SB Nation User Blog: Gopherballs

More Thoughts on Davies

Gopherballs — Thu, 28 May 2009 16:42:19 -0000

Davies 2009 (through May 27 start):

4.94 FIP
6.33 K/9, 3.86 BB/9, 1.64 K/BB, 8.9% swinging strike rate
1.23 HR/9, 11.8% HR/FB
17.4% LD%, 44.4% GB%, 288 BABIP

The K/BB is almost exactly his 2008 1.65 K/BB, so he has added a few strikeouts at the expense of a proportional number of walks. A good sign is that his swinging strike rate is up almost a 1% from 8.0% in 2008. As expected, his HR/FB rose from an unsustainable 6.7% in 2008 to a tick above average in 2009. He has also turned about 5% of batted balls from line drives (21.6% LD% 2008) into groundballs (38.8% GB% 2008).

Going from a flyball pitcher to a neutral pitcher helps, but Davies needs to push his K/BB closer to 2.0 to become an effective middle of the rotation starter.

There also seems to be a commonality between the teams from last September and his first start this year when Davies was dominant -- he faced the Mariners, White Sox, Twins, and White Sox again -- all teams that generally employ an aggressive approach at the plate and like to swing at pitches out of the zone. On the other hand, he tends to have mixed results against teams with better patience and discipline.

Davies for now is an adequate back of the rotation starter who still has some work to do to become a middle of the rotation starter. He has the raw tools to work with, but it will not come automatically. If it does not happen, he probably could become a pretty decent reliever who could go multiple innings -- relievers usually add velocity and control is not as important. With Cortes battling his own control problems on a second tour of duty in AA, the Royals do not have any major league ready pitching prospects to push anyone out of the rotation. But the clock is still ticking on Davies, as he will already be in his second year of salary arbitration this offseason, when teams start having to pay starters real money. He has come a long ways from 2006 and 2007 when he was one of the worst starters in baseball, but the Royals will have a tough call in the not-so-distant future if Davies does not show more improvement.

When Is A Small Sample Size No Longer Small?

Gopherballs — Fri, 22 May 2009 15:37:15 -0000

When Is A Small Sample Size No Longer Small?

Eric Seidman at Fangraphs has a nice article summarizing research on when sample sizes become (somewhat) reliable for hitters. The research suggests that the following stats begin to stabilize with the corresponding number of plate appearances:

50 PA: Swing %
100 PA: Contact Rate
150 PA: Strikeout Rate, Line Drive Rate, Pitches/PA
200 PA: Walk Rate, Groundball Rate, GB/FB
250 PA: Flyball Rate
300 PA: Home Run Rate, HR/FB
500 PA: OBP, SLG, OPS, 1B Rate, Popup Rate
550 PA: ISO
Did not stabilize by 650 PA (the cutoff): BA, BABIP

The underlying work was done by a well respected analyst who goes by the handle Pizza Cutter at the Statistically Speaking blog. His research on pitchers is here.

Brayan Pena DFA'd

Gopherballs — Sat, 25 Apr 2009 22:09:34 -0000

Brayan Pena DFA'd

Or so says one of the totally legitimate rumor sites. Considering Pena historically has struggled against RHP and has a questionable defensive reputation, he might actually clear waivers.

This clears a spot for the return of Jose Guillen.

So What Are K/9, BB/9, and K/BB Anyway?

Gopherballs — Wed, 25 Mar 2009 16:42:05 -0000

Continuing with the series on explaining useful statistics as requested in the excellent For All the Dummies post, this entry looks at three key pitching statistics that both scouting and statistical analysts consider very important: strikeout rate (K/9), walk rate (BB/9), and strikeout-to-walk ratio (K/BB, or alternatively K:BB). Again, the purpose here is provide a brief introduction and (likely oversimplified) explanation of these stats, not to start any debates or provide any in-depth analysis.

So what are K/9, BB/9, and K/BB? K/9 is strikeouts per nine innings. The formula is simple and the same as ERA except using strikeouts instead of earned runs: K/IP x 9. Some analysts prefer to use strikeouts per game pitched (K/G) or strikeout percentage (K%), which provide slightly sharper accuracy as they are based on strikeouts per batters faced instead of innings pitched. But given the minor differences, K/9’s simpler formula, and its wider usage, this post will use K/9.

BB/9 is bases on balls (walks) per nine innings. The formula, as you can probably guess, is BB/IP x 9. As with strikeouts, BB/G and BB% are alternatives.

K/BB is simply the ratio between a pitcher’s strikeouts and walks. K/BB is the formula too (total strikeouts divided by total walks).

Why do K/9 and BB/9 matter? K/9 and BB/9 are what are known as rate stats (just like ERA, FIP, batting average, OBP, etc.). Rate stats are excellent for evaluating and comparing players because they neutralize playing time (the number of innings pitched varies widely from pitcher to pitcher) and focus on the player’s level of performance. Counting stats, like raw totals of strikeouts and walks, usually need some context to provide meaning.

For example, consider that in 2008, Brett Tomko (41 Ks) recorded one more strikeout than Robinson Tejeda (40 Ks). This gives an initial (and misleading) impression that Tomko and Tejeda were roughly equal at striking hitters out. But Tejeda (39.1 IP) pitched 20 fewer innings than Tomko (60.2 IP). Using the K/9 rate stat immediately conveys the superiority of Tejeda (9.38 K/9) over Tomko (5.93 K/9) in this aspect of the game. Because K/9 rate provides this immediate context, it is much better than just listing the raw number of innings pitched and strikeouts. And, frankly, most readers (maybe even without realizing it) will probably end up trying to do the math in their head to compare the number of strikeouts per inning.

The same goes for BB/9.

The other great thing about strikeout and walk rates is that based on evaluating years and years of data, strikeout and walk rates correlate very strongly from year-to-year and are two of the most repeatable skills in baseball. In other words, past strikeout and walk rates are very good indicators of future strikeout and walk rates.

So what are good K/9 and BB/9 rates? As a preliminary note, what is considered good and bad strikeout and walk rates will depend on whether the pitcher is a starter or a reliever. On average, starters have lower K/9 rates and BB/9 rates than relievers. One of the main reasons is that starters generally do not throw as hard as relievers because starters must pace themselves, while relievers usually can throw at maximum effort given that they will only face a handful of batters. Throwing with less velocity and force generally will decrease strikeouts but improve control, leading to fewer walks.

The major league averages for the last three seasons are:

2008

Overall: 6.8 K/9, 3.4 BB/9

Starters: 6.5 K/9, 3.1 BB/9

Relievers: 7.5 K/9, 3.9 BB/9

2007

Overall: 6.7 K/9, 3.3 BB/9

Starters: 6.3 K/9, 3.1 BB/9

Relievers: 7.4 K/9, 3.7 BB/9

2006

Overall: 6.6 K/9, 3.3 BB/9

Starters: 6.2 K/9, 3.1 BB/9

Relievers: 7.3 K/9, 3.7 BB/9

Thus, in general, the average pitcher should be expected to post rates around 6.7 K/9 and 3.3 BB/9, with the average starter closer to 6.3 K/9 and 3.1 BB/9 and the average reliever closer to 7.4 K/9 and 3.7 BB/9.

Among the Royals regulars last year, Greinke (8.14 K/9) and Meche (7.83) led the starters, while Tejeda (9.38) and Soria (8.82) led the relievers in strikeout rate. Davies (5.65), Bannister (5.57), and Hochevar (5.02) were all below average.

As for walk rate, Greinke (2.49 BB/9) and Bannister (2.86) led the full-time starters, swingman Tomko (1.93 BB/9) led the regulars overall, while Peralta (2.39) and Soria (2.54) led the relievers. Meche was average (3.12), and Hochevar (3.28) and Davies (3.42) were moderately below average.

What about K/BB? As the major league averages listed above suggest, the average is right around 2.0 K/BB. In other words, the average pitcher should post twice as many strikeouts as walks. K/BB is useful because it is a good indicator of a pitcher’s command. A low BB/9 rate may show that a pitcher can throw strikes, but a good K/BB shows that a pitcher can throw quality strikes.

While not conclusive, good pitchers tend to post K/BB above 2.0 while bad pitchers tend to post ratios below 2.0. Last year, Soria (3.47 K/BB) and Greinke (3.27) led the Royals, while Duckworth (1.05) and Gobble (1.17) were the worst among the regulars. Among the other starters, Meche was solid (2.51), Bannister was just about average (1.95), while Davies (1.65) and Hochevar (1.53) fell below average. In the majors overall, the four pitchers who posted K/BB rates above 5.0 were four of the best pitchers in baseball – Halladay (5.28), Haren (5.15), Beckett (5.06), and Lee (5.00), while the bottom four were generally lousy – Greg Smith (1.28), Barry Zito (1.18), Kenny Rogers (1.15), and Daniel Cabrera (1.06).

So all you really need to look at is K/BB, right? While K and BB rates are the most important components in evaluating pitchers, a pitcher can still struggle if his K/BB is above 2.0, but it usually takes a disaster involving batted balls (allowing lots of home runs or line drives) to do so. For example, last year, Tomko (3.08 K/BB), Farnsworth (2.77 K/BB), and Peralta (2.71 K/BB) all posted good K/BB ratios, but struggled because they posted horrendous home run rates – Tomko (1.63 HR/9), Farnsworth (2.24 HR/9), Peralta (2.56 K/9). Home run rates are a topic for another day, but for context, an average home run rate is somewhere around 1.0 to 1.1 HR/9, and among AL pitchers with 50+ IP, Peralta finished dead last, Farnsworth was second to last, and Tomko was fifth to last.

On the flip side, a pitcher can still contribute with a poor K/BB if he can, among other things, limit home runs – despite below average K/BB rates, Davies (0.80 HR/9) and Hochevar (0.84 HR/9) both posted decent home run rates last year (although Hochevar is more likely to repeat that result than Davies for reasons that will become apparent when home run per fly ball rate is discussed).

In summary, the advantages provided by high strikeout or low walk rates should be obvious. Last year, Greinke had one of the best strikeout rates (8.14 K/9) among AL starters – 21.5% of all batters that Greinke faced made an out before even putting the ball in play – which is approximately 5% better than the average starter. In contrast, Horacio Ramirez (3.13 K/9) struck out only 8.5% of the batters he faced, one of the worst rates in all of baseball. Forcing an out without giving the hitter an opportunity to get a hit or even advance runners with an out is the best possible outcome for pitchers, as it essentially eliminates any chance of the other team scoring as a result of that at-bat.

Again, the same goes for BB/9 – the fewer walks surrendered, the fewer base runners, and the fewer opportunities to allow runs.

Kevin Goldstein Hates the Royals and Is Stupid

Gopherballs — Tue, 17 Mar 2009 17:35:28 -0000

Kevin Goldstein Hates the Royals and Is Stupid

Baseball Prospectus (subscription may be required) ranks the Royals minor league system No. 16, only one ahead of the no-talent Seattle Mariners.

So What is FIP Anyway?

Gopherballs — Wed, 04 Mar 2009 19:33:19 -0000

As discussed in the recent and most excellent For All the Dummies post, it appears that one reason some readers resist using statistics in evaluating baseball players is that the discussion of a particular stat often skips an explanation and goes straight to the conclusion. The following is an attempt to provide a brief introduction and background to Fielding Independent Pitching or FIP, one stat that frequently gets discussed on Royals Review (and has generally gained acceptance by those who study baseball stats). This post probably oversimplifies FIP a bit, but the main point here is to provide the general idea, not to start any debates, or link a bunch of research articles and studies (other than to note that current MLB consultant Tom Tango invented FIP and that I found this article helpful in explaining the basics of FIP). For further reading, see devil_fingers epic post of links.

In short, what is FIP? FIP is essentially a defense-neutral ERA that only considers three things demonstrably within a pitcher’s control – strikeouts, walks (plus HBP), and home runs allowed. Defense – which plays a large part in deciding whether a batted ball is a hit – is intentionally removed from the equation, and thus corrects one of the main problems with ERA.

FIP calculates runs allowed based on Ks, BBs, and HRs and then to make it more accessible, expresses the runs allowed on the same scale as ERA (the calculation for FIP actually sets the league average FIP as the league average ERA for that year). Thus, just like ERA, a 3.50 FIP is good, a 5.50 FIP is bad, and a 4.50 FIP is roughly average.

Strikeouts, walks, and home runs also happen to be three things that correlate strongly for a pitcher from year to year (strikeouts and walks more so than home runs). Because Ks, BBs, and HRs tend to repeat, so does FIP, making FIP a much better indicator for projecting a pitcher’s future performance than ERA.

What is the formula? Not that you would ever need to use it personally, the formula is (HR*13+(BB+HBP-IBB)*3-K*2)/IP, plus a factor (usually around 3.20) to scale FIP to match league average ERA for a given year.

Why does FIP use only Ks, BBs, and HRs, and not hits? The answer to this question is another question: what are the things actually within the pitcher’s control? The four major things under a pitcher’s control are strikeouts, walks (plus hit batters), home runs, and the type of batted ball allowed (groundball, flyball, and line drive) that does not leave the park. Obviously, defense is not one of them, and we know that whether a batted ball goes for a hit or an out depends in large part on the defense, so FIP intentionally removes hits (other than HRs) from the equation.

OK, that makes sense to not include hits, but what about the type of batted ball allowed, which pitchers do have some control over? This starts to get into the more advanced aspects, but the basic (and perhaps too simplistic) answer is that FIP generally assumes that a pitcher will have a league average batted ball profile (in other words, a pitcher will have league average groundball, flyball, and line drive rates) and thus allow a league average hit rate. This removes a bit of accuracy for simplicity, but for many if not the majority of pitchers, the balls in play rates fall close enough to the league averages to not create a big difference one way or other.

An easy solution is to simply look at the pitcher’s batted ball rates (groundball rate, flyball rate, line drive rate) to determine if a pitcher is performing at, above, or below his FIP. Other than to say that groundballs are generally good, flyballs are generally bad, and line drives are very bad, an evaluation of batted ball rates is probably best saved for a separate post. An advanced version of FIP that does factor in batted ball data is tRA, which again, probably deserves its own post at some point.

So how do the Royals stack up? As suggested in the comments, here is the list for the regulars in 2008. Please note that scoring was down last year throughout MLB, so the league average ERA and thus league average FIP was 4.32. In short, Ramon Ramirez, Soria, Greinke, and Tejeda were very good last year, while Bannister, Yabuta, Gobble, and Peralta had trouble.

Fangraphs, now with wOBA

Gopherballs — Wed, 26 Nov 2008 06:09:12 -0000

Fangraphs, now with wOBA

Fangraphs has added wOBA (weighted On-Base Average), one of the stats that improves on OPS by essentially weighing the relative value of on-base and slugging more accurately and scaling it to league average OBA (.335 is average, .400 is awesome, .300 is virtually replacement value). The link above explains the basics in more detail. The Fangraphs version of wOBA does incorporate SB/CS.

The Royals last year were 25th in MLB with a team wOBA of .314. Mike Aviles led the Royals (100 PA min.) with a .360 wOBA, with only David DeJesus (.355) and Alex Gordon (.344) also above average.

Jeremy Affeldt, Free Agent Bargain (No, Seriously)

Gopherballs — Mon, 27 Oct 2008 18:36:23 -0000

Jeremy Affeldt, Free Agent Bargain (No, Seriously)

I was kicking around drafting a post on impending free agent Jeremy Affeldt based on his surprisingly good numbers this year (in a hitter's park), but Fangraphs beat me to it. Affeldt added a couple miles per hour to his fastball and curve, which resulted in a jump in his strikeout rate and drop in his walk rate without affecting his high groundball rate.

While his past struggles might keep the price down, hard throwing lefty relievers who can get righties out tend to do well on the free agent market, so he will not come dirt cheap. But paying a high-leverage reliever mid-leverage reliever money could make for a nice bargain. Relief pitching is obviously not an immediate priority for the Royals, but if the Royals are not going to spend big elsewhere, this is the type of moderate deal that makes the team better without breaking the budget. Plus, adding bullpen depth would allow the Royals to move a reliever for help in a different area of need (which the free agent market likely overvalues).

Royals Swing First, Take Pitches Infrequently

Gopherballs — Fri, 24 Oct 2008 21:05:17 -0000

Royals Swing First, Take Pitches Infrequently

Fangraphs has a new article up on ranking teams by plate discipline, finding a strong correlation between a team's philosophy toward statistical analysis and the rate at which its players chase pitches out of the strike zone. The Royals finished 26th with German and Callaspo not playing enough to offset Pena and Olivo.

Meet Josh Bard, or Yes, the Market for John Buck Insurance is That Limited

Gopherballs — Tue, 07 Oct 2008 05:14:29 -0000

As expected, the San Diego Padres non-tendered catcher Josh Bard, making him an unrestricted free agent. With San Diego the last three seasons, Bard had tremendous success as a semi-regular in 2006 (partially with Boston where he infamously struggled catching Tim Wakefield’s knuckler), a good year as a regular in 2007 (San Diego), and a horrible, injury-filled season in 2008:

2006 Age 28 (284 PA) 333/404/502 .410 wOBA

2007 Age 29 (443 PA) 285/364/404 .344 wOBA

2008 Age 30 (178 PA) 202/279/270 .255 wOBA

Career 265/333/395

For those who only look at the most recent season’s results, you can stop reading now. For everyone else, Bard has very good plate discipline for a catcher and acceptable power. Bard spent half his playing time the last three years at Petco, an extreme pitcher’s park, which makes his 2006 and 2007 more impressive, and his 2008 more explainable. His 2008 stats were also hurt by severe bad luck – despite a healthy line drive rate of 21.6%, his BABIP was .230 (about .100 below his expected BABIP). Moving to the American League would hurt his stats, but the move away from Petco (career 245/346/366) would more than offset that.

Bard, who turns 31 before next opening day, probably will never come close to repeating his 2006 and is not good enough to start 135 games each year, but he is a good bet to bounce back and hit well enough to play regularly in a Gregg Zaun role. He is a switch hitter, and while he historically has been better from the right side (career 288/341/443), he is adequate from the left side for a catcher (256/330/376). Defensively, he has problems with the running game, but more importantly, does fine with blocking balls. He made $2.2 million last year, and considering he just cleared waivers after the Padres took him off their 40-man roster, he likely will sign in the $1-$2 million range next season. Bard would represent an immediate upgrade over Miguel Olivo and could serve as a stopgap starter if John Buck does not rebound from his lousy 2008.