The Fundamentals of Baseball Analysis
By Chris Liss
RotoWire Managing Editor
After writing a quick primer for the RotoWire baseball staff, I figured it would be worth posting the information for our readers as well. After all, the best owners don't just rely on the word of an expert or two, but look at the underlying facts and do the analysis themselves.
A. On Base Percentage and Slugging Percentage
The basic goal of a hitter is to get on base and move himself and others
around the bases. That might be painfully obvious, but it's a point that's
often lost when people look at metrics like batting average and RBI to
evaluate players. Batting average is just one component of on-base
percentage, and RBI is but one component of advancing oneself and others
around the bases. To get a more complete picture of how well a hitter is
achieving his basic goals, you should look at on base percentage (which
includes walks) and slugging percentage (which is calculated independently
of whether his teammates got on base ahead of him or not).
On-base and slugging do a better job of measuring whether a hitter is
helping his team score runs.
B. Batting average, Hit Rate, Contact Rate
A player's batting average is merely hits divided by at-bats, and it is not
necessarily a reliable predictor of near term future batting average. Why?
Because a lot of hits are due to luck. Bloop singles, ground balls that find
a hole. Also, a lot of outs are due to luck. Line drives right at people, a
sure base hit up the middle that becomes a double play because the shortstop
was moving to cover second on a hit and run. A hitter typically establishes
a career hit rate for him (e.g., 34 percent) but will deviate from that
number from year to year, often due to luck. If a career .300 hitter is
batting .260, and his contact rate is the same (i.e., he's not striking out
any more than usual), that means he's got a low hit rate. Batted balls are
not dropping in for him. Usually, you would expect a correction going
forward. So, instead of saying (as many were last year) that Victor
Martinez's poor first half was a "sophomore slump", you could look at his
very low hit rate and conclude that it was just bad luck. And that he would
be likely to turn it around.
A player's contact rate is an important predictor of batting average because
strikeouts are automatic zeroes, and any time a batter puts a ball in play,
he's got a chance to get a hit. The lower your contact rate, the harder it
is to maintain a good batting average because you need a higher hit rate on
batted balls to make up for your automatic outs when you swing and miss.
C. Plate Discipline
This is a batter's ratio of walks to strikeouts. If a batter has nearly as many walks as strikeouts, he can be said to have an excellent batting eye. Unless he's got very few of both, in which case he's an excellent contact hitter. Batting eye is important for two reasons: first, every walk a hitter takes more than offsets every automatic out (strikeout) he gives up, and second, because if he lets enough bad pitches go to draw frequent walks, it
is presumed that he's swinging at more hittable pitches more often than the player with poor plate discipline. This usually leads to better production. A ratio of better than 3:5 Walks to Ks is decent. 4:5 or better is very good plate discipline.
D. RBI and Runs
These are both team dependent stats for obvious reasons. In terms of scoring runs, getting on base and being a good base runner is the best a hitter can do. Whether or not someone
drives him in is largely luck. Though you can obviously expect more runs from a guy with good hitters behind him. RBI is the same thing - depends who's on base ahead of you. One roto note: Guys who draw a lot of walks like Barry Bonds, Todd Helton and Brian Giles help the team but cost themselves RBI. Hackers like Garrett Anderson or Juan Gonzalez in his prime are sometimes better RBI men.
E. Doubles and Triples
Triples are just doubles for fast guys. When young players have a lot of
either, they often develop home run power as they get stronger physically
F. Stolen Bases and Caught Stealing
In real baseball, you need to steal about three bases for every caught
stealing to break even (the addition of an out and the loss of a base runner
is far more significant than merely advancing one base). For that reason
sabermetrically inclined teams like the A's and Red Sox don't run a whole
lot. Also, if a guy has a low success rate, his manager should not be
sending him, and presumably, if he has half a brain, won't be. In fantasy
baseball, a stolen base is the rarest of the typical 5 x 5 hitting categories, and
thus steals are disproportionately valuable.
G. Real Baseball vs. Roto
On base and slugging percentage largely define the quality of a hitter in
real ball, but guys who walk too much lose HR and RBI in fantasy, and guys
who steal bases are disproportionately valuable in fantasy. We need to be
fluent in both languages - how a player's performance affects his status
with his team (that often controls how many at-bats he'll get) and how his
performance will affect a fantasy team. It's important to be aware of this
distinction and to understand the factors which contribute to both. Of
course, there is a lot of overlap.
H. Pitfalls to Avoid
Avoid the following types of pseudo analysis:
1. Hot and Not
Every player in the league gets hot at some point. And every one gets cold,
too. But it's impossible to know whether a hot streak will last three games
or six. Or when a slump will end. Essentially, that someone HAS BEEN on a
tear or HAS BEEN SLUMPING doesn't make it more or less likely that they will
streak or slump in the future. So all you have to go on is a hitter's
skills. The hitter with the better skills (power, speed, plate discipline,
contact rate, etc.) is the one to play regardless of recent performance. No
one can "time the market" for when slumps and streaks begin and end. For
every success you'll have playing that way, you'll have just as many
failures. Go with the best guys and ignore short term fluctuations. And
advise readers accordingly, i.e., don't start or sit players because they're
There are two exceptions: (1) When the streak coincides with the DEVELOPMENT
OF A SKILL (improved plate discipline, e.g.) or (2) When the streak CHANGES
A PLAYER'S ROLE on the team. But when Pujols, Abreu, Jeter or some other
fully formed, full time hitter has a streak or a slump, it usually doesn't
2. Vacuous Statements That You Hear in the Mainstream Media
"Sophomore Slump" - what does that mean? That means a player is having a
worse second year than his first year. Does it shed light on the cause? No,
it actually implies that the cause is some kind of curse that players have
after a good rookie year. That is not a cause. In Victor Martinez's case, the
cause was a low hit rate, i.e., he wasn't really hitting any worse, just
"He's a first half player" - Unless a guy has been in the league five years
or more and had consistent first half/second half splits every single year
(and moreover there's some attributable cause), avoid this kind of
non-analysis. Out of the 750 major leaguers, 25-50 or so will randomly show
first half/second half patterns for a few years. Don't make something of it
unless it persists for a LONG time, or unless there's some cause (e.g., a
catcher in Texas getting worn down by the heat every year).
3. Law of Averages/Regression to the Mean
There's a proper and an improper way to apply this:
The improper way is to assume that a career .300 hitter who hit .250 in the
first half will somehow hit .350 in the second half to make up for it.
The proper way is to assume, that a career .300 hitter who hit .250 in the
first half will hit .300 in the second half and hit .275 for the season.
A hitter can exceed his averages in the second half, and he can continue to
fall below them. But you should assume regression to the mean, all things
A. Four Possibilities in an At-Bat
3. Home Run
4. Ball in Play
This is very important to keep in mind because studies have shown that for
the most part BALLS IN PLAY AFFECT EVERY PITCHER THE SAME WAY. Essentially,
there is a 30 percent chance that a ball in play off of anyone (whether it's
Aaron Sele or Pedro Martinez) falls in for a hit.
So the difference between one pitcher and another depends SOLELY on how good
they are at preventing home runs and walks and GETTING STRIKEOUTS.
B. Strikeout Rate
The reason why strikeout rate is so important is simple: Balls in play fall
in for hits 30 percent of the time. But strikeouts are outs 100 percent of the
time. So if you strike out 10 guys per nine innings, you only need 17 more
batted ball outs. If you strikeout 5 guys per nine innings, you need 22 more
batted ball outs.
But for every 10 balls in play, three fall in for hits, and seven are outs.
So to get 22 ball-in-play outs, you should expect to give up 9.4 hits (3/7
*22). But to get 17 ball-in-play outs, you should only expect to give up 7.3
hits (3/7 *17). That means that the elite strikeout pitchers give up two
less hits per game. Now add walks and home runs allowed into the mix, and
you can see that an extra walk and a half, an extra home run and two more
hits per nine innings make the difference between giving up 2.8 runs per
nine innings and 5.8, i.e., the difference between an elite pitcher and a
So the key to doing pitcher analysis is to note (1) Strikeout rate, (2)
Strikeout to walk ratio and three (3) Home run rate.
C. Strikeout to Walk Ratio.
This measures a pitcher's command. The strikeouts show that a pitcher's
pitches are moving and/or that he's changing speeds effectively, and walks
partly show how well a pitcher is locating. A K/BB ratio of 2/1 is the bare
minimum for a good pitcher. 2.5/1 or more is preferable, and 3/1 or more is
D. Home Run Rate
Not allowing home runs is a skill just like not walking batters is a skill.
If a pitcher is wild, he'll miss off the plate and walk people, but he might
just as easily miss over the plate and give up home runs. The top pitchers
in the league keep the ball in the park.
E. Wins and ERA
There are four components to predicting pitcher wins. I list them in order
(1) Run Support - completely team dependent, and fluctuates wildly from
pitcher to pitcher and game to game even on a given team.
(2) Pitcher Quality - Strikeout rate, K/BB ratio and Home run rate
(3) Bullpen - Do the guys come in and blow the win?
(4) Team defense - not just errors, but defensive range.
A pitcher is NOT a winner per se. If he pitches well game after game, he's
got a leg up, but he needs the other three components to work in his favor.
When doing analysis, do NOT emphasize a pitcher's won-loss record as
evidence that he's good or bad, because there are much better measures of
pitcher performance that are not so luck and team dependent than wins.
ERA is more on the pitcher, but remember that 30 percent hit rate on all
balls in play? Sometimes, a pitcher will go through a stretch where 40
percent of batted balls drop in. That will affect his ERA significantly
through no fault of his own. He may also go through a phase where only 18
percent of batted balls drop in. That, too, will affect ERA. In other words,
ERA has some luck built into it and as such is not as good a predictor of
future ERA as a combination of the big three factors (K rate, K/BB ratio and
When doing updates, team previews, articles or other analysis, do not use
wins and ERA as your primary evidence of pitcher performance.
Saves are mostly about opportunity, as it's not that hard to protect a two
or three run lead for one innning. Even a one run lead is safe with most
pitchers, since no full time hurler has an ERA of 9.00. But for the reasons
above, closers with big strikeout rates are preferred, as well as low walk
and home run rates. Occasionally, there are players like LaTroy Hawkins who
pitch very well as setup guys, but don't have the mentality to close, but
all things being equal, look at the skill set first and make judgments about
a pitcher's mentality second.
A. Park Effects
Pitchers fare much better in places like Shea, Dodgers Stadium and Petco
Park. Hitters in Coors, Arlington and Bank One Ballpark. Check out Jason and
Darin Brown's article on park effects to get familar with the different parks. Also, some parks
are tough on particular players - e.g., Yankee stadium is traditionally
tough on right-handed power hitters (though apparently not so much last
B. Minor Leaguers
Stats and skills are important, and should be analyzed in the ways suggested
above, but age also plays a big part. If a player is tearing up Double-A at
age 24, it's not nearly as big a deal as if he does it at 21. Minor leagues
have park and league effects (some leagues are hitter friendly, and others
are pitcher friendly), so modify your assessments accordingly. Also,
remember that minor league doubles (especially for 18-20 year olds) often
translate into home runs as those players mature.
C. Peak Ages
Hitters typically peak around age 27, though they often have power spikes in
their early 30s. Pitchers tend to peak in their early 30s.
See Will Carroll's article for some injuries that tend to linger:
Of course, I'm probably leaving plenty of important stuff out, but hopefully
this is a good starting point and can get everyone on the same page when
doing baseball analysis.
From a Subsequent Email from Jeff Erickson
When looking at park effects, a player's groundball/flyball ratio (both
as hitters and as pitchers) can serve as a decent indicator for other
metrics, most notably HR rate. This is especially true in the more extreme
parks. In other words, a pitcher with a high flyball rate getting signed by
the Reds (say, for example, Eric Milton) is a pretty obvious indicator of
more homers allowed to come.
When it's appropriate, we welcome the use of "scouty" analysis. It's
appropriate to talk about a pitcher's repertoire, for instance,
particularly when he gets called up to the majors. It can even serve as a
counterweight to otherwise deceptive stat lines (think Rich Hill, e.g.).
Just make sure not to get overly caught up with it, especially if it flies
in the face of other criteria, mentioned above.
Article first appeared 02/15/06