that threshold. The only quibbles I have with the system are the playing time projections for Punto, Bartlett, and Kubel. Between the three of them, I believe that at least two will make it through an entire season as a starter, bringing the overall change in value up into the black rather than a net loss in runs above replacement level. The remaining plate appearance deficit will probably be filled by recycling undistinguished subs- the Tyners and Heinzes of the world- who will add little to no value over replacement level.

The run prevention side of the ledger poses an entirely different conundrum. Retaining Liriano’s services, even at a more reasonable level than his unbelievable 2.10 ERA from last year, would make it much easier to fill out the rotation. As it stands, the Twins have to make up 50 runs above replacement within the rotation, and the reinforcements they are asking to do it- Ramon Ortiz and Sidney Ponson- exemplify replacement level. Additionally, PECOTA sees last season’s bullpen success as non-replicable, especially for players like Nathan and Reyes who obliterated reasonable expectations. Nathan could conceivably replicate his value from last season, since he’s one of those truly elite relief pitchers who don’t fall victim to random performance fluctuations so frequently. Reyes is precisely the type of pitcher for whom projection metrics overcompensate; Nathan’s projection must show some attrition so that the system can also dismiss Reyes without being able to draw a quantitative boundary between the two. My point here is that I am not entirely confident that PECOTA recognizes that Nathan, Rincon, and Crain really are as good as they were last year. Perhaps they will not exactly repeat their value, but I think that the bullpen as a whole will come close to its value from last year rather than experiencing the several-win decline predicted here.

The rotation is a horse of a different color. First, a methodological note: I lumped Scott Baker’s innings into the bullpen, even though some of those innings will come from starts. Just like Garza, Bonser, Silva, Lohse, Liriano, and Perkins all saw part-time starting action last year, many of these pitchers will have multiple roles through the season. Whether those innings and prevented runs come as a starter or a reliever, it’s on the ledger, and counts toward victories or losses.

The real problem comes from replacing Liriano. Santana is still the best pitcher in the game, Bonser should reasonably stand in for Radke, Silva will be no worse than he was last year, and the gaggle of young pitchers (Garza, Perkins, Baker, Durbin, Slowey) will likely be better than they were last year at the end of the rotation. The problem is that Liriano got hurt and Ramon Ortiz has to pitch in his stead. Ortiz is not the worst pitcher in the majors- he’s certainly no Jose Lima. But replacing a legitimate phenom is not an easy task, and Ortiz could easily be a full five wins worse than he was last year, making a rotation that would otherwise approximate its 2006 performance look like a far worse unit.

With the offense and defense looking comparable, or even a bit better, and the bullpen getting somewhere close to its great season from last year, the onus is on the rotation to find a way to keep the team at the top of the division. PECOTA sees a grand total of 78 runs lost over replacement level, but due to the considerations I have listed above, I think that number will end up looking more like 50 runs below, almost the exact deficit created by replacing Liriano with what was found on the top of the scrap heap. And even though the Twins won 96 games last year, their run differential predicted a 93 win season, which is more indicative of future performance. Losing 50 runs from one year to the next usually amounts to about five fewer wins, putting the Twins at about 88 wins going into the season, which may not be enough to stay in the race with three other very strong teams in the division.

It is fair to note that the Twins have outperformed their run differential in four of Ron Gardenhire’s five seasons at the helm for a total of 20 additional wins (four per season). A consistently dominant bullpen contributes to that seeming statistical anomaly, and that does not figure to change drastically this season. Thus, predicted 89-91 wins seems very reasonable. A catastrophic scenario in which Mauer, Morneau, or Santana goes down for an extended period of time could easily shave four or five wins off of that projection, though, so the Twins have to strike while the iron is hot. Coming out of the gate well enough so as not to require a Promethean effort from a new edition of Liriano must be the case, because nobody on this roster will repeat that season.

Position	2006 Starter	PA	VORP	Dial DEF	2007 Starter	PA	PECOTA VORP	DEF (+/-)	(+/-)
C	Mauer	608	66.9	4	Mauer	623	53.4	4	-13.5
1B	Morneau	661	52	4	Morneau	612	28.5	3	-24.5
2B	Castillo	652	21.2	-2	Castillo	586	19.5	0	0.3
3B	Batista/Punto	719	5.1	8	Punto	480	7.1	8	2
SS	Castro/Bartlett	536	8.7	10	Bartlett	471	16.2	10	7.5
LF	Stewart/Kubel/Tyner	657	-4.2	NA	White	275	-0.6	NA	3.6
CF	Hunter	611	32.6	0	Hunter	538	20.9	0	-11.7
RF	Cuddyer	635	36.3	-7	Cuddyer	584	13.3	-4	-20
DH	White	355	-13.1	0	Kubel	405	11	0	24.1
C2	Redmond	191	9.4	NA	Redmond	184	3.3	NA	-6.1
OF4	Ford	255	-11.1	NA	Ford	303	0.9	NA	12
UTIL	Rodriguez	132	-3.3	NA	Cirillo	193	1	NA	4.3
UTIL	Nevin/Tiffee	103	-2.9	NA	Casilla	200	8	NA	10.9
TOTAL	NA	6115	197.6	17	NA	5454	182.5	21	-11.1

Pos	2006 Starter	IP	VORP	2007 Start	IP	VORP	(+/-)	IP Diff
SP1	Santana	233.7	79.6	Santana	218.3	64.5	-15.1
SP2	Radke	162.3	23.3	Bonser	155	15.6	-7.7
SP3	Liriano/Garza	171	51.8	Silva	156	9	-42.8
SP4	Silva	180.3	-7.6	Ortiz	110.7	3.6	11.2
SP5	Bonser/Baker	183.6	10.7	Garza	150.3	15.6	4.9
Swing	Lohse	63.7	-6.7	Perkins	116	5.6	12.3
		994.6	151.1		906.3	113.9	-37.2	-88.3
Pos	2006 Starter	IP	VORP	2007 Starter	IP	VORP	(+/-)
CP	Nathan	68.3	34.5	Nathan	62.3	24.3	-10.2
SU	Rincon	74.3	20.6	Rincon	61.3	14.8	-5.8
SU	Crain	76.7	21.1	Crain	66.3	17.1	-4
MR	Reyes	50.7	26.5	Reyes	64	13.5	-13
MR	Neshek	37	16.2	Neshek	76.7	19.9	3.7
LR	Guerrier	69.7	18.3	Guerrier	63	9.9	-8.4
LR	Eyre	59.3	4.2	Baker	140	12.3	8.1
		436	141.4		533.6	111.8	-29.6	97.6