Sabermetric Primer: SWIP

February 25, 2010 2:36pm CST

Sometimes we take things for granted. On the offensive side of the game, batting average was used for years as one of the "best" ways to evaluate a batter's performance. Eventually, slowly, people have come around to the idea that OBP is actually a much better way to judge a hitter's effectiveness. In this article, and the two that follow, I will add what I hope is another piece to that toolbox of yours by playing off an ingrained measure of the fantasy world.

WHIP: A Review
My contention is that WHIP, while a nice measure of a pitcher's general level of success, could easily be augmented by a similar to figure metric which I believe might be more reflective of a pitchers ability to dominate hitters. Perhaps I should start by defining WHIP before I move on to the "new" option - and sorry if this seems like a review of how to add two plus two.

WHIP has become all the rage in the past ten years - especially in fantasy baseball. You measure WHIP by adding walks and hits together and then dividing by innings pitched.

WHIP = (Walks + Hits) / Innings Pitched

WHIP is but one of the many measures that can be used to address the success, or lack thereof, for any pitcher. But I believe that the vagaries of hits allowed compromises the measure somewhat. Think about the type of factors that can effect whether or not a batted ball falls for a hit: a ball lost in the sun, a wind gust, a misstep by a fielder or a coach positioning players in the wrong spot can often offer a "false" indicator of a pitcher's performance by charging a hit to his ledger. Should a pitcher be punished because a ball gets lost in the sun if the fielder didn't have his shades on?

If you think about it, the pitcher is directly in control of few things during the game. Two of those "events" that a pitcher is directly able to influence the outcome of deal with whether or not he strikes out the batter or walks him (I will not worry about the vagaries of umpires strike zones or batters willingness to swing at pitches in our current discussion). About the only other event that the pitcher relies solely upon himself and not his fielders, coaches or terrain is the hit by pitch, but since WHIP does not count HBP, I won't consider it here either.

The fact of the matter is that the fewer times a batter puts the ball in play the fewer hits he gives up (how's that for undeniable logic?). Therefore, the pitcher who successfully limits the amount of balls put in play stands a greater chance of given up fewer baserunners, and by inference, fewer runs. So to address the issue of the importance of keeping batters off base while limiting free passes I invented the idea of SWIP.

In PART I of this three-part series I will define what SWIP is and why it is useful. In PART II, I will break down the SWIP leaders amongst starting pitchers from 2009. PART III I will deal with the plight of relief pitchers last season.


PART I - WHAT IS SWIP?

Following the simple methodology of WHIP, I have created a new measure of a pitchers dominance called SWIP.

S- Strikeouts (abbreviated as K)
W- Walks (abbreviated as BB)
IP- Innings Pitched

Numerically speaking, the formula for SWIP works along the same lines as WHIP. Therefore SWIP is determined by the following equation:

Strikeouts minus Walks divided by Innings Pitched equals SWIP.

SWIP = (K - BB) / IP

Another way to look at this is to say that for each positive result, the recording of an out in the form of a strikeout, the pitcher receives a (+1). For each negative encounter, in the form of a walk, he receives a (-1). Simple enough right? Here is an example of how you can figure out SWIP so you can see what I'm talking about (and yes, it really is as simple as it sounds).

EXAMPLE

Matt Cain had 171 Ks and 73 BBs in 217.2 IP for the Giants in 2009. Therefore:

(171-73) / 217.2
98 / 217.2
0.45 SWIP

Cain's SWIP for the 2009 season was therefore 0.45.

Though SWIP is recorded in the same manner as WHIP, the way to read the results is slightly different. Whereas the lower the WHIP the better one has performed, SWIP works in the opposite direction: the higher the SWIP the better.

Here is a rough estimate of what the results mean to help you to put things in perspective, a key if you will.

.90 and Up: Excellent season. Hall of Fame level.
.70 to .89: An all-star performance. Worthy of Cy Young consideration.
.50 to .69: Borderline all-star to decent starting pitcher. A guy you'd like to have on your staff.
.35 to .50: A guy who should be nothing more than the 3rd or 4th starter with his club.
.20 to .34: His major league days are likely numbered.
Below .20: Minor leaguer in training.

Before going any further here, let me briefly detail the performance of every hurler who tossed a pitch during the 2009 season. Last season major league pitchers put up the following numbers:

33,591 Strikeouts
16,620 Walks
43,272 IP

So in order to find out the major league average for SWIP during the 2009 season we simply plug the numbers into our simple equation.

SWIP = (K - BB) / IP
(33591-16620) / 43272
16971 / 43272
0.3922
SWIP = 0.39

Not surprisingly, last year's 0.39 mark mirrors the major league SWIP totals of the past few years though it is a 10-year high.

2009: 0.39 SWIP
2008: 0.38 SWIP
2007: 0.37 SWIP
2006: 0.37 SWIP
2005: 0.36 SWIP
2004: 0.36 SWIP
2003: 0.34 SWIP
2002: 0.35 SWIP
2001: 0.38 SWIP
2000: 0.30 SWIP

If you really care, and if you have read this far you probably do, here are the breakdowns by league for the 21st century.

Year	NL	AL
2009	0.40	0.39
2008	0.39	0.37
2007	0.37	0.37
2006	0.37	0.36
2005	0.36	0.35
2004	0.37	0.34
2003	0.36	0.33
2002	0.36	0.33
2001	0.41	0.36
2000	0.32	0.28

Conventional wisdom is that the NL is the "pitcher's league", and while SWIP certainly is too basic a metric to support or reject such a claim, on its surface at least SWIP appears to support that conventional wisdom - at least in slight.

Now that I have enumerated what SWIP is, and how it is figured, I will spend a brief moment detailing its major flaw.

Limit's of SWIP

As almost every metric out there that measures anything, SWIP is limited, in this case because of its simplicity. The major flaw of SWIP is that it favors pitchers with strikeout potential while often shortchanging those pitchers who might actually be "better" real world pitchers. We all know that this type of pitcher, the one who gets by more on guile than stuff (a guy like Jamie Moyer instantly leaps to mind), but SWIP is concerned with "stuff" so it favors pitchers with power arms.

Starting pitchers have multiple innings to set up batters and vary pitch sequences, not to mention the time needed to work themselves out of trouble. This freedom allows SP's to pitch with a variety of styles, all of which can be successful. At one end of the pitching spectrum there are "stuff" guys like Tim Lincecum who dominate hitters. However, you have hurlers like knuckleballer Tim Wakefield and soft-tossers like Aaron Cook who can be very successful as well. Obviously these pitchers do not record strikeouts at the same rate as their power pitching compatriots - their stuff simply isn't overpowering enough. Still, all pitching styles can be successful whether they rely upon the strikeout or the groundball if the pitcher has enough time to work out of jams and if he knows how to pitch. However, the more often that a pitcher can limit a batters ability to put the ball in the field of play, the more often he has "control" over the at-bat. So everything being equal, a pitcher is better off by not allowing the batter to hit the ball - it's as simple as that. Again, that doesn't mean there is only is only one path a pitcher must follow for success, there is just a better, or should I say safer way to achieve it. As a result, some of the pitchers that you will read about with a poor SWIP marks will have been successful in 2009 even if their path to success was slightly unconventional.

As far as relief pitchers, they operate under a different set of "rules." Relievers usually don't have multiple innings to set up batters and they often come into games with runners already on base. They don't have time to find their grove and work on touch pitches like change-ups and curveballs, they need to come in and throw strikes - immediately. As a result it appears that SWIP might be an even more useful tool to pass judgment on pitchers who rely mostly on "hard stuff" (fastballs, sliders and forkballs) than those soft throwers that I mentioned above. These hard throwing pitchers, as a general rule, tend to congregate more toward the bullpen than in the starting rotation since relievers can come in, throw gas, and not have to worry about pacing themselves to last multiple frames.

Therefore, SWIP can be termed a "dominance stat" in that it helps us to track which pitchers are best at limiting hitters ability to hit that ball. So in what you are about to read don't take the analysis to mean that I necessarily think that Ted Lilly is a better pitcher than Felix Hernandez because he had a better SWIP mark in 2009. Take the analysis for what it is, and remember this simple axiom...

Everything being equal, the pitcher who limits hitters ability to hit the ball, as well as limiting the free passes he allows, is the pitcher more likely to be consistent from year-to-year.
In PART II, you can read about how this SWIPmeasure is applied and we can go about trying to find a few pitchers who could be potential sleepers for the 2009 season.