LIMA rises: elite non-closing relieversby Jeffrey Gross
March 28, 2012
It is a foregone conclusion at this point in spring training, with regular season games gearing up this week, but I had an epiphany this offseason in preparing for my fantasy auctions. And that epiphany is this: Now, more than ever, the LIMA (Low Investment Mound Aces) strategy, explained below, is the most viable option to winning your fantasy league. With robust pitching aplenty in the return of the era of the pitcher, why invest in elite arms?
Now let's get one thing straight: Leagues are not won at the auction/draft table. They can be lost there (coming back from a poor draft can be a devastatingly difficult uphill battle), but one cannot sit idly by after the league drafts and hope to go for gold. The waiver wire is a key component to a successful fantasy season, both in acquiring players for your team and preventing other teams from acquiring players that help theirs. My motto is that drafting is only 60 percent of the battle. The other 40 percent is a diligent and careful use of the waiver wire (in addition to trading).
That said, drafting is undeniably important. In the context of auctions, you are given a limited budget—usually $260—to acquire a 28-man roster out of 336 players out of the given baseball universe, give or take a few spots based on your league's size and roster requirements. Economizing is key, as is game theory. You need to maximize your resources by guessing not only what a player's expected relative value will be, but how other participants in your league are going to value said player.
If you have to pay $50 to acquire $38 of production, is that the wisest use of your limited budget (even assuming the value of concentrated production in a single roster spot)? Likewise, if the next highest valuator in the league values that $38 of production at $30, then do you really need to bid $38 to acquire him? This game of bid chicken, price enforcing and value-seeking is fun, complex, and nerve-wracking at times. Guessing all of these moving parts is no easy task, and a high-priced mistake can be devastating. Likewise, leaving $25 at the table could have meant the difference between having Matt Kemp instead or Drew Stubbs, or upgrading multiple positions.
Given all these complexities, limitations, and dangers of mistakes on spending, why spend money when you do not have to? Enter the LIMA strategy.
While most fantasy drafters tend to split their auction budget somewhere between 60/40 and 70/30 on hitting and pitching, the LIMA method splits the "excess" budget something like 85/15, or, in its most extreme form, 100/0.
Excess budget is defined by the amount of money you can spend on a player in excess of $1 per position, the minimum required to be kept unspent until all slots are filled. So in a league with $260 auction budgets and 28 players to be drafted per roster, the excess budget is $232.
The theory behind LIMA is that a good-enough-to-compete pitching staff can be cobbled together through playing match-ups, streaming and economizing late-round sleeper picks. LIMA revolves around the notion that pitching is so deep and volatile that it is a better use of resources to invest in elite hitting. LIMA requires a lot more micromanagement of your team, knowledge of up-and-coming pitchers, activity on the waiver wire, and the use of a value-seeking lens that ignores brand name and bankable production in favor of those "unsexy options" that no one else likes. (An example: drafting a pitcher like Edinson Volquez this year with the intent of streaming him at home.) LIMA is almost an extreme version of playing stars and scrubs, with almost all of your stars being hitters and almost all of your scrubs being pitchers.
The version of LIMA I usually play is to draft one ace and two to three cheap closers and surround them with $1 pitchers that I draft late. Last year, those $1 pitchers included Brandon Beachy, Michael Pineda, Javier Vazquez and Brandon McCarthy. In 2010, they included Colby Lewis, Phil Hughes and Kris Medlen (pre-injury). This year, with the exception of Yu Darvish and Anibal Sanchez in a couple of leagues, I almost exclusively drafted $1 pitchers.
Noting that the LIMA strategy has been around a long time—popularized and given its name by, Ron Shandler—why is this strategy more viable now than ever before? And why do I feel so comfortable drafting $1 pitchers across the board when I have never done anything so extreme before? The answer lies in recent trends in pitching.
It is no secret that over the past few years, the pitching standard has evolved towards a lower mean. Look at the 10-year trend in ERA, WHIP, strikeouts, walks, swings-and-misses (SwStr%) and first pitch strikes (F-Strike%):
A few things worth noting: While the league average ERA and WHIP have hardly been steady over the past 10 years, there is a clear difference between the pitching results from 2010 and 2011 compared to 2002-2009. From 2002 to 2009, the league ERA fluctuated from a low of 4.28 to a high of 4.53. That's a maximum variation of a quarter of a run per nine innings. Noting that, it is pretty huge that the league average ERA from 2009 to 2010 varied by 0.24 runs per nine innings. That's an ERA difference of 0.20 from the previous low mark in the 10-year sample, which came all the way back in 2002 (oddly, the season that Barry Bonds was walked to death). That's a huge outlier right there, and it was thought of as just that by many people heading into 2011.
But then 2011 happened, and the league average ERA dropped to 3.94. That's almost as big of a drop between 2010 and 2011 as the drop between 2010 and 2002 (again, the prior low league average ERA in the 10-year sample). The 2011 league average ERA of 3.94 is just over a third of a run per nine innings, or approximately a run per game, lower than the 2002-2009 league low. That's a pretty substantial 8 percent change in runs allowed against the lowest league average ERA between 2002 and 2008. The 3.94 major league average ERA is a 10-plus percent—almost half a run—difference from the 2002-2008 major league average ERA. That is pretty substantial.
A similar trend is observable in league average pitcher WHIPs, though to a much lesser degree. What is the cause of this? The answer is pretty murky, as detailed below
Second, walk rates have been more or less stable over this time. They have varied by no more than five percent in any year-to-year change, and the changes have moved in both directions an equal number of times in this set. You'll notice also a similar stability in first pitch strike rates, with only one season (2005) having a change in rate greater than one percent (and in that year, it was 1.2 percent). This is unsurprising given my rudimentary findings a few years ago that first pitch strike rates are highly relevant to a pitcher's overall walk rate.
Third, strikeout rates overall seem to be on the rise, despite a decline in swinging strike rates and the decline of big-strikeout, big home run hitters that defined the 1990s and early 2000s. Given the relationship between swinging strike percentage and strikeout percentage, this seems a bit odd. But this trend is not new. It is part of a larger overall trend that has been going on in baseball for more than 50 years, as eloquently detailed by Christina Kahrl at Baseball Prospectus in an excerpt from the upcoming book Extra Innings" (a sequel to the must-read Baseball Between the Numbers).
Kahrl points to the rise of micromanaging relief pitcher usage (bullpen assembly and management) against the decline of the "nine-inning starter" as one probable cause for this spike. Fresh arms that haven't been out there earlier in the game are harder to guess.
Steve Treder, in the 2011 Hardball Times Annual, had this to add about strikeout trends, as relayed to me in a recent email:
In July of this year, on his website The Diamond Appraised, the iconic sabermetrician Craig Wright published an excellent article titled “What is Behind the Offensive Decline?” Wright methodically considers various factors potentially causing scoring rates to decline over the past several years, including the crackdown on steroids. His take is that PED testing is at most a minor explanation, for two good reasons:
[First, a]s best we can tell, pitchers have made at least as much use of steroids and other PEDS as hitters. Thus to the extent that drug testing has removed “enhanced” performance from the game, there’s no reason to assume that the removal would necessarily result in a net reduction in scoring.
[Second, t]he serious, penalty-laden drug testing regime was introduced in 2004—a season in which scoring increased. Then following a decline in 2005, scoring climbed to even higher levels in 2006. There is no basis for concluding that the imposition of PED-testing would somehow result in a time-released inhibition of offense, and definitely not one still slowly unfolding seven years down the road.
Wright assesses other issues as well: improved fielding, a less lively ball, and a wider strike zone. In each case, he dispassionately weighs the available evidence and dismisses them as meaningful factors.
His conclusion is that the gradual decline in offense can be traced to two interrelated causes: a general improvement in the tactical approach of pitchers attacking the strike zone, and a lack of a corresponding general adjustment by hitters. Here’s the money quote:
'Let me just talk about the most common count, the one in every plate appearance, the 0-0 count, the first pitch. In the 20 years of data I have on this, there has been a fairly steady trend of the average batter taking more first pitches. The pitchers have wised up to this and starting around 2002 they began throwing more first-pitch strikes. That adjustment by the pitchers hasn’t changed the mentality of the average hitter. In fact, the general response of batters has been to take even more of the first pitches. They have set a new record in that category—in that 20-year database—each of the last four years. And again, that is with the pitchers throwing even more first-pitch strikes than they did back in the 1991-2001 period.
'For the first time in the 20-year history of the database, batters in 2010 are taking more than half of the first-pitch strikes thrown to them (.517). Twenty years ago it was .453. In a sample that covers the full variety of all hitters, a shift of 14.1% is a remarkable change.
'That means there are more plate appearances going to an 0-1 count without a fight. As many of you know, an 0-1 count radically affects the ultimate effectiveness of a plate appearance. In 2009, the final result of plate appearances after reaching an 0-1 count was an OPS of .629 as compared to .860 in the other plate appearances. It is not hard to overdo the wisdom of taking a tough first-pitch strike, and the general trend for batters may have crossed that line.'
I find this to be a penetrating insight on Wright’s part. The never-ending battle between pitcher and hitter always revolves around each probing to see what the other is willing to give, and taking advantage accordingly. Wright’s thesis is that through the 1990s batters were steadily gaining the edge by exploiting pitchers’ cautiousness, but since the early 2000s pitchers have been effectively turning the tables on hitters in the same manner. He makes quite a compelling argument.
That said, I remain a bit skeptical that the picture is as simple as Wright paints it. I tend to believe that trends so deep that they’re visible league-wide over a period of several years likely reflect changing underlying conditions beyond just the marketplace of competing tactical pitcher-batter approaches: in this case, given that first-pitch called strikes are the centerpiece of Wright’s profound observation, I’m not nearly as willing as he is to remove the umpires from the equation. My guess is that the tactical trends that Wright has identified are being enhanced by a cultural shift in umpiring, with the men in blue increasingly inclined to give the pitcher the benefit of the doubt on a borderline first-pitch take (and likely on other takes as well).
There is no shortage of recent literature on the recent trends in pitching, and the specific causes are still up for debate. Regardless of exactly why pitchers are showing better results, the fact is that they are, and some research shows that this trend might be more sustainable than one would have thought at the end of the 2010 season.
Let me point out some additional key observations. First, the difference between the top starters and the average major league starter has not changed by too much over the past five years. Below is a chart that compares the difference between the ERA, WHIP, strikeout rate and walk rate of the cumulative top 10 qualified starting pitchers by ERA against the major league average:
You'll notice that not too much looks out of place comparing 2010 and 2011 to 2007-2009. Likewise, the variation between qualified starting pitcher ERA and WHIP, a measure of "clustering," do not seem to indicate any substantial trend over time either:
Getting the top guys is still much better for your team than grabbing the league average pitcher, assuming you can successfully identify them (there is a lot of turnover).
So why LIMA? Does ignoring pitching really make more sense when the differentials between the top and middle are not overly substantial at best?
The answer, looping back to Kahrl's Baseball Prospectus article, lies in the increased availability of elite non-closing relievers. Check out the stats of qualified relievers over the past decade:
For whatever reason—better micromanagement and usage of bullpens, better evaluation in bullpen composition, or something else—it seems that over the past few years, relievers as a whole are putting up more elite numbers. I have written about the value of drafting elite, non-closing relievers on many occasions.
Saves are only one statistic. There are plenty of players like Jonny Venters, Tyler Clippard, David Robertson, Addison Reed, Mike Adams and Kenley Jansen out there who have the potential to be worth double-digit dollar values by the end of the season despite a lack of saves. The additional upshot of some of these guys, like Sergio Romo and Sean Marshall (before Ryan Madson's injury) is that they are often "next in line" for saves. Some can even net you five to 10 wins to boot.
As more and more of these guys are available, or as more of these elite non-closer relievers are more often and better used, they become more and more useful for fantasy teams. No one really drafts these guys, and fewer people pick up emerging elite non-closing relievers in the middle of the season. You'll of course have to watch the leverage usage statistics (if your elite non-closing reliever is coming in only to face guys like Adrian Gonzalez, you might want to opt for someone else), but why not invest a dollar or two into a couple of players who can arguably turn Bud Norris into Tim Lincecum.
It might burn an extra roster space, but with innings caps and only so many hitters you can start in one day, maximization can take on many forms. They include playing the LIMA method by drafting elite non-closing relievers to boot. I call this method of maximization LIMAPER (LIMA Plus Elite Non-closing Relievers).
Winning in fantasy baseball is about adjusting, and it is essential to adjust to this trend to win in more competitive leagues. Every advantage counts. Fantasy baseball is about finding market inefficiencies and exploiting them. The biggest names in elite non-closing relievers—Jonny Venters (60 percent Yahoo ownership rate), Kenley Jansen (74 percent), Tyler Clippard (44 percent) and Aroldis Chapman (45 percent)—are are pretty heavily owned, but outside the name brands, and the farther the player is from vulturing saves, the lower the ownership rates.
Notice the drop in ownership rates to even Mike Adams (31 percent), David Robertson (32 percent), Sergio Romo (33 percent), Fernando Salas (33 percent), and Vinnie Pestano (26 percent), and then compare them to Koji Uehara (8 percent), Eric O'Flaherty (13 percent), Joaquin Benoit (13 percent) and Kris Medlen (1 percent, and he might even take the fifth starter role for April in Tim Hudson's absence).
Take Oliver's projections and THT Forecast's customizable auction price guide set for a deep, 12-team league with nine pitching spots. Assume a $260 budget heavily skewed 70/30 in favor of hitting. Guys like Adams, Benoit, Uehara, Romo and Joel Peralta are pegged to be worth $6-8, while Craig Kimbrell, arguably the most valuable closer in baseball for 2012, is worth only about $12. That's more than double what Oliver thinks you can expect out of Johan Santana, but he's owned in over three-quarters of leagues.
Elite non-closing relievers are a market inefficiency. How much is one of these guys going to cost you? A buck? Two bucks? Maybe three to five dollars if you get a top name early on? Plenty will be available on the waiver wire. Some, like Luke Gregerson, Rafael Soriano and Hong-Chih Kuo last year, will surely bust. But you can drop them without much worry. They did not cost you much of your budget—maybe 1 percent of it. And there will be plenty of guys like Takashi Saito and Kerry Wood to replace them with even in league-only formats.
Let's come full circle and note a few things:
Hitting is about concentrating production. You can play the match-ups and try to Frankenstein your way to success, but the best strategy is to install the best players possible at each hitting position. It is harder to "stream" hitters day-to-day than it is pitchers. By using the LIMAPEN strategy, you are freeing a greater amount of your resources to confidently invest in positional upgrades. It could mean the difference between making a few Mark Reynoldses into Pablo Sandovals, or upgrading Asdrubal Cabreras into Hanley Ramirez. It's harder to play the matchups to replicate Ramirez at shortstop than it is to combine breakout potential players like Juan Nicasio, Mike Minor, Brian Matusz, Jeff Samardzija, Luke Hocheaver or flawed/risky pitchers on the cheap like Brandon Morrow, Bud Norris, Carlos Zambrano with elite non-closing relievers like Robertson, Adams, and Benoit to replicate the production of Daniel Hudson, James Shields or C.J. Wilson types.
It cannot be any riskier than drafting Adam Wainwright or Josh Johnson at plus-market rates. You might have more wins risk, but there's plenty of Alfredo Aceveses in the season. Further considering the lower bust/flop rates of hitters compared to pitchers, it makes more sense to invest in the former over the latter. And the LIMAPER strategy enables you to do that with less worry. Combine that with the "do not pay for saves" strategy* and you have a lot of extra money to spend maximizing lesser risks.
*I do advocate paying something for saves, just not too much, because you by and large cannot punt any category in a non-H2H format and reasonably expect to end up on top. I suppose you could even "overpay" for saves from guys like Craig Kimbrell, who will bolster your bottom pitching line in addition to netting you saves, if you want to play it safest. I covered this issue, very crudely, in one of my very first fantasy articles.
Now if only I had written this article a month ago... at least you still have plenty of time to scour the waiver wire!
Check out part two of this article for the quantification of this argument.
As always, leave the love/hate in the comments below.
Jeffrey Gross is an attorney (and die-hard Cubs fan) who currently resides in the suburbs of Chicago. In addition to writing for The Hardball Times, he also writes about craft beer as part of a side project blog titled "saBEERmetrics." He previously worked for The Daily Illini and Northern Star newspapers as a film critic and sportswriter (respectively). You can reach him by email at saBEERmetrics AT gmail DOT com.
<< Return to Article