"Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: 'There are three kinds of lies: lies, damned lies, and statistics.'" Mark Twain, Chapters from My Autobiography
Way back when the first fantasy baseball league, the Rotisserie Baseball League, was established in 1980 in New York, New York, the owners relied on ERA and WHIP as two of the four fundamental statistical categories used to evaluate pitching performance (along with wins and saves). Since those early days of roto fantasy baseball, those ratio stats have formed an important basis for the evaluation of pitchers to draft and pick up off the waiver wire when putting together an owner’s roster. While the “standard” fantasy league now employs a 5x5 format (with strikeouts being added as an additional counting stat), there is a growing albeit marginal at present impetus to substitute some other stat category in place of what must be acknowledged as original pitching performance evaluation benchmarks. Certainly, some of this movement can be attributed to a desire that exists within humanity to avoid stagnation. As Alvy Singer tells Annie Hall in the movie adorned with her moniker:
“A relationship, I think, is like a shark. It has to constantly move forward or it dies. And I think what we got on our hands is a dead shark.” Woody Allen, Annie Hall.
Certainly, no fantasy player, or more precisely, League Commissioner, wants to have a dead shark on his or her hands, right? So how do we go about shaking up the traditional format of fantasy baseball and still gauge the relative value of mound aces, whether starters or relievers, or that new breed of pitcher, the opener? Are there better, more predictive ratio statistics that could be turned to in determining the usefulness of an arm on a fantasy baseball squad? Certainly, if baseball knows how to do one thing well, it is to analyze the heck out of player performance on the field. Ever hear the term sabermetrics? It would be hard to avoid the influence of that sort of number crunching that was once dismissed as nerd-speak but now is much more readily accepted as a valid method of figuring out a player’s impact on the diamond.
One of the first such statistical measuring sticks to arrive on the scene was BABIP (batting average on balls in play). BABIP measures how often a ball hit by a batter in play goes for a hit of whatever sort. A “ball in play” is defined as an instance where a hitter’s plate appearance ends in a result other than a walk, strikeout, hit by pitch, sacrifice bunt, catcher interference or home run. On average, approximately 30% of all balls in play land safely for hits. Be advised, however, that three factors largely influence this statistic; those being luck, defense, and talent level. BABIP is used to assess both hitters and pitchers, with both sets of players’ average rates settling in at .300. In the case of hurlers, if the BABIP is below the league average, he could benefit from pitching in front of a superior defense, exhibit exceptional talent on the bump, or experience good luck. Or benefit from all three of those factors, or two of the three. Conversely, an inflated BABIP generally tends to indicate an unlucky pitcher, in general terms or having to pitch with a porous defense backing him up. Although some pitchers consistently have low or high BABIP ratios over their careers, frequently a pitcher with a bloated or severely depressed ratio will eventually see his BABIP trend toward the league norm. Beware that it takes a significant amount of pitching performance to establish a pitcher’s baseline BABIP, generally three seasons of action, so momentary fluctuations in the ratio are not necessarily good indicators that a pitcher is regressing either negatively or positively. Coupled with the fact that most pitcher’s BABIP ratios fall in the range of .290-.310, this ratio stat is poorly suited for use as a fantasy stat category. Not that it is not useful in determining a player’s projected future results, especially since once the ball leaves a hitter’s bat, it is almost entirely outside of the pitcher’s control, and those pitchers with extreme BABIP ratios, unless saddled with a truly horrific defense, will likely see their performance on balls in play return to normal levels over time.
Perhaps it would be helpful to examine another set of advanced metrics that provide analysis of the events that a pitcher does control while tossing the ball toward the plate. BABIP is driven by events that happen once a hitter makes contact and puts the ball into play. FIP and its close relative, xFIP take a different view of a pitcher’s efforts on the hill, emphasizing strikeouts, walks, homeruns and batters hit by pitches and their effect on a pitcher’s effectiveness at preventing runs from scoring. These two stat categories take into account those occurrences a pitcher has within his control. As a result, FIP indicates what a particular pitcher’s ERA would be assuming league average results over a period of time with regard to balls put in play and timing of events in a particular half inning. Timing or sequencing are significant as depending on the order that outs and hits take place while he is toeing the rubber, a pitcher could shut down an offense or be charged with multiple runs.
xFIP develops on the FIP idea by substituting the actual home run rate for a particular pitcher with an additional calculation taking into account league-average home run rate on fly balls, in conjunction with the hurler’s own fly ball rate. This change is meant to determine just how many crash-bombs a particular SP or RP should allow, all things adjusting to the league norm. xFIP, then, in addition to peeling away the influence of defense, luck and timing/sequencing from a pitcher’s performance, also attempts to strip away the randomness associated with the number of home runs a pitcher will allow over time.
Both FIP and xFIP are better used as predictive metrics, as their short-term forecasting tendency can be wildly fluctuating. In other words, these stats need larger sample sizes to be effectively extrapolative, as small inning trials are far less reliable than a half-season or better yet, a full season’s set of numbers for an arm on your fantasy staff. Also, neither metric is league nor park adjusted, working with MLB averages instead. Pitchers who have a “pitchers’ park” as their home stadium will consistently produce lower FIPs, all other things being equal (which they never are, but certain general groupings of like-skilled players can be compared with effective results, naturally). Of course, pitchers can outperform their FIP (or conversely, underperform their FIP) if they are extraordinarily successful at limiting homers while generating more than their “fair share” of fly balls (which go for hits less often than ground balls and line drives) or exceeding average rates of holding runners on base, but these arms are the exception.
Another predictive metric has arrived for us stat-heads to ponder and consider: SIERA. The acronym stands for Skill-Interactive ERA, and in addition to exploring the role of the pitcher-influenced events during a mound-appearance, namely strikeouts, walks, batters struck by pitches and home runs, it also considers how balls in play (groundballs and flyballs primarily) have an effect on a pitcher’s success. It factors in a great value for high-strikeout pitchers, while also recognizing that the ability to limit walks is a skill that benefits mound aces. While certainly an ERA estimation tool, it should be noted that SIERA works best as a predictive metric that is backward looking as opposed to forward prophesying. Still, because of the intention to evaluate the how and why of pitching, it has the potential to be both accurate and predictive in its approach to statistically probe the extraordinarily process of throwing a baseball. SIERA has the additional benefit of taking into account ballpark variance and the effect a particular stadium will have on a pitcher’s performance on the hill.
The problems with swapping out ERA and WHIP for any of the stat systems discussed above is that while ERA and WHIP report the actual ratios derived from a pitcher’s activities on the mound, FIP, xFIP and SIERA work best at predicting what lies in store for a pitcher based on past performance, either by the pitcher himself or the league as a whole. There is also the major impediment that none of the metrics above are available on the popular fantasy baseball platforms (Yahoo, ESPN, CBS), so even if you decide to do away with the “old tried and true” ERA and WHIP categories, finding other statistical scoring options is difficult.
Do not abandon all hope for means to shake up your fantasy baseball leagues, though, as other ratio stat categories are available to employ in your leagues. Above we noted that the events that the pitcher controls (strikeout, walks, home runs) are more predictive of a pitcher’s skill set. It is certainly an option to add a K/9, BB/9, K/BB or HR/9 stat category to track how effective or ineffective your pitchers are at keeping runners off the bases or from scoring. Analyzing a pitcher’s strikeout, walk or dinger-allowed rate is an effective manner to determine those arms that help themselves with their own efforts on the bump. There is no need to swap out the strikeout (K) counting stat, although if you are emphasizing pitchers’ control and dominance, then perhaps you have made a decision to stop simply counting the number of punchouts accumulated and are gazing instead at your pitchers’ overall skill at missing opposing hitters’ bats, or in the case of walks allowed, throwing strikes. Home run rate, as we noted above, is subject to variance based on fly ball rate, home park and luck, and tosses in a random factor that may not prove to be an accurate reflection of a pitcher’s abilities, and therefore is not seen in many league setups. (If you peruse the Two-Start Pitcher articles during the season on our site, you will see frequent references to all these ratio stats. End of self-promotion.)
Aside from avoiding “dead shark” stagnation, are there any other viable reasons for swapping out the traditional, long-standing stat categories ERA and/or WHIP for some of the newer developed sabermetric analysis tools? Certainly, as several come to mind:
- Attracting more sophisticated/cerebral owners. Baseball stat-lovers who have embraced the plethora of data available from the MLB and other sources will adore a fantasy scheme that rewards their willingness to delve deep below the surface of just simply determining how many hits, walks and runs a pitcher allows in his appearance(s).
- Making your league unique. As noted a few paragraphs above, it will be out of the norm to incorporate SIERA/FIP/BABIP or K/9 categories in the pitching columns. Aside from the rare custom league structure (generally requiring a fee paid to the site provider, ala a “commissioner league”), you are challenged to find a league that does not use both ERA and WHIP, even if the settings include rarer categories such as K/9 or K/BB ratios.
- Also as pointed out in the body of this article, the various prognostic stats discussed offer a far better means of predicting future performance of your pitchers than utilizing stats which simply report what has happened during the course of the season to date.
As a counterpoint to the first bullet point above, using the traditional ratio stats plays to the interests of newer fantasy owners, as most baseball fans have a fair grasp of what ERA and, although to a lesser extent, WHIP mean in terms of a pitcher’s production on the field. Making your owners comfortable in drafting and managing their teams should be one of your primary considerations as a Commissioner (aka benevolent dictator) of a league. Plus, it is a simple matter to incorporate ERA and WHIP as part of your settings on ALL platforms that you are likely to set up a league.
As ever, good luck and Godspeed in your fantasy efforts. Further questions about this article can be sent to this writer via e-mail addressed to ia@fantasyalarm.com