Forecaster's Toolbox: Pitchers

Click here to return to the Education page.

Click here to return to the Forecaster's Toolbox home page.

General Concepts | Batting | Pitching | Minor League Toolbox | Strategy Toolbox | Team Toolbox


Strikeouts and Walks

Fundamental skills

Unreliable pitching performance is a fallacy driven by the practice of attempting to project pitching stats using gauges that are poor evaluators of skill.

How can we better evaluate pitching skill? We can start with the three statistical categories that are generally unaffected by external factors. These three stats capture the outcome of an individual pitcher versus batter match-up without regard to supporting offense, defense or bullpen:  

Walks Allowed, Strikeouts and Ground Balls

Even with only these stats to observe, there is a wealth of insight that these measures can provide. 

Control rate (Ctl, bb/9), or opposition walks per game 

BB allowed x 9 / IP

Measures how many walks a pitcher allows per game equivalent. BENCHMARK: The best pitchers will have bb/9 of 2.8 or less.

Dominance rate (Dom, k/9), or opposition strikeouts/game 

Strikeouts recorded x 9 / IP

Measures how many strikeouts a pitcher allows per game equivalent. BENCHMARK: The best pitchers will have k/9 levels of 7.0 or higher.

Command ratio (Cmd)

(Strikeouts / Walks)

A measure of a pitcher’s ability to get the ball over the plate. There is no more fundamental a skill than this, and so it is used as a leading indicator to project future rises and falls in other gauges, such as ERA. BENCHMARKS: Baseball’s best pitchers will have ratios in excess of 3.0. Pitchers with ratios less than 1.0—indicating that they walk more batters than they strike out—have virtually no potential for long-term success. If you make no other changes in your approach to drafting pitchers, limiting your focus to only pitchers with a command ratio of 2.5 or better will substantially improve your odds of success. 

Command ratio as a leading indicator

The ability to get the ball over the plate—command of the strike zone—is one of the best leading indicators for future performance. Command ratio (K/BB) can be used to project potential in ERA as well as other skills gauges.

1. Research indicates that there is a high correlation between a pitcher’s Cmd ratio and his ERA.

Earned Run Average

Command     2011    2012    2013    2014    2015
0.0 - 1.0   5.45    6.22    5.98    6.81    6.31
1.1 - 1.5   4.84    5.03    4.91    4.97    5.23
1.6 - 2.0   4.35    4.48    4.42    4.37    4.54
2.1 - 2.5   3.89    4.09    3.96    3.80    4.19
2.6 - 3.0   3.66    3.88    3.81    3.78    3.87
3.1 - 3.5   3.58    3.67    3.46    3.43    3.51
3.6 - 4.0   3.00    3.34    3.32    3.16    3.56
4.1+        2.95    3.12    2.86    2.92    3.07

On the pitching flipside, the number of arms comprising the 4.1+ group has nearly doubled since 2012. That year, 58 pitchers made up this group; in 2014 there were 93 and 90 this year.

We can create percentage plays for the different levels:

For Cmd % with ERA of

Levels of    3.50-    4.50+   
0.0 - 1.0     0%        87% 
1.1 - 1.5     7%        67%
1.6 - 2.0     7%        57%
2.1 - 2.5    19%        35%
2.6 - 3.0    26%        25%
3.1 +        53%         5% 

Pitchers who maintain a Cmd over 2.5 have a high probability of long-term success. For fantasy drafting purposes, it is best to avoid pitchers with sub-2.0 ratios. Avoid bullpen closers if they have a ratio less than 2.5.

2. A pitcher’s Command in tandem with Dominance (strikeout rate) provides even greater predictive abilities. 

Earned Run Average

Command    -5.6 Dom    5.6+ Dom
0.0-0.9        5.36     5.99
1.0-1.4        4.94     5.03
1.5-1.9        4.67     4.47
2.0-2.4        4.32     4.08
2.5-2.9        4.21     3.88
3.0-3.9        4.04     3.46
4.0+           4.12     2.96

This helps to highlight the limited upside potential of soft-tossers with pinpoint control. The extra dominance makes a huge difference. 

3. Research also suggests that there is a strong correlation between a pitcher’s command ratio and his propensity to win ballgames. Over three quarters of those with ratios over 3.0 post winning records, and the collective W/L record of those command artists is nearly .600.

The command/winning correlation holds up in both leagues, although the effect was more pronounced in the NL. Over four times more NL hurlers than AL hurlers had Cmd over 3.0, and higher ratios were required in the NL to maintain good winning percentages. A ratio between 2.0 and 2.9 was good enough for a winning record for over 70% of AL pitchers, but that level in the NL generated an above-.500 mark slightly more than half the time.

In short, in order to have at least a 70% chance of drafting a pitcher with a winning record, you must target NL pitchers with at least a 3.0 command ratio. To achieve the same odds in the AL, a 2.0 command ratio will suffice.

Swinging strike rate as leading indicator (Stephen Nickrand)

An emerging indicator for predicting starting pitching performance is swinging strike rate (SwK%), which measures the percentage of total pitches against which a batter swings and misses. SwK% can help us validate and forecast a SP’s Dominance (K/9) rate, which in turn allows us to identify surgers and faders with greater accuracy.

Follow these rules of thumb when targeting starting pitchers based on SwK%: SwK% baselines for SP are 8.0% in AL, 8.4% in NL; Expected Dom (xDom) can be estimated from SwK%; and a pitcher’s individual SwK% does not regress to league norms.

The few starters per year who have a 12.0% or higher SwK% are near-locks to have a 9.0 Dom or greater. In contrast, starters with a 7.0% or lower SwK% have nearly no chance at posting even an average Dom. Finally, use an 8.5% SwK% as an acceptable threshold when searching for SP based on this metric; raise it to 9.5% to begin to find SwK% difference-makers.

Fastball velocity and Dominance rate (Stephen Nickrand)

It is intuitive that an increase in fastball velocity for starting pitchers leads to more strikeouts. But how much? We analyzed the historical link between fastball velocity and Dominance (K/9) rate. Among the findings: 

The vast majority of SP with significant fastball velocity gains 

•    experience a significant Dom gain during the same season. 

•    are likely to give back those gains during the following season.

•    are likely to increase their Dom the following season, but the magnitude of the Dom increase usually is small. 

The vast majority of SP with significant fastball velocity losses

•    are likely to experience a significant Dom decrease during the same season. 

Those SP with significant fastball velocity losses from one season to the next are just as likely to experience a fastball velocity or Dom increase as they are to experience a fastball or Dom decrease, and the amounts of the increase/decrease are nearly identical.

First-pitch strike rate as leading indicator (Stephen Nickrand)The measurement of a pitcher’s rate of first-pitch strikes (FpK%) can help us validate and forecast a pitcher’s Control (BB/9) rate. As first-pitch strike rate increases, walks are very likely to go down, and WHIP will follow. As it goes up, walks are likely to increase, as will WHIP.  So if you’re wondering if a pitcher’s newfound good control is likely to hold, check out his FpK%.

The FpK% baseline is 60% for starting pitchers and does not vary significantly by league. Expected Ctl (xCtl) can be estimated from FpK%, and a starting pitcher’s individual FpK% does not regress to league norms. BENCHMARKS: Elite pitchers will have a FpK% above 68% and most of them will have a Ctl below 2.0. Avoid pitchers with a FpK% below 55%, as they are likely to have a Ctl at or above 4.0.

Power/contact rating 

(BB + K) / IP

Measures the level by which a pitcher allows balls to be put into play. In general, extreme power pitchers can be successful even with poor defensive teams. Power pitchers tend to have greater longevity in the game. Contact pitchers with poor defenses behind them are high risks to have poor W-L records and  ERA. BENCHMARKS: A level of 1.13+ describes pure throwers. A level of .93 or less describes high contact pitchers.  

Balls in Play

Balls in play (BIP)

(Batters faced – (BB + HBP + SAC)) + H – K 

The total number of batted balls that are hit fair, both hits and outs. An analysis of how these balls are hit—on the ground, in the air, hits, outs, etc.—can provide analytical insight, from player skill levels to the impact of luck on statistical output.

Batting average on balls in play  (Voros McCracken) 

(H – HR) / (Batters faced – (BB + HBP + SAC)) + H – K – HR

Abbreviated as BABIP; also called hit rate (H%). The percent of balls hit into the field of play that fall for hits. BENCHMARK: The league average is 30%, which is also the level that individual performances will regress to on a year to year basis. Any +/- variance of 3% or more can affect a pitcher’s ERA. 

BABIP as a leading indicator  (Voros McCracken)

In 2000, Voros McCracken published a study that concluded that “there is little if any difference among major league pitchers in their ability to prevent hits on balls hit in the field of play.” His assertion was that, while a Johan Santana would have a better ability to prevent a batter from getting wood on a ball, or perhaps keeping the ball in the park, once that ball was hit in the field of play, the probability of it falling for a hit was virtually no different than for any other pitcher.

Among the findings in his study were:

•     There is little correlation between what a pitcher does one year in the stat and what he will do the next. This is not true with other significant stats (BB, K, HR).

•     You can better predict a pitcher’s hits per balls in play from the rate of the rest of the pitcher’s team than from the pitcher’s own rate.

This last point brings a team’s defense into the picture. It begs the question, when a batter gets a hit, is it because the pitcher made a bad pitch, the batter took a good swing, or the defense was not positioned correctly?  

Pitchers will often post hit rates per balls-in-play that are far off from the league average, but then revert to the mean the following year. As such, we can use that mean to project the direction of a pitcher’s ERA.  

Subsequent research has shown that ground ball or fly ball propensity has some impact on this rate.  

Hit rate (See Batting average on balls in play)  

Opposition batting average (OBA)

Hits allowed / (Batters faced – (BB + HBP + SAC))

The batting average achieved by opposing batters against a pitcher. BENCHMARKS: The best pitchers will have levels less than .250; the worst pitchers levels more than .300.

Opposition on base average (OOB)

(Hits allowed + BB) / ((Batters faced – (BB + HBP + SAC)) + Hits allowed + BB)

The on base average achieved by opposing batters against a pitcher. BENCHMARK: The best pitchers will have levels less than .300; the worst pitchers levels more than .375.

Walks plus hits divided by innings pitched (WHIP)

Essentially the same measure as opposition on base average, but used for Rotisserie purposes. BENCHMARKS: A WHIP of less than 1.20 is considered top level; more than 1.50 indicative of poor performance. Levels less than 1.00—allowing fewer runners than IP—represent extraordinary performance and are rarely maintained over time.

Ground ball, line drive, fly ball percentage (G/L/F)

The percentage of all balls-in-play that are hit on the ground, in the air and as line drives. For a pitcher, the ability to keep the ball on the ground can contribute to his statistical output exceeding his demonstrated skill level.

Ground ball tendency as a leading indicator (John Burnson)

Ground ball pitchers tend to give up fewer HRs than do fly ball pitchers. There is also evidence that GB pitchers have higher hit rates. In other words, a ground ball has a higher chance of being a hit than does a fly ball that is not out of the park. 

GB pitchers have lower strikeout rates. We should be more forgiving of a low strikeout rate (under 5.5 K/9) if it belongs to an extreme ground ball pitcher.

GB pitchers have a lower ERA but a higher WHIP than do fly ball pitchers. On balance, GB pitchers come out ahead, even when considering strikeouts, because a lower ERA also leads to more wins.

Groundball and strikeout tendencies as indicators  

(Mike Dranchak)

Pitchers were assembled into 9 groups based on the following profiles (minimum 23 starts in 2005):

Profile        Ground Ball Rate
Ground Ball      higher than 47%
Neutral          42% to 47%
Fly Ball         less than 42%

Profile         Strikeout Rate (k/9)
Strikeout         higher than 6.6 k/9
Average           5.4 to 6.6 k/9
Soft-Tosser       less than 5.4 k/9 

Findings: Pitchers with higher strikeout rates had better ERAs and WHIPs than pitchers with lower strikeout rates, regardless of ground ball profile. However, for pitchers with similar strikeout rates, those with higher ground ball rates had better ERAs and WHIPs than those with lower ground ball rates.

Pitchers with higher strikeout rates tended to strand more baserunners than those with lower K rates. Fly ball pitchers tended to strand fewer runners than their GB or neutral counterparts within their strikeout profile.

Ground ball pitchers (especially those who lacked high-dominance) yielded more home runs per fly ball than did fly ball pitchers. However, the ERA risk was mitigated by the fact that ground ball pitchers (by definition) gave up fewer fly balls to begin with.

Extreme GB/FB pitchers  (Patrick Davitt)

Among pitchers with normal strikeout levels, extreme GB pitchers (>3–7% of all batters faced) have  ERAs about 0.4 runs lower than normal-GB% pitchers but only slight WHIP advantages. Extreme FB% pitchers (32% FB) show no ERA benefits.

Among High-K (>=24% of BF), however, extreme GBers have ERAs about 0.5 runs lower than normal-GB pitchers, and WHIPs about five points lower. Extreme FB% pitchers have ERAs about 0.2 runs lower than normal-FB pitchers, and WHIPs about 10 points lower.

Revisting Flyballs  (Jason Collette)

The increased emphasis on defensive positioning is often associated with infield shifting, but the same data also influences how outfielders are positioned. Some managers are positioning OFs more aggressively than just the customary few steps per a right- or left-handed swinging batter. found that five of the top 10 defensive efficiency teams in 2013 —OAK, STL, MIA, LAA and KC—also had parks among the top 10 in HR suppression. 

Before dismissing flyball pitchers as toxic assets, pay more attention to park factors and OF defensive talent. In particular, be a little more willing to roster fly ball pitchers who pitch both in front of good defensive OFs and in good pitchers’ parks.

Line drive percentage as a leading indicator  (Seth Samuels)

Also beyond a pitcher’s control is the percentage of balls-in-play that are line drives. Line drives do the most damage; from 1994-2003, here were the expected hit rates and number of total bases per type of BIP.

            |——------— Type of BIP —--—-—|

                 GB    FB     LD
H%              26%    23%    56%
Total bases    0.29   0.57   0.80

Despite the damage done by LDs, pitchers do not have any innate skill to avoid them. There is little relationship between a pitcher’s LD% one year and his rate the next year. All rates tend to regress towards a mean of 22.6%. 

However, GB pitchers do have a slight ability to prevent LDs (21.7%) and extreme GB hurlers even moreso (18.5%). Extreme FB pitchers have a slight ability to prevent LDs (21.1%) as well.

Home run to fly ball rate (hr/f)

HR / FB    

The percent of fly balls that are hit for home runs. 

hr/f as a leading indicator  (John Burnson)

McCracken’s work focused on “balls in play,” omitting home runs from the study. However, pitchers also do not have much control over the percentage of fly balls that turn into HR. Research shows that there is an underlying rate of HR as a percentage of fly balls of about 10%. A pitcher’s HR/FB rate will vary each year but always tends to regress to that 10%. The element that pitchers do have control over is the number of fly balls they allow. That is the underlying skill or deficiency that controls their HR rate.

Pitchers who keep the ball out of the air more often correlate well with Roto value. 

Opposition home runs per game (hr/9)

(HR Allowed x 9 / IP)

Also, expected opposition HR rate = (FB x  0.10) x 9 / IP

Measures how many HR a pitcher allows per game equivalent. Since FB tend to go yard at about a 10% rate, we can also estimate this rate off of fly balls. BENCHMARK: The best pitchers will have hr/9 levels of less than 1.0.


Expected earned run average  (xERA)

Gill and Reeve version: (.575 x H [per 9 IP]) + (.94 x HR [per 9 IP]) + (.28 x BB [per 9 IP]) – (.01 x K [per 9 IP]) – Normalizing Factor 

John Burnson version (used in this book):

(xER x 9)/IP, where xER is defined as xER% x (FB/10) + (1-xS%) x  [0.3 x (BIP – FB/10) + BB] where xER% = 0.96 – (0.0284 x (GB/FB)) and xS% = (64.5 + (K/9 x 1.2) – (BB/9 x (BB/9 + 1)) / 20) + ((0.0012 x (GB%^2)) – (0.001 x GB%) - 2.4)

xERA represents the an equivalent of what a pitcher’s real ERA might be, calculated solely with skills-based measures. It is not influenced by situation-dependent factors.  

Expected ERA variance


The variance between a pitcher’s ERA and his xERA is a measure of over or underachievement. A positive variance indicates the potential for a pitcher’s ERA to rise. A negative variance indicates the potential for ERA improvement. BENCHMARK: Discount variances that are less than 0.50. Any variance more than 1.00 (one run per game) is regarded as a indicator of future change.  

Projected xERA or projected ERA? 

Which should we be using to forecast a pitcher’s ERA?  Projected xERA is more accurate for looking ahead on a purely skills basis. Projected ERA includes situation-dependent events—bullpen support, park factors, etc.—which are reflected better by ERA. The optimal approach is to use both gauges as a range of expectation for forecasting purposes.

Strand rate (S%)

(H + BB – ER) / (H + BB – HR)

Measures the percentage of allowed runners a pitcher strands (earned runs only), which incorporates both individual pitcher skill and bullpen effectiveness. BENCHMARKS: The most adept at stranding runners will have S% levels over 75%. Those with rates over 80% will have artificially low ERAs which will be prone to relapse. Levels below 65% will inflate ERA but have a high probability of regression. 

Expected strand rate (Michael Weddell)

73.935 + K/9 - 0.116 * (BB/9*(BB/9+1)) + (0.0047 * GB%^2 - 0.3385 * GB%) + (MAX(2,MIN(4,IP/G))/2-1) + (0.82 if left-handed)

This formula is based on three core skills:  strikeouts per nine innings, walks per nine innings, and groundballs per balls in play, with adjustments for whether the pitcher is a starter or reliever (measured by IP/G), and his handedness.  

Strand rate as a leading indicator (Ed DeCaria)

Strand rate often regresses/rebounds toward past rates (usually 69-74%), resulting in Year 2 ERA changes:

% of Pitchers with Year 2 Regression/Rebound 

Y1 S%    RP     SP     LR
<60%    100%    94%    94%
65       81%    74%    88%
70       53%    48%    65%
75       55%    85%   100%
80       80%   100%   100%
85      100%   100%   100%

Typical ERA Regression/Rebound in Year 2

Y1 S%     RP       SP       LR
<60%    -2.54    -2.03    -2.79
65      -1.00    -0.64    -0.93
70      -0.10    -0.05    -0.44
75       0.24     0.54     0.75
80       1.15     1.36     2.29
85       1.71     2.21      n/a 

Starting pitchers (SP) have a narrower range of strand rate outcomes than do relievers (RP) or swingmen/long relievers (LR). Relief pitchers with Y1 strand rates of <=67% or >=78% are likely to experience a +/- ERA regression in Y2. Starters and swingmen/long relievers with Y1 strand rates of <=65% or >=75% are likely to experience a +/- ERA regression in Y2. Pitchers with strand rates that deviate more than a few points off of their individual expected strand rates are likely to experience some degree of ERA regression in Y2. Over-performing (or “lucky”) pitchers are more likely than underperforming (or “unlucky”) pitchers to see such a correction.


Projecting/chasing wins

There are five events that need to occur in order for a pitcher to post a single win...

1.    He must pitch well, allowing few runs.

2.     The offense must score enough runs.

3.     The defense must successfully field all batted balls.

4.     The bullpen must hold the lead.

5.    The manager must leave the pitcher in for 5 innings, and not remove him if the team is still behind.

Of these five events, only one is within the control of the pitcher. As such, projecting or chasing wins based on skills alone can be an exercise in futility.

Home field advantage (John Burnson)

A 2006 study found that home starting pitchers get credited with a win in 38% of their outings. Visiting team starters are credited with a win in 33% of their outings.



Batters faced per game (Craig Wright) 

((Batters faced – (BB + HBP + SAC)) + H + BB) / G

A measure of pitcher usage and one of the leading indicators for potential pitcher burnout. 


Research suggests that there is a finite number of innings in a pitcher’s arm. This number varies by pitcher, by development cycle, and by pitching style and repertoire. We can measure a pitcher’s potential for future arm problems and/or reduced effectiveness (burnout): 

Sharp increases in usage from one year to the next. Common wisdom has suggested that pitchers who significantly increase their workload from one year to the next are candidates for burnout symptoms. This has often been called the Verducci Effect, after writer Tom Verducci. analyst Michael Weddell tested pitchers with sharp workload increases during the period 1988-2008 and found that no such effect exists. 

Starters’ overuse. Consistent “batters faced per game” (BF/G) levels of 28.0 or higher, combined with consistent seasonal IP totals of 200 or more may indicate burnout potential, especially with pitchers younger than 25. Within a season, a BF/G of more than 30.0 with a projected IP total of 200 may indicate a late season fade.

Relievers’ overuse. Warning flags should be up for relievers who post in excess of 100 IP in a season, while averaging fewer than 2 IP per outing.

When focusing solely on minor league pitchers, research results are striking:

Stamina: Virtually every minor league pitcher who had a BF/G of 28.5 or more in one season experienced a drop-off in BF/G the following year. Many were unable to ever duplicate that previous level of durability.

Performance: Most pitchers experienced an associated drop-off in their BPVs in the years following the 28.5 BF/G season. Some were able to salvage their effectiveness later on by moving to the bullpen.

Protecting young pitchers (Craig Wright)

There is a link between some degree of eventual arm trouble and a history of heavy workloads in a pitcher’s formative years. Some recommendations from this research: 

Teenagers (A-ball): No 200 IP seasons and no BF/G over 28.5 in any 150 IP span. No starts on three days rest.

Ages 20-22: Average no more than 105 pitches per start with a single game ceiling of 130 pitches.

Ages 23-24: Average no more than 110 pitches per start with a single game ceiling of 140 pitches.

When possible, a young starter should be introduced to the majors in long relief before he goes into the rotation.

Overall Performance Analysis

Base Performance Value (BPV)

((Dominance Rate - 5.0) x 18) 

+  ((4.0 - Walk Rate) x 27)) 

+ (Ground ball rate as a whole number - 40%) 

A single value that describes a player’s overall raw skill level. This is more useful than traditional statistical gauges to track player performance trends and project future statistical output. The formula combines the individual raw skills of power, control and the ability to keep the ball down in the zone, all characteristics that are unaffected by most external factors. In tandem with a pitcher’s strand rate, it provides a more complete picture of the elements that contribute to ERA, and therefore serves as an accurate tool to project likely changes in ERA. BENCHMARKS: A BPV of 50 is the minimum level required for long-term success. The elite of the bullpen aces will have BPVs in excess of 100 and it is rare for these stoppers to enjoy long term success with consistent levels under 75.

Base Performance Index (BPX)

BPV scaled to league average to account for year-to-year fluctuations in league-wide statistical performance. It’s a snapshot of a player’s overall skills compared to an average player. BENCHMARK: A level of 100 means a player had a league-average BPV in that given season. 

Runs above replacement (RAR)

An estimate of the number of runs a player contributes above a “replacement level” player. 

Batters create runs; pitchers save runs. But are batters and pitchers who have comparable RAR levels truly equal in value? Pitchers might be considered to have higher value. Saving an additional run is more important than producing an additional run. A pitcher who throws a shutout is guaranteed to win that game, whereas no matter how many runs a batter produces, his team can still lose given poor pitching support.

To calculate RAR for pitchers:

1.    Start with the replacement level league ERA.

2.    Subtract the pitcher’s ERA. (To calculate projected RAR, use the pitcher’s xERA.)

3.    Multiply by number of games played, calculated as plate appearances (IP x 4.34) divided by 38.

4.    Multiply the resulting RAR level by 1.08 to account for the variance between earned runs and total runs.


1    LHers tend to peak about a year after RHers.

2.    LHers post only 15% of the total saves. Typically, LHers are reserved for specialist roles so few are frontline closers.

3.    RHers have slightly better command and HR rate.

4.    There is no significant variance in ERA.

5.    On an overall skills basis, RHers have ~6% advantage.

Skill-Specific Aging Patterns for Pitchers  (Ed DeCaria)

Baseball forecasters obsess over “peak age” of player performance because we must understand player ascent toward and decline from that peak to predict future value. Most published aging analyses are done using composite estimates of value such as OPS or linear weights. By contrast, fantasy GMs are typically more concerned with category-specific player value (K, ERA, WHIP, etc.). We can better forecast what matters most by analyzing peak age of individual baseball skills rather than overall player value.

For pitchers, prior research has shown that pitcher value peaks somewhere in the late 20s to early 30s. But how does aging affect each demonstrable pitching skill?

Strikeout rate (k/9): Declines fairly linearly beginning at age 25.

Walk rate (bb/9): Improves until age 25 and holds somewhat steady until age 29, at which point it begins to steadily worsen. Deteriorating k/9 and bb/9 rates result in inefficiency, as it requires far more pitches to get an out. For starting pitchers, this affects the ability to pitch deep into games.

Innings Pitched per game (IP/G): Among starters, it improves slightly until age 27, then tails off considerably with age, costing pitchers nearly one full IP/G by age 33 and one more by age 39.

Hit rate (H%): Among pitchers, H% appears to increase slowly but steadily as pitchers age, to the tune of .002-.003 points per year.

Strand rate (S%): Very similar to hit rate, except strand rate decreases with age rather than increasing. GB%/LD%/FB%: Line drives increase steadily from age 24 onward, and outfield flies increase beginning at age 31. Because 70%+ of line drives fall for hits, and 10%+ of fly balls become home runs, this spells trouble for aging pitchers.

Home runs per fly ball (hr/f): As each year passes, a higher percentage of a pitcher’s fly balls become home runs allowed increases with age.

Catchers’ effect on pitching (Thomas Hanrahan)

A typical catcher handles a pitching staff better after having been with a club for a few years. Research has shown that there is an improvement in team ERA of approximately 0.37 runs from a catcher’s rookie season to his prime years with a club. Expect a pitcher’s ERA to be higher than expected if he is throwing to a rookie backstop. 

First productive season (Michael Weddell)

To find those starting pitchers who are about to post their first productive season in the majors (10 wins, 150 IP, ERA of 4.00 or less), look for:

•     Pitchers entering their age 23-26 seasons, especially those about to pitch their age 25 season. 

•     Pitchers who already have good skills, shown by an xERA in the prior year of 4.25 or less. 

•     Pitchers coming off of at least a partial season in the majors without a major health problem. 

•     To the extent that one speculates on pitchers who are one skill away, look for pitchers who only need to improve their control (bb/9).

Overall pitching breakout profile (Brandon Kruse)  

A breakout performance is defined here as one where a player posts a Rotisserie value of $20 or higher after having never acheieved $10 previously. These criteria are primarily used to validate an apparent breakout in the current season but may also be used carefully to project a potential breakout for an upcoming season.

•     Age 27 or younger

•     Minimum 5.6 Dom, 2.0 Cmd, 1.1 hr/9 and 50 BPV

•     Maximum 30% hit rate

•     Minimum 71% strand rate

•     Starters should have a H% no greater than the previous year; relievers should show improved command

•     Maximum xERA of 4.00

Career year drop-off (Rick Wilton)

Research shows that a pitcher’s post-career year drop-off, on average, looks like this:

•     ERA increases by 1.00

•     WHIP increases by 0.14.

•     Nearly 6 fewer wins

Pitchers crossing leagues (Bob Berger)

The AL has higher league-wide ERA and lower K/9 when compared to the NL. Fantasy owners should consider adjusting their ERA, WHIP, and K/9 expectations for pitchers moving to the “other” league. Pitchers moving to the NL may perform better than expected based on their recent career trends; pitchers moving to the AL may perform worse than expected.



There are six events that need to occur in order for a relief pitcher to post a single save:

1.    The starting pitcher and middle relievers must pitch well.

2.    The offense must score enough runs.

3.    It must be a reasonably close game.

4.    The manager must put the pitcher in for a save opportunity.

5.    The pitcher must pitch well and hold the lead.

6.    The manager must let him finish the game.

Of these six events, only one is within the control of the relief pitcher. As such, projecting saves for a reliever has little to do with skill and a lot to do with opportunity. However, pitchers with excellent skills may create opportunity for themselves.

Saves conversion rate (Sv%)

Saves / Save Opportunities

The percentage of save opportunities that are successfully converted. BENCHMARK: We look for a minimum 80% for long-term success.

Leverage index (LI) (Tom Tango)

Leverage index measures the amount of swing in the possible change in win probability indexed against an average value of 1.00. Thus, relievers who come into games in various situations create a composite score and if that average score is higher than 1.00, then their manager is showing enough confidence in them to try to win games with them. If the average score is below 1.00, then the manager is using them, but not showing nearly as much confidence that they can win games. 

Saves chances and wins (Patrick Davitt)

Some fantasy owners think that good teams get more saves because they generate more wins. Other owners think that poor teams get more saves because more of their wins are by narrow margins. The “good-team” side is probably on firmer ground, though there are enough exceptions that we should be cautious about drawing broad inferences.

The 2014 study confirmed what Craig Neuman found years earlier: The argument “more wins leads to more saves” is generally correct. Over five studied seasons, the percentage of wins that were saved (Sv%W) was about 50%, and half of all team-seasons fell in the Sv%W range of 48%-56%. As a result, high-saves seasons were more common for high-win teams.

That wins-saves connection for individual team-seasons was much less solid, however, and we observed many outliers. Data for individual team-seasons showed wide ranges of both Sv%W and actual saves. 

Finally, higher-win teams do indeed get more blowout wins, but while poorer teams had a higher percentage (73%) of close wins (three runs or fewer) than better teams (56%), good teams’ higher number of wins meant they still had more close wins, more save opportunities and more saves, again with many outliers among individual team-seasons. 

Origin of closers

History has long maintained that ace closers are not easily recognizable early on in their careers, so that every season does see its share of the unexpected. Shawn Tolleson, A.J. Ramos, Roberto Osuna, Ken Giles, Carson Smith, Wade Davis, Brad Ziegler…  who would have thought it a year ago? 

Accepted facts, all of which have some element of truth:

•    You cannot find major league closers from pitchers who were closers in the minors. 

•    Closers begin their careers as starters. 

•    Closers are converted set-up men. 

•    Closers are pitchers who were unable to develop a third effective pitch. 

More simply, closers are a product of circumstance.

Are the minor leagues a place to look at all?

From 1990-2004, there were 280 twenty-save seasons in Double-A and Triple-A, accomplished by 254 pitchers.

Of those 254, only 46 ever made it to the majors at all.

Of those 46, only 13 ever saved 20 games in a season.

Of those 13, only 5 ever posted more than one 20-save season in the majors: John Wetteland, Mark Wohlers, Ricky Bottalico, Braden Looper and Francisco Cordero.

Five out of 254 pitchers, over 15 years—a rate of 2%. 

One of the reasons that minor league closers rarely become major league closers is because, in general, they do not get enough innings in the minors to sufficiently develop their arms into big-league caliber.

In fact, organizations do not look at minor league closing performance seriously, assigning that role to pitchers who they do not see as legitimate prospects. The average age of minor league closers over the past decade has been 27.5.

Elements of saves success

The task of finding future closing potential comes down to looking at two elements:

Talent: The raw skills to mow down hitters for short periods of time. Optimal BPVs over 100, but not under 75.

Opportunity: The more important element, yet the one that pitchers have no control over.

There are pitchers that have Talent, but not Opportunity. These pitchers are not given a chance to close for a variety of reasons (e.g. being blocked by a solid front-liner in the pen, being left-handed, etc.), but are good to own because they will not likely hurt your pitching staff. You just can’t count on them for saves, at least not in the near term.

There are pitchers that have Opportunity, but not Talent. MLB managers decide who to give the ball to in the 9th inning based on their own perceptions about what skills are required to succeed, even if those perceived “skills” don’t translate into acceptable metrics. 

Those pitchers without the metrics may have some initial short-term success, but their long-term prognosis is poor and they are high risks to your roster. Classic examples of the short life span of these types of pitchers include Matt Karchner, Heath Slocumb, Ryan Kohlmeier, Dan Miceli, Joe Borowski and Danny Kolb. More recent examples include Tom Wilhelmsen, Kevin Gregg and Jim Johnson.  

Closers’ job retention (Michael Weddell)

Of pitchers with 20 or more saves in one year, only 67.5% of these closers earned 20 or more saves the following year.  The variables that best predicted whether a closer would avoid this attrition: 

•    Saves history: Career saves was the most important factor. 

•    Age: Closers are most likely to keep their jobs at age 27.  For long-time closers, their growing career saves totals more than offset the negative impact of their advanced ages. Older closers without a long history of racking up saves tend to be bad candidates for retaining their roles.

•    Performance: Actual performance, measured by ERA+, was of only minor importance. 

•    Being right-handed: Increased the odds of retaining the closer’s role by 9% over left-handers.

How well can we predict which closers will keep their jobs? Of the 10 best closers during 1989-2007, 90% saved at least 20 games during the following season. Of the 10 worst bets, only 20% saved at least 20 games the next year.

Closer volatility history

        Closers                 Closers
Year    Drafted     Avg R$      Failed      Failure %      New Sources
2008      32        $17.78        10          31%              11
2009      28        $17.56         9          32%              13
2010      28        $16.96         7          25%              13
2011      30        $15.47        11          37%               8
2012      29        $15.28        19          66%              18
2013      29        $15.55         9          31%              13
2014      28        $15.54        11          39%              15
2015      29        $14.79        13          45%              16

Drafted refers to the number of saves sources purchased in both LABR and Tout Wars experts leagues each year. These only include relievers drafted for at least $10*, specifically for saves speculation. Avg R$ refers to the average purchase price of these pitchers in the AL-only and NL-only leagues. Failed is the number (and percentage) of saves sources drafted that did not return at least 50% of their value that year. The failures include those that lost their value due to ineffectiveness, injury or managerial decision. New Sources are arms that were drafted for less than $10 (if  drafted at all) but finished with at least double-digit saves.

The failed saves investments in 2015 were Cody Allen, Joaquin Benoit, Brett Cecil, Steve Cishek, Sean Doolittle, Neftali Feliz, Greg Holland, Jenrry Mejia, Joe Nathan, Glen Perkins, Addison Reed, Fernando Rodney and Drew Storen. The new sources in 2015 were John Axford, Santiago Casilla, Wade Davis, Jeurys Famila, Ken Giles, Jason Grilli, Kevin Jepsen, Jim Johnson, Andrew Miller, Roberto Osuna, A.J. Ramos, Carson Smith, Joakim Soria, Shawn Tolleson, Tom Wilhelmsen and Brad Ziegler.

*The 2015 season represented the most justifiably risk-averse year since we began tracking closer volatility in 1999. Pre-season pricing dropped to the lowest on record and for the first time included five frontline closers whose average draft price was below $10 (Boxberger, Doolittle, Mejia, Nathan, Rodriguez). The 45% failure rate was the second highest in 12 years (behind the 66% in 2012) as was the number of new sources of saves. However, five of those 16 new sources failed to hold the job during the year (Axford, Grilli, Johnson, Soria and Wilhelmsen).

Closers and multi-year performance (Patrick Davitt)

A team having an “established closer”—even a successful one—in a given year does not affect how many of that team’s wins are saved in the next year. However, a top closer (40-plus saves) in a given year has a significantly greater chance to retain his role in the subsequent season.

Research of saves and wins data over several seasons found that the percentage of wins that are saved is consistently 50%-54%, irrespective of whether the saves were concentrated in the hands of a “top closer” or passed around to the dreaded “committee” of lesser closers. But it also found that about two-thirds of high-save closers reprised their roles the next season, while three-quarters of low-save closers did not. Moreover, closers who held the role for two or three straight seasons averaged 34 saves per season while closers new to the role averaged 27.

BPV as a leading indicator (Doug Dennis)

Research has shown that base performance value (BPV) is an excellent indicator of long-term success as a closer. Here are 20-plus saves seasons, by year:


Year    No.    100+    75+    <75
1999    26      27%    54%    46%
2000    24      25%    54%    46%
2001    25      56%    80%    20%
2002    25      60%    72%    28%
2003    25      36%    64%    36%
2004    23      61%    61%    39%
2005    25      36%    64%    36%
2006    25      52%    72%    28%
2007    23      52%    74%    26%
MEAN    25      45%    66%    34%

Though 20-saves success with a 75+ BPV is only a 66% percentage play in any given year, the below-75 group is composed of closers who are rarely able to repeat the feat in the following season:

           No. with    No. who followed up

Year    BPV < 75    20+ saves <75 BPV
1999     12              2
2000     11              2
2001      5              2
2002      7              3
2003      9              3
2004      9              2
2005      9              1
2006      7              3
2007      6              0

Other Relievers

Projecting holds (Doug Dennis)

Here are some general rules of thumb for identifying pitchers who might be in line to accumulate holds. The percentages represent the portion of 2003’s top holds leaders who fell into the category noted.

1.    Left-handed set-up men with excellent BPIs. (43%)

2.    A “go-to” right-handed set-up man with excellent BPIs. This is the one set-up RHer that a manager turns to with a small lead in the 7th or 8th innings. These pitchers also tend to vulture wins. (43%, but 6 of the top 9)

3.    Excellent BPIs, but not a firm role as the main LHed or RHed set-up man. Roles change during the season; cream rises to the top. Relievers projected to post great BPIs often overtake lesser set-up men in-season. (14%)

Reliever efficiency percent (REff%)

(Wins + Saves + Holds) / (Wins + Losses + SaveOpps + Holds)

This is a measure of how often a reliever contributes positively to the outcome of a game. A record of consistent, positive impact on game outcomes breeds managerial confidence, and that confidence could pave the way to save opportunities. For those pitchers suddenly thrust into a closer’s role, this formula helps gauge their potential to succeed based on past successes in similar roles. BENCHMARK: Minimum of 80%.


A pitcher, typically a middle reliever, who accumulates an unusually high number of wins by preying on other pitchers’ misfortunes. More accurately, this is a pitcher typically brought into a game after a starting pitcher has put his team behind, and then pitches well enough and long enough to allow his offense to take the lead, thereby “vulturing” a win from the starter.

In-Season Analysis

Pure Quality Starts

We’ve always approached performance measures on an aggregate basis. Each individual event that our statistics chronicle gets dumped into a huge pool of data. We then use our formulas to try to sort and slice and manipulate the data into more usable information. 

Pure Quality Starts (PQS) take a different approach. It says that the smallest unit of measure should not be the “event” but instead be the “game.” Within that game, we can accumulate all the strikeouts, hits and walks, and evaluate that outing as a whole. After all, when a pitcher takes the mound, he is either “on” or “off” his game; he is either dominant or struggling, or somewhere in between. 

In PQS, we give a starting pitcher credit for exhibiting certain skills in each of his starts. Then by tracking his “PQS Score” over time, we can follow his progress. A starter earns one point for each of the following criteria:

1.    The pitcher must go a minimum of 6 innings. This measures stamina. If he goes less than 5 innings, he automatically gets a total PQS score of zero, no matter what other stats he produces. 

2.    He must allow no more than an equal number of hits to the number of innings pitched. This measures hit prevention. 

3.    His number of strikeouts must be no fewer than two less than his innings pitched. This measures dominance. 

4.    He must strike out at least twice as many batters as he walks. This measures command. 

5.    He must allow no more than one home run. This measures his ability to keep the ball in the park.  

A perfect PQS score is 5. Any pitcher who averages 3 or more over the course of the season is probably performing admirably. The nice thing about PQS is it allows you to approach each start as more than an all-or-nothing event. 

Note the absence of earned runs. No matter how many runs a pitcher allows, if he scores high on the PQS scale, he has hurled a good game in terms of his base skills. The number of runs allowed—a function of not only the pitcher’s ability but that of his bullpen and defense—will tend to even out over time.

It doesn’t matter if a few extra balls got through the infield, or the pitcher was given the hook in the fourth or sixth inning, or the bullpen was able to strand their inherited baserunners. When we look at performance in the aggregate, those events do matter, and will affect a pitcher’s peripherals and ERA. But with PQS, the minutia is less relevant than the overall performance. 

In the end, a dominating performance is a dominating performance, whether Clayton Kershaw is hurling a 4-hit shutout or giving up three runs while striking out 10 in 6 IP. And a disaster is still a disaster, whether Kyle Lohse gets a 5th inning hook after giving up 5 runs on 10 hits, or “takes one for the team” and gets shelled for 8 runs in 3.1 innings.

Skill versus consistency

Two pitchers have identical 4.50 ERAs and identical 3.0 PQS averages. Their PQS logs look like this: 

PITCHER A:     3    3    3    3    3 
PITCHER B:     5    0    5    0    5 

Which pitcher would you rather have on your team? The risk-averse manager would choose Pitcher A as he represents the perfectly known commodity. Many fantasy leaguers might opt for Pitcher B because his occasional dominating starts show that there is an upside. His Achilles Heel is inconsistency—he is unable to sustain that high level. Is there any hope for Pitcher B?

•    If a pitcher’s inconsistency is characterized by more poor starts than good starts, his upside is limited.

•    Pitchers with extreme inconsistency rarely get a full season of starts.

•    However, inconsistency is neither chronic nor fatal.

The outlook for Pitcher A is actually worse. Disaster avoidance might buy these pitchers more starts, but history shows that the lack of dominating outings is more telling of future potential. In short, consistent mediocrity is bad. 

PQS DOMination and DISaster rates (Gene McCaffrey)

DOM% is the percentage of a starting pitcher’s outings that rate as a PQS-4 or PQS-5.  DIS% is the percentage that rate as a PQS-0 or PQS-1. 

DOM/DIS percentages open up a new perspective, providing us with two separate scales of performance. In tandem, they measure consistency. 


A pitcher’s DOM/DIS split can be converted back to an equivalent ERA. By creating a grid of individual DOM% and DIS% levels, we can determine the average ERA at each cross point. The result is an ERA based purely on PQS. 

Quality/consistency score (QC)

(DOM% – (2 x DIS%)) x 2) 

Using PQS and DOM/DIS percentages, this score measures both the quality of performance as well as start-to-start consistency.  

PQS correlation with Quality Starts (Paul Petera) 

    PQS           QS%
     0             0%
     1             3%
     2            21%
     3            51%
     4            75%
     5            95%

Forward-looking PQS (John Burnson)

PQS says whether a pitcher performed ably in a past start—it doesn’t say anything about how he’ll do in the next start. We built a version of PQS that attempts to do that.  For each series of five starts for a pitcher, we looked at his average IP, K/9, HR/9, H/9, and K/BB, and then whether the pitcher won his next start. We catalogued the results by indicator and calculated the observed future winning percentage for each data point. 

This research suggested that a forward-looking version of PQS should have these criteria:

•    The pitcher must have lasted at least 6.2 innings.

•    He must have recorded at least IP – 1 strikeouts.

•    He must have allowed zero home runs.

•    He must have allowed no more hits than IP+2.

•    He must have had a Command (K/BB) of at least 2.5.

In-season ERA/xERA variance as a leading indicator 

(Matt Cederholm)

Pitchers with large first-half ERA/xERA variances will see regression towards their xERA in the second half, if they are allowed (and are able) to finish out the season. Starters have a stronger regression tendency than relievers, which we would expect to see given the larger sample size. In addition, there is substantial attrition among all types of pitchers, but those who are “unlucky” have a much higher rate. 

An important corollary: While a pitcher underperforming his xERA is very likely to rebound in the second half, such regression hinges on his  ability to hold onto his job long enough to see that regression come to fruition. Healthy veteran pitchers with an established role are more likely to experience the second half boost than a rookie starter trying to make his mark.

Pure Quality Relief (Patrick Davitt)

A system for evaluating reliever outings. The scoring :

1.    Two points for the first out, and one point for each subsequent out, to a maximum of four points.

2.    One point for having at least one strikeout for every four full outs (one K for 1-4 outs, two Ks for 5-8 outs, etc.).

3.    One point for zero baserunners, minus one point for each baserunner, though allowing the pitcher one unpenalized runner for each three full outs (one baserunner for 3-5 outs, two for 6-8 outs, three for nine outs)

4.    Minus one point for each earned run, though allowing one ER for 8– or 9-out appearances.

5.    An automatic PQR-0 for allowing a home run.

Avoiding relief disasters (Ed DeCaria)

Relief disasters (defined as ER>=3 and IP<=3), occur in 5%+ of all appearances. The chance of a disaster exceeds 13% in any 7-day period. To minimize the odds of a disaster, we created a model that produced the following list of factors, in order of influence: 

1.    Strength of opposing offense

2.    Park factor of home stadium

3.    BB/9 over latest 31 days (more walks is bad)

4.    Pitch count over previous 7 days (more pitches is bad)

5.    Latest 31 Days ERA>xERA (recent bad luck continues)

Daily league owners who can slot relievers by individual game should also pay attention to days of rest: pitching on less rest than one is accustomed to increases disaster risk.

Sample size reliability (Russell Carleton)  

At what point during the season do statistics become reliable indicators of skill? Measured in batters faced:

150:    K/PA, ground ball rate, line drive rate

200:    Fly ball rate, GB/FB

500:    K/BB

550:    BB/PA

Unlisted stats did not stabilize over a full season of play. (Note that 150 BF is roughly equivalent to six outings for a starting pitcher; 550 BF would be 22 starts, etc.)

Pitching streaks

It is possible to find predictive value in strings of DOMinating (PQS 4/5) or DISaster (PQS 0/1) starts:

Once a pitcher enters into a DOM streak of any length, the probability is that his next start is going to be better than average. The further a player is into a DOM streak, the higher the likelihood that the subsequent performance will be high quality. In fact, once a pitcher has posted six DOM starts in a row, there is greater than a 70% probability that the streak will continue. When it does end, there is less than a 10% probability that the streak-breaker is going to be a DISaster. 

Once a pitcher enters into a DIS streak of any length, the probability is that his next start is going to be below average, even if it breaks the streak. However, DIS streaks end quickly. Once a pitcher hits the skids, odds are low that he will post a good start in the short term, though the duration itself should be brief. 

5-game PQS predictability (Bill Macey)

5-Game avg PQS        Avg PQS       DOM%       DIS%
Less than 1            2.1          27%        40%
Between 1 and 2        2.4          32%        32%
Between 2 and 3        2.6          36%        26%
Between 3 and 4        3.0          47%        19%
4 or greater           3.5          61%        12%

Pitchers with higher PQS scores in their previous 5 starts tended to pitch better in their next start. But the relative parity of subsequent DOM and DIS starts for all but the hottest of streaks warn us not to put too much effort into predicting any given start. That more than a quarter of pitchers who had been awful over their previous 5 starts still put up a dominating start next shows that anything can happen in a single game.

High pitch counts and PQS (Patrick Davitt)

Starting pitcher matchups are vital for both daily fantasy owners and owners in longer formats that allow “streaming” of starters. In making SP decisions, owners might be tempted to sit a starter coming off a high-pitch-count (PC) start, even a good start, believing his next-game performance is bound to suffer from fatigue.

We studied starts from 2010-12 that had both high PCs (100+, 110+ and even 120+) and high scores in the Pure Quality Start (PQS) metric. The study showed such starters had good results in starts after high-PC starts: 

    1st Game Pitches               Next PQS Ave
    90- 99                            3.0
    100-109                           3.1
    110-119                           3.3
    120+                              3.6

And 120+ pitch-count starters were actually better than their peers after posting high PQS scores:

    1st      2nd     2nd
    PQS     All      120+
     3      3.0      3.6
     4      3.1      3.6
     5      3.3      3.7 
Thus, we can safely ignore the conventional wisdom that a high-PC game will make a pitcher “tired” or “worn out” and therefore less likely to be effective. The opposite is true—especially if the high-PC outing was also a strong PQS performance. It appears these workhorse starters and their teams know what they’re doing, and that they are highly likely to deliver a solid outing the next time out.

Days of rest as a leading indicator

Workload is only part of the equation. The other part is how often a pitcher is sent out to the mound. For instance, it’s possible that a hurler might see no erosion in skill after a 120+ pitch outing if he had enough rest between starts: 

PITCH COUNTS                       N E X T   S T A R T

Three days rest    Pct.       PQS       DOM       DIS      qERA 
< 100              72%       2.8       35%       17%       4.60
100-119            28%       2.3       44%       44%       5.21

Four Days rest     
< 100              52%       2.7       36%       27%       4.82
100-119            45%       2.9       42%       22%       4.56
120+                3%       3.0       42%       20%       4.44

Five Days rest
< 100              54%       2.7       38%       25%       4.79
100-119            43%       3.0       44%       19%       4.44
120+                3%       3.2       48%       14%       4.28

Six Days rest
< 100              58%       2.7       39%       30%       5.00
100-119            40%       2.8       40%       26%       4.82
120+                3%       1.8       20%       60%       7.98

20+ Days rest
< 100              85%       1.8       20%       46%       6.12
100-119            15%       2.3       33%       33%       5.08

Managers are reluctant to put a starter on the mound with any fewer than four days rest, and the results for those who pitched deeper into games shows why. Four days rest is the most common usage pattern and even appears to mitigate the drop-off at 120+ pitches. 

Perhaps most surprising is that an extra day of rest improves performance across the board and squeezes even more productivity out of the 120+ pitch outings. 

Performance begins to erode at six days (and continues at 7-20 days, though those are not displayed). The 20+ Days chart represents pitchers who were primarily injury rehabs and failed call-ups, and the length of the “days rest” was occasionally well over 100 days. This chart shows the result of their performance in their first start back. The good news is that the workload was limited for 85% of these returnees. The bad news is that these are not pitchers you want active. So for those who obsess over getting your DL returnees activated in time to catch every start, the better percentage play is to avoid that first outing. 

Post-DL Pitching Performance  (Bill Macey)

One question that fantasy baseball managers frequently struggle with is whether or not to start a pitcher when he first returns from the disabled list. A 2011 study compared each pitcher’s PQS score in their first post-DL start against his average PQS score for that year (limited to pitchers who had at least 15 starts during the year and whose first post-DL appearance was as a starter). The findings:

•    In general, exercise caution with immediate activations. Pitchers performed worse than their yearly average in the first post-DL start, with a high rate of PQS-DIS starts.

•    Avoid pitchers returning from the DL due to an arm injury, as they perform significantly worse than average.

•    If there are no better options available, feel comfortable activating pitchers who spent near the minimum amount of time on the DL and/or suffered a leg injury, as they typically perform at a level consistent with their yearly average.

April ERA as a leading indicator (Stephen Nickrand)

A starting pitcher’s April ERA can act as a leading indicator for how his ERA is likely to fare during the balance of the season. A study looked at extreme April ERA results to see what kind of in-season forecasting power they may have. From 2010-2012, 42 SP posted an ERA in April that was at least 2.00 ER better than their career ERA. The findings:

•    Pitchers who come out of the gates quickly have an excellent chance at finishing the season with an ERA much better than their career ERA.

•    While April ERA gems see their in-season ERA regresses towards their career ERA, their May-Sept ERA is still significantly better than their career ERA.

•    Those who stumble out of the gates have a strong chance at posting an ERA worse than their career average, but their in-season ERA improves towards their career ERA.

•    April ERA disasters tend to have a May-Sept ERA that closely resembles their career ERA.

Second-half ERA Reduction Drivers (Stephen Nickrand)

It’s easy to dismiss first-half-to-second-half improvement among starting pitchers as an unpredictable event. After all, the midpoint of the season is an arbitrary cutoff.  Performance swings occur throughout the season.

A study of SP who experienced significant 1H-2H ERA improvement from 2010-2012 examined what indicators drove second-half ERA improvement. Among the findings for those 79 SP with a > 1.00 ERA 1H-2H reduction:

•    97% saw their WHIP decrease, with an average decrease of 0.26

•    97% saw their strand (S%) rate improve, with an average increase of 9%

•    87% saw their BABIP (H%) improve, with an average reduction of 5%

•    75% saw their control (bb/9) rate improve, with an average reduction of 0.8

•    70% saw their HR/9 rate improve, with an average decrease of 0.5

•    68% saw their swinging strike (SwK%) rate improve, with an average increase of 1.4%

•    68% saw their BPV improve, with an average increase of 37

•    67% saw their HR per fly ball rate (hr/f) improve, with an average decrease of 4%

•    53% saw their ground ball (GB%) rate improve, with an average increase of 5%

•    52% saw their dominance (k/9) rate improve, with an average increase of 1.3

These findings highlight the power of H% and S% regression as it relates to ERA and WHIP improvement. In fact, H% and S% are more often correlated with ERA improvement than are improved skills. They also suggest that improved control has a bigger impact on ERA reduction than does increased strikeouts.

Pitcher Home/Road Splits (Stephen Nickrand)

One overlooked strategy in leagues that allow frequent transactions is to bench pitchers when they are on the road. Research reveals that several pitching stats and indicators are significantly and consistently worse on the road than at home.

Some home/road rules of thumb for SP:

•    If you want to gain significant ground in ERA and WHIP, keep all your average or worse SP benched on the road.

•    A pitcher’s win percentage drops by 15% when on the road, so don’t bank on road starts as a means to catch up in wins.

•    Control erodes by 10% on the road, so be especially careful with keeping wild SP in your active lineups when they are away from home.

•    NL pitchers at home produce significantly more strikeouts than their AL counterparts and vs. all pitchers on the road.

•    hr/9, groundball rate, hit rate, strand rate, and hr/f do not show significant home vs. road variances.

Other Diamonds

The Pitching Postulates

1.    Never sign a soft-tosser to a long-term contract.

2.    Right-brain dominance has a very long shelf life.

3.    A fly ball pitcher who gives up many HRs is expected. A GB pitcher who gives up many HRs is making mistakes.

4.    Never draft a contact fly ball pitcher who plays in a hitter’s park.

5.    Only bad teams ever have a need for an inning-eater.

6.    Never chase wins.

Dontrelle Willis List

Pitchers with BPIs so incredibly horrible that you have to wonder how they can possibly draw a major league paycheck year after year. 


Having the ability to post many saves despite sub-Mendoza BPIs and an ERA in the stratosphere. 

Vintage Eck Territory

A BPV greater than 200, a level achieved by Dennis Eckersley for four consecutive years.

ERA Benchmark

A half run of ERA over 200 innings comes out to just one earned run every four starts.

Gopheritis (also, Acute Gopheritis and Chronic Gopheritis)

The dreaded malady in which a pitcher is unable to keep the ball in the park. Pitchers with gopheritis have a FB rate of at least 40%. More severe cases have a FB% over 45%.

The Knuckleballers Rule

Knuckleballers don’t follow no stinkin’ rules.

Brad Lidge Lament

When a closer posts a 62% strand rate, he has nobody to blame but himself.

LOOGY (Lefty One Out GuY)

A left-handed reliever  whose job it is to get one out in important situations. 

Vin Mazzaro Vindication

Occasional nightmares (2.1 innings, 14 ER) are just a part of the game.


Any game in which a starting pitcher allows more runs than innings pitched.

The Five Saves Certainties

1. On every team, there will be save opportunities and someone will get them. At a bare minimum, there will be at least 30 saves to go around, and not unlikely more than 45.

2. Any pitcher could end up being the chief beneficiary. Bullpen management is a fickle endeavor.

3. Relief pitchers are often the ones that require the most time at the start of the season to find a groove. The weather is cold, the schedule is sparse and their usage is erratic.

4. Despite the talk about “bullpens by committee,” managers prefer a go-to guy. It makes their job easier.

5. As many as 50% of the saves in any year will come from pitchers who are unselected at the end of Draft Day.


A pitcher with a strikeout rate of 5.5 or less.

Soft-tosser land

The place where feebler arms leave their fortunes in the hands of the defense, variable hit and strand rates, and park dimensions. It’s a place where many live, but few survive.