Sure, batting average has some value, as long as its limitations are also recognized. The year-to-year fluctuation, the element of chance that is inherent (bloopers that fall in one year but not the next, for example) . . . and it should always be placed in the context of OBP and SLG. A guy with a .250 batting average who also has a .300 OBP and a .350 SLG is useless (think Tony Womack). A guy with a .250 batting average who also has a .395 OBP and a .590 SLG is very valuable (think Adam Dunn).

Its simplicity of calculation is probably why it endures. I'd guess many of us first learned of division because of batting average.

Dear BG

As a fan I use batting avg as a way to gage odds.

Let's say its a tight game and player x is at bat and he is batting .290 for the season and is 1-4 so far that night, the odds, based on his avg is that he should get a hit in the next ab. I doesn't always work of course but it helps me atleast get a feel for what could be next. I also apply that thinking when we are pitching to size up if its an easy out or not. I guess I use the b.a. as a way to find who is due :)

Law of averaging yours..
mike

My own theory is that statheads don't like it because it's too easy to calculate and everyone can do it once you've mastered long division.

I don't agree that it's worthless. It's remained the one way to judge a hitter's relative effectiveness at just a glance. That's fine for most fans but it's obviously not enough to go by if you're a manager filling out a lineup card or making strategic decisions during a game or a general manager deciding things like trades and other personnel decisions.

It's just one of many important figures in those situations and probably not the most important. I admit that through all my years of watching, coaching and playing baseball, I never really considered just how fine a line it was between being a .300 hitting "star" and mediocre player until I saw Bull Durham for the first time and had Crash Davis explain it to me. One ground ball with eyes or flare dropping in every week or two can be the difference between Cooperstown and a career in the minors.

Dear Mr. Crikket,

Long Division?

Curiously,
Batgirl

This is an uneducated guess, but my feeling is that BA persists as a popular measuring stick because of the legacy of former players in the managerial and broadcasting roles.

A significant percentage (there's the math again, damn!) of coaches and broadcasters used to play the game professionally. And while not having any experience of being inside a clubhouse, I would surmise that the focus is largely on how many hits a player has in a particular game, series, week etc. Among people who play the game for a living, number of hits is the definition of success and not who has the most walks, bases per time at-bat, etc.

There is a definite disconnect between the ability to play the game, and the ability to rationally analyze it and decide what's important as a skill set. Not to say the two are mutually exclusive by any means, (see Beane, Billy) but it definitely ties into the Moneyball argument between scouting vs. stats. People who favor BA tend to be scouts who've been around the game forever and know what's important, batting average dadgummit! Whereas fans who don't play, but want to know the whys and hows of to best build a team tend to favor skills more statistically correlative with success.

BA is great in context but, yes, overused in valuation.

The large sample size of AB's makes it an attractive stat. It's humorous every year to see the hype of people chasing .400 only to have the law of averages catches up. And it's telltale when players with studly BA's are shut out in the limited sample size of postseason AB's.

The best use of BA is when looking at a player's entire career and comparing him to himself.

Sorry, one more comment on my comment. I didn't mean to imply that everyone who doesn't play the game loves stats, or that one approach is better to frame a personal reference to the game. I just happen to prefer the statistical approach in player evaluation.

Don't get me wrong, I love the sass. Love it!

Mike:

Noooo!!! If you're a .290 hitter and you're 1-4 for the day, you're not due, you've got a 29% chance of a hit in the next AB, same as the last (assuming BA is probability of getting a hit, not factoring in conditions and pitching).

Being "Due," and a "Law of Averages" like that is a misunderstanding of statistics. Small sample sizes will not revert to the mean, they will only revert over a season (and that's assuming constant conditions).

But if you do believe in such voodoo, say your batter is a .290 hitter and he's 1-4 for the day. In the fifth AB, he gets a hit, his BA for the day is .400, while if he makes an out, his AB is .200, which is closer to .290. If he gets a BB, HBP, or SAC, he's still at .250, so that would be most likely in that scenario. But unless you're Barry Bonds or Corey Koskie, getting hit or passed is probably less likely than a hit or an out.

Oh, and to the question, I'd guess BA is popular because it's easy to judge. People have a good feel for it, like the Fahrenheit temperature scale.

Compare:
BA:
Under .200 = very bad (unless pitcher)
.200-.250 = bad
.250-.275 = okay
.275-.300 = good
.300-.350 = great
.350-.399 = MVP
Over .400 = Ted Williams

Deg Fahrenheit"
Under 0 = Very Cold
0-30 = Cold
30-50 = Cool
50-60 = Lukewarm
60-75 = Warm
75-90 = Hot
90-100 = Very Hot
Over 100 = Death Valley

Using Celsius or OPS would make more sense, but not enough people have an ingrained feeling for it.

I think pitching peripherals (WHIP, K/9) may be easier to pick up, because ERA is a less-ingrained stat. (AL/NL differences makes it less straightforward in a way BA doesn't).

It's useful mostly as a component of OBP and (to a lesser degree) slugging percentage (neither of which, by the way, is significantly harder to calculate than a player's batting average). For example, if all you have is a stat sheet showing that Player A and Player B were both shortstops with a .360 OBP and a .430 SLG on the season, you'd probably think that they were equally likely to continue at that pace. However, upon discovering that one is batting .260 and the other .330, you have a much clearer picture of the shape of their performance and a better idea of which player is likely to be better going forward (ie, the ".330 hitter" is likely to see his average drop and his other rates along with it; visa versa with your ".260 hitter").

OBP is equally easy to judge; you just have to be raised with it. I read a little table with the league leaders every morning in USA Today; now it's instinctual, much the way a grasp of the metric system would be (if I had one):

Under .300 = very bad
.300-.320 = bad
etc. . . .
.450-.499 = MVP
500+ = Barry Bonds

The idea that there's something inherently geeky -- and by inference bad, since using your brain is definitely not valued in this country -- about wanting to use a more accurate measure of value to think about a player's performance borders on the comical.

See I always thaought that people that used ob% or slg% were doing so to defend players with low batting averages. Maybe I'm just not a moneyball guy but I still look at batting average first, hrs second, rbis third, and then depending on the player i will look at things like stolen bases or sacs.

Until the triple crown is replaced with obp, slug, & runs created, I'm going to let the sabr worry about those and I will stick with the big three.

BA persists beause it's what people are used to. It's on the bubble gum cards. It's what we used as kids to evaluate players. Players win batting titles not based on OBP or OPS but on BA. To paraphrase Wittgenstein, it's our established practice, an integral part of our form of life.

The fact that it's inadequate in separating luck from skill, or even in defining success in terms of a player's impact on a season, has made a negligible dent in the practice. We still see lists of player stats in the papers that neglect OBP and SLG. BA, HR and RBI are still the prevalent published stats. Perhaps most people don't care whether a player's success was a matter of luck or skill?

It will take a radical shift in practice before that changes. The evolution is underway, but it will take years before the practice of listing BA next to OBP and OPS is established in the mainstream.

mike is succumbing to the gambler's fallacy. Regardless of previous at bats, a batter batting .290 has a 29% chance of getting a hit in any at bat and the odds are always against him.

Also, if the other team has a batter who's 0-fer-whatever, he's in a slump. It's only when your player is 0-fer-whatever that he's due. This is known as the First Law of Subjective Analysis and is supported by centuries of mathelogical history.

I'm a believer in the stat one simple reason. A hit is a hit, some aren't pretty, but they are still hits and a good hitter is going to get some hits that help his average but not his slugging percentage. Is there a stat that compares how many runs a player scores following a line drive vs. bloop singles, more often than not a good hitter is fooled but still gets the bat on the ball and gets a hit . . . and more importantly gets on base? I've seen games won when a good hitter was fooled but still managed to get the bat on the ball and find a hole thus driving in the winning run.

I used a .290 hitter just as example, law of averages tell me a .290 hitter is most like to get 3 hits per 10 at bats. When you break it down to a game by game stitiution if a good hitter has one hit in four at bats I say his odds of getting a hit in his next ab are good. Not a perfect science by any means.

Like wise if a .240 hitter is 2-2 one night i'm saying that law of averages tell me he is not getting a hit in his next ab. It's not perfect its playing the odds.

Just try it some time. Also there are times were even those theories are thrown out, like if Jeter is up in the 9th in a big game it's a hit, if it's A-Rod it's an out.

Batting average is useful to get a first-glance general picture. It tells you how likely a guy is to get a hit--it doesn't tell you how he got out, or how many times he walks/sacrifices, or how many extra-base hits he got. It just simply says: when a player comes to the plate, this is how likely he is to hit the ball. It's not a horrible way to keep track of how a guy is doing offensively. (It's pathetic if you're offensive style is "all power", because it's missing the power hitting stats. It's not bad if you're trying to play small ball--it tells you the guys who get hits and thus are able to advance runners/get on base (sure, on-base % would be better, but batting average isn't bad).

But batting average doesn't tell the whole story; like every other stat in baseball, you need to see them all to get an accurate picture. Batgirl's "Less Stats, More Sass" philosophy likely doesn't want to get into that. There are way too many stats--especially when you take into account "versus left-handed pitcher/versus right-handed pitcher".

If only given one stat, I wouldn't mind batting average. It tells part of the story.

BG: yes, long division. I know I'm dating myself but calculating batting averages was my first real life application of what my elementary school teacher in Albert Lea tought me. I realize the use of the term dates me but then I've admitted before that I'm probably too old to be spending time in discussions with you youngsters.

CJoseph: I agree with you. In fact, when I began coaching youth baseball many years ago, I calculated two hitting percentages for my players, and only two. BA, because the kids knew what that was and kept asking for it, and OBP, because that's what I wanted to know in terms of how productive each kid really was.

Dear Mr. Mathewson,

I'm curious about the luck comment--can you talk a little more about it? It seems every statistic in baseball can have some luck behind it. And if .OBP is just batting average and walks, how does it make it less prone to luck?

Curiously,
Batgirl

My own personal pet peeve (besides people writing checks for purchases of $1.18) is how the Strib uses batting average as a ranking tool for team offense. Don't they understand that the most important statistic for a team's offense is runs scored?

It doesn't work that way. Batting averages take months, and hundreds of at bats, to normalize. A guy who got a hit in his last at bat is no more or less likely to get one in his next at bat. A guy who gets hits 29% of the time will have several arbitrary stretches of 10 at-bats in which he'll get no hits, and several where he gets six. The "Law of Averages", as you're applying it here, doesn't exist. If it did, it would require a lot more math than anything your run-of-the-mill stat-minded fan does.

The idea that an entire community of diverse persons, most of whom have never met, are using OBP in some kind of quest to "defend" players with low batting averages (why? because we don't like hits?), is pretty silly. There's no conspiracy of dunces, believe me.

Mike,

I don't think that is the proper use of the law of averages. The law of averages (law of large numbers, actually), says that a ".290 hitter" will get hits in approximately 3 out of every 10 at-bats. That will reflect streaks, slumps, bloops, line-drive outs, and everything else. Being 1-4 or 0-5 in a game does not mean they are statistically due. Its like in flipping a coin. If you flip 10 straight heads, the law of large numbers doesn't say that the next flip will be a tails or a heads. The probability is still 50%. The law of large numbers says that if you flip the coin 1000 more times, the number of heads will be pretty close to 500. If a .250 hitter is 3-3 the law of averages doesn't tell you he will get out, the fact that he is a .250 hitter tells you that. Same for a .300 hitter or a .400 hitter; they will still perform the same regardless of the past. Although if you believe in streaks and slumps (which most people do) you would actually think his odds of getting a hit might be higher than .25 since he is on a roll, which would be different than a fair coin flip.

BatGirl,

If I may respond for Mr. Mathewson, what I think he means is that batting average is subject to much wider fluctuations from year to year. As Crash Davis explained, an extra bloop here or there can change your batting average pretty significantly in both small samples and in longer periods of time.

A lot of the statistical research shows that there isn't a lot of luck, ie wide fluctuation, in things like on-base percentage, slugging and isolated power (which is slugging minus batting average). The stats with less fluctuation from year to year and level to level have much greater predictive power, and you can be much more certain about their skills translating to the major league level. Which is why every GM worth his salt has called TR this offseason to ask about trading Dr. Morneau.

It's a dumb point, but one worth mentioning: I think baseball as a whole enjoys the argument that 'a great baseball player only succeeds 1/3 of the time' as a measure of how difficult the sport is at the pro level. Everyone's heard that cheese analogy from someone over the years, and it's completely based on batting average.

It's antiquated in this era of Sabrmetrics, but it sounds more impressive than '42% of the time'.

To directly answer the questions, I think, yes, BA is of use to value players. However, the level of value put on it is too high. A .250 hitter is not as good as a .300 hitter at getting hits. That is all. Just like a 25 HR hitter isn't as good as a 40 HR hitter. However, to properly view a player, you need to look at stats in tandem with each other. Is a .250 hitter with 40 HR better than a .300 hitter with 25 HR. What if the .250 hitter walks 100 times and the .300 hitter only walks 10 times. What if one steals lots of bases, plays good defense, and rescues cats from trees? The use of new-fangled stats attempts to quantify more of these variables. Obviously, they are more subjective based on relative weights and values (doubles 2x singles, walk = hit??) but they incorporate more of the overall aspects of a player's game. Historically, it was virtually impossible to perform these calculations, with the limited data and calculation availabilities. So the early stats, such as BA caught on. And as baseball is both a very historical game (in terms of historical respect) and numbers-based game (based on the individuality of many of the outcomes), this has not changed. Baseball, much more than any other sport and society in general, is opposed to change. You see this in DH, Intraleague play, the Wild-Card, and Instant Replay. Eventually, changes are made to incorporate them into the fabric of the game, but the transition is slow and painful. In 20 years, other stats will probably be a greater part of the game, but the change should not be expected to occur quickly.

It may not be the true law of averages,and I'm not saying it's 100% correct all I'm saying is when I am at a game and I look at things like this, playing the odds, I tend to be right more times then not.

Odds and averages tell me a good that has a low average that already has two hits is more likely to not get one his next at bat, and I will agree that there are more things to take in to account,but Occam's razor also tells me that the simpliest answer is the best answer.

Of course you are right more times than not, if a .250 hitter is 3-3. I bet you are right about 75% of the time. But you would be just as right if a .250 hitter was 0-3. That is the point we are making. And, as I previously stated, you are actually probably right somewhat less than 75% of the time given streaks and slumps.

Re: The application of Occam's Razor,

It's not really the simplest method, however. Depending on how you want to evaluate a player's likelihood to get a hit in his next at bat (by how many he had in X number of previous at bats, or how well he's made contact on X number of previous pitches, or what have you), the weight of previous performance coming to bear on future performance is in constant flux. It's actually a very complicated way of doing the thing. It also gets more complex when you think about that an at-bat is not a truly binary event, though it's sometimes useful to think of them that way. How do you take into account walks? Hit-by-pitches? It gets hairy in there. And you're lost if a guy comes up who's batting .330 and he's 1 for 3 already.

It seems to me that the simplest way to think about it is just to use the rates a batter has over a longer period, and not try to factor in immediate recent performance unless you know something (injury, personal distress) is influencing it.

Mike,

Not to sound like I'm antagonizing you, but isn't it possible your observations are reinforced when outcomes are positive, and discarded when they're not?

Respectfully,
cubsfan

But Occam's Razor tells me to forget about walks, hit by pitch, sacflys, fielder choice, safe on an error, or AJ bluffing the umpires into making an out a hit. Occam tells me to base it off of he either gets a hit or he makes an out. Find the easiest solution to the problem.

Mike,

I guess we just differ as to what the easiest solution to the problem is.

Mr Cubbiefan

I don't in anyway keep track, and I know for fact that I am wrong alot, but I also feel as though most times i get it right. It's not some super power or anything that a brag about it's just me trying to play off the odds. try it sometime when you are at a game. It could be just dumb luck but I try to use what has already happened to get a feel for the future.

It most amazes my young nephew.

Thank you, everyone, for going after the "due" fallacy! In the unlikely event that you flip a coin 99 times and each time it comes up heads, the chance of it coming up tails the 100th time is STILL only 50%. That will never, ever change, no matter how many times you flip the coin. Same's true for batting average.

But most people do think like that. Which is how casinos make money.

And back on topic: Can't I just like all the numbers? I like batting average (and OBP, for that matter) because I can quickly and easily explain what it means. # of times out of 100 that a player will get a hit (get on base). These are measurements. OPS, while a useful number and one I'll look at, isn't a measurement of an outcome. The number doesn't correspond to some probability or frequency that can be explained. Hence, it's a statistic.

OPS is a moderately useful junk stat. And OBP is equally easy to explain:

OBP: Number of times out of one hundred that a player gets on base through any means.

Slugging is slightly more complicated:

SLG: Number of bases out of one hundred a player collects on hits.

Both expressed as decimals. Fairly simple.

I'm rethinking the slugging definition. I screwed it up.

Dear Mr. Joseph,

Does that include errors?

Curiously,
Batgirl

When I use the term "Law of Averages," I refer to the batter's OWN average.

And applying probability statistics to BA is a "misunderstanding of statistics."

I would hit exactly 0 of 100 Major League fastballs. Not 50 out of 100.

BG,
OBP doesn't include errors. It's walks, hits and hit by pitches.

And Twink's right about OPS not being a measurement, but rather a statistic. OBP and SLG are both measurements, but once added together, while useful in predictive power, they don't measure anything. It's two measurements added together. Still pretty useful though.

So, in summary, OBP is easy to explain, and I'm a moron.

An error is counted as an "out" in OBP, just as in batting average. I guess I should have clarified that.

Why is applying probability to a batter's average a misunderstanding of statistics? Isn't it exactly what it is calculating? If a batter in a large amount of at-bats has been historically a .300 hitter, you could accurately say he has a 30% chance of getting a hit. Of course, walks and HBP would be left out (using OBP would include them). Of course, this would imply that his past at-bats are no different than future at-bats (rookie = veteran = OMM). I don't see where the fallacy is here, but I may just be missing it.

Isn't slugging = total bases/at bats?
So a player with a slugging of .500 means that on average they get a half a base per at bat, or a single per 2 at bats (or HR per 8, etc.)

So as a number it would be the number of bases per 1000 at bats.

With all due respect, this is soundling like a BASEBALL BLOG!

OBP how can I explain it?
I'll take you frame by frame it
To have y'all jumpin' shall we be singin' it
Just get on base and then be wingin' it.

I'm down with OBP (Yeah, you know me.)

Ron, you are Naughty.

Almost all of these statistics are valuable, they just have varying degrees of detail to accomodate different uses. It doesn't have to be an either/or proposition.

A casual fan doesn't need something as detailed as OPS to determine whether they like a player or not. BA is simple enough and accurate enough for the casual fan to both understand and to make a reasonable determination as to the ability of a given player to hit the ball.

A manager (or sabermatrician) needs (or just craves) more information, so additional statistics are created that incorporate more forms of relevant data.

- Freez
P.S. - Mike, we should play Poker. Tell you what, you bring lots of cash, and I'll bring something to carry it home with. =)

Oh my god, we've been over run by actuaries... 'tis the end of the world.

I guess, for me personally, whether we're talking major leagues or T-Ball, the thing that makes the game interesting is the human factor more so than the statistical or predictive nature of the numbers.

With a .250 hitter coming up in the 9th with the tying run on 2nd, I don't put much stock in whether he's oh-fer and thus "due" or whether he's already got a hit today and therefore unlikely to get another. I think back to consider if he's had good at-bats.. making contact.. or lunging at everything he's thrown. Is he "seeing the ball well"? Were I privy to the information, I'd consider things like whether he's had a fight with his wife/girlfriend earlier today, found out his accountant has been embezzling from him or whether he has been sleeping soundly to be of greater interest and importance than his batting average.

If it were all merely about statistics and probabilities, we could just play simulated games forever. But the players are humans with widely varying personalities and flaws. Thank God.

OBP includes sac flies (but not bunts or grounders) as plate appearances not resulting in on-base, while BA excludes all plate appearances of which the result is a sacrifice. Therefore, you can have a lower OBP than BA. This happened to Castro for the month of August, i think, where he had no BB or HBP but one Sac Fly. His month ended something like this:
BA: .249
OBP: .246

SLG is Total Bases / At-Bats. Total Bases is the first stat I go to in the box for batters (for Pitchers, it's IP).

Applying pure laws of probability to baseball statistics isn't always helpful, because probability is about the exact same event being repeated while their are countless variations for each plate appearance. The biggest I can think of is the pitcher faced: Is it Jose Lima or Andrew Sisco or Roger Clemens? What are the weather conditions? What point in the game, and are there runners on? Did the batter skip breakfast? Is he a switch hitter? Does he notice holes in the defense that he might try to take advantage of? Are there 25 different people telling him how to fix his swing? How does the pitcher approach him? Is the ump squeezing the pitcher or giving him first base as part of the strike zone?

I think of OPS as Non-outs + Bases per AB. A Single is 2, a walk is 1, a triple is 4. That's not exactly it, but close. With SLG and OPS, you can't do mike's (mathematically incorrect but still fun) prediction game and say "I bet he gets 1.5 Bases & non-outs!"

Jim Crikket:
I am an actuary (who should be working), and you are exactly right.

Dear Mr. Davis,

It is not my fault if no one read the book.

Sincerely,
Batgirl

I, personally, think that batting average can still be important, simply because a poor batter will never draw any walks. A hitter has to have something for the pitcher to be scared of. Some people think slugging takes care of that, but power is too expensive in this league as it is.

Batgirl,

You're absolutely right. There's nearly as much variation between batting averages as there is between on base percentages. Take Derek Lee as an example--this year, he got on base at a .418 clip, compared to his career average of .363. Did he suddenly learn to draw walks once he turned 29? Of course not. His AVG increased by nearly the same amount (from .276 career to .335)--accounting for the increase in OBP.

However, the ability to draw a walk is still more reliable than the ability to get a hit, because there are more variables to take into account once a ball is hit into play--how hard it's hit, where it's hit, the quality of the field, the speed of the runner, the mobility of the fielders, the arm of the person fielding the ball, etc.--whereas the only real variable in trying to draw a walk is the strike zone of the umpire.

So how can we tell how good a player is at drawing walks is OBP is all tied up in AVG? BB/PA is a pretty good measure. Let's look at Derek Lee again. For his career, 10.4% of his plate appearances have resulted in unintentional walks. More recently, his BB/PA was at 9.6% in 2004. This season, his BB/PA was 9.5%. That's pretty darn consistent, and various studies have shown that Derek Lee is not an abberation--that there's not much variation in how much players walk from year to year.

Back onto the issue of luck, some staticians subscribe to the Three True Outcomes theory, which states that the three outcomes that batters truly control are home runs, walks and strikeouts. Everything else is subject to luck. From this theory comes (of course!) another stat--BABIP, or batting average on balls in play (Hits / AB - K - HR). BABIP is subject to wild variation from year to year. However, if you have a couple years worth of BABIP, generally a pattern emerges. Looking at any one particular season in the context of this larger pattern can help explain if that year is an abberation or part of a trend.

Let's move on to Justin Morneau, coming off a miserable 2005, where he hit .238 with only 22 HR. His BABIP in 2005 was .254--a pretty significant drop from the .275 he posted in 2004. Even more notable is how BOTH years are staggering drop-offs from his minor league career average of .346.

While transitioning from AAA to MLB almost always results in a drop-off, Morneau's is extreme, and likely at least in part due to bad luck, making him a good candidate to rebound next year.

Dear BG

Book ? What Book ?

Unprepairedily yours
mike

BA's a good way to determine a player's value if you don't mind having a very incomplete picture to look at.

How many at-bats does that .333 hitter have? He could be 1-for-3 or he could be 100-for-300.

How many years has he been playing? This year's .250 rookie might be next year's .300 hitter.

How many of those hits meant something to the team? Some guys can hit like mad with an empty pond but totally choke with RISP.

Ichiro hit .372 in 2004 but the Mariners went 63-99. Ok, that's a poor example because it has very little to do with Ichiro's value and a whole lot to do with the fact that the Mariners sucked in '04. I just think it's a pretty sick stat ;-)

Then again, only five of the nine players listed by baseball-reference.com as the 2005 Twins' regulars posted an average above .250, whereas ALL NINE of the 2005 White Bitches hit over .250, so it's not a totally useless stat. Yeah, I did just call Carl Everett a White Bitch.

I guess the point is that looking at someone's BA alone tells you about as much about their value as a player as looking at their appearance alone tells you about their value as a human being.

ps- Dear Mr. Everett: For what it's worth, I cheered very loudly when you broke up Mike Mussina's no-no bid.

Hey, there's a thought- mike, your method goes to pot when you're watching someone toss a no-hitter, eh?

Wether you use average or obp ted williams is the greatest hitter of all time and we shoudl just him as the standard

Well, I think that OBP is better than BA, but it takes some time to get used to. I think that some players contribute more by walking than others, and that OBP takes that into consideration, so it is better. A walk can walk in a run, move someone on first to second, or get someone on base to allow them to score later. I think that just focusing on hits or homeruns isn't telling the whole story. also, Slugging confuses me, so i don't use it. simple enough...

"While transitioning from AAA to MLB almost always results in a drop-off, Morneau's is extreme, and likely at least in part due to bad luck, making him a good candidate to rebound next year."

From your keyboard to God's ears, Santans Second Cy.

(I say that at the risk of adding "Devine Intervention" in to the statistical formulas which should just about make all the sabremetricians' heads explode.)

Dear BG

The SABR perspective, as far as I understand it, is to evaluate players on ability rather than luck. For example, ERA is supposed to depend more on luck than other metrics because whether or not balls have eyes or bloops fall in is beyond the control of a pitcher. The other metrics, which focus primarily on HRs allowed, BBs, and SOs, measure what a pitcher does have control over.

Similarly BA is supposed to depend more on luck than OBP because a batter can control whether he swings at ball 4 of not. But he can't control whether he hits a lot of line drives right at people or whether his balls have eyes. (Uh, that sounds funny, but I'll leave it.) Because walks are as close to an absolute measure of a hitters batting eye, OBP is closer to an absolute measure of a hitters skill than BA and less a measure of his luck.

I hope that explains it.

Sincerely,

cmathewson

Dearest BG,

I was just in an old-school frame of mind.

Naughtily by Nature, RD

p.s. I'm working on a rap for the 50 Cent sequel. You know, "Sit Down Bitch, or Die Trying."

"the ability to draw a walk is still more reliable than the ability to get a hit"

So theoretically, nobody ever swings at a 3-2 pitch that would have been ball 4.

"...the only real variable in trying to draw a walk is the strike zone of the umpire."

I don't agree with this, either... how many outs are there, how many bases are open, who's at the plate, who's coming up behind the batter at the plate, does the pitcher have his stuff, has the pitcher's arm turned to jell-o, is the batter working the count or just hacking at anything near the strike zone... you can go on and on.

As for Three True Outcomes, I can only assume that you're being satirical, as TTO is described by Baseball Prospectus as "a long-time inside joke at rec.sport.baseball".

In fact, here's BR's glossary entry:

"The Three True Outcomes are a walk, strikeout or home run. They are called this because the three supposedly are the only events that do not involve the defensive team (other than the pitcher). This is clearly untrue, as outfielders can take away potential homers and inside-the-park home runs are influenced by the defense. Additionally, if a defensive player fails to catch a foul pop-up, the batter can wind up with a defense-influenced walk or strikeout (as the at-bat could have been ended earlier)."
(http://www.baseball-reference.com/bullpen/Three_True_Outcomes)

This stat's posterboy is generally considered to be Rob Deer, which makes me think that it's about as useful as the Curse of the Balboni was, even before it was broken by Gonzo: little more than a source of amusement.

Interesting stuff. Let me add my $0.02US.

Batting average is still important. If you don't believe me, consider a player with an OPS of 800, generally considered a decent total. It is far more likely that a player with an OPS of 800 will have a batting average of nearly .300 than one near .200.

The reason this is so is that both components of OPS, OBP and SLUG, implicitly contain batting average: OBP is basically batting average with walks included, while slugging percentage is batting average given 'extra credit' for extra-base hits.

The practical upshot is that for pretty much every real player in MLB, his OPS is determined largely by his batting average: a .275/800 hitter is 11/16th (about 69%) batting average, while a .200/500 player is 80% batting average. In this sense, it's hard for a player to hit .300 and not still be valuable, just as it's hard for a player to hit .200 and still be valuable.

The way in which batting average is overrated (and I do believe it's overrated) is when batting average is used as a 'tiebreaker' between players who otherwise seem fairly equivalent offensively: for instance, consider two players with roughly an 800 career OPS. One of them was a first-ballot Hall-of-Famer, while the other will probably never be considered a Hall-of-Famer. Granted, there are other factors involved here, but one of the big ones is that the HoFer's career BA is .300 (Paul Molitor), while the other player's career BA is just .272 (Jermaine Dye).

I will say that the primary value of batting average rests in its large sample sizes: taking any subset of batting average (with RISP, in Close and Late situations, etc.) is basically worthless, as it's almost always impossible to state with any certainty that the difference between a player's situational BA and his overall BA is due to anything other than simple random chance.

The challenge of getting a round bat to connect with a round ball (and make it go somewhere) fascinates me, so I find batting averages interesting.

I think batting average is an incomplete picture by itself, but, at the moment, that's what shows up on the scoreboard at RFK Stadium for Nationals home games. I hope OBP and/or OPS go on the scoreboard later, and this may be part of the transition several people have mentioned. But right now, I'll settle for scoreboard lineups at the park that don't list two shortstops, and check other statistics at home.

The more I think about it, the more I'd rather see OBP supplant BA as the yardstick. Sure, hitting's great and hitting wins more games than simply reaching base... which is why hitting counts for the bulk of OBP. To me, it's worth more to know how frequently the batter's successful at reaching base, not just successful at hitting.

Picture this- 2006 ALCS, Game 7, at the Dome vs. the defending champs, bottom of the 9th, tied at 0, bases loaded, two out, the Chairman's up and he's worked the count full.

Using this year's stats (because my internet provider has a filter in place preventing me from accessing next year's), do you care that he batted .294 or that he successfully reached base .372 of the time (and walked 61 times vs. 64 strikeouts)?

Or hell, look at the 1999 NLCS- not many Mets fans cared HOW game 5's winning run was driven in in the bottom of the 15th (but Robin Ventura's Grand Slam Single sure does make one for the storybooks). Similarly, it didn't matter to Braves fans that the WS berth-clinching run was plated in game 6 when Kenny Rogers walked Andruw Jones with the bases juiced in the bottom of the 11th.

Even better- 7/1/90, Andy Hawkins pitches a CG no-hitter (for the Yankees, against the White Sox) and still manages to lose FOUR TO NOTHING thanks to the yellow and white uniforms provided by Chico's Bail Bonds.

The Sox's BA was .000 that day but hey- a win's a win.

Crap, I forgot to mention this- click my name for an article I found while researching that last post... tells some very entertaining stories about Dock Ellis, including the one about how he pitched a CG no-hitter while on LSD.

Maybe it's because the only access most people have to baseball statistics is what is printed in a basic box score: at-bats, runs, hits and rbi. If that's what you are used to seeing, then batting average makes the most sense for evaluating how well a player has done over the season.

Dang, I wish I had gotten into this discussion earlier. Now all my points have been made. I don't even get to vent my horror at the belief that a player's past 3 at-bats can tell you as much about the likelihood of a hit in the next one as his past 300 at-bats can.

I value fluency in both sass and stats. It's fun to find a post here in a different language. I hope to return to sass soon, though. My native tongue.

Here's another question for discussion: Is "streakiness" real in sports, or can it simply be explained by random variations that even out in the long haul? That is, would you expect the same degree of streaks when rolling for 1's and 6's on a standard die as you would expect when watching a "streaky" .333 hitter over the course of his season (assuming he's consistently healthy)?

Click my name to visit a university professor's stats-oriented webpage dedicated to determining whether having a "hot hand" is a real phenomenon in sports.

I think that the reason that BA has survived is that it is easy to calculate and it does have some correlation to offensive production (runs).

The reason that statheads don't much care for it is that numbers like OBP and OPS have a much BETTER correlation to runs. If it has a better correlation to runs than BA or OBP, which is does, I would hardly call OPS a "junk stat."

Harmon Killebrew once said that he didn't really pay attention to his batting average. Good call, Harmon. Although he had a batting average in the mid .250s for his career, he walked a ton and hit those 573 HRs. His OPS was *pretty* and it more truly described his contribution.

BA fans can think of the Killer as a .256 hitter. I prefer to think of him as someone with an .884 OPS, back when that was a gaudy number (143 OPS+, which is adjusted for park effects and compared to the league average -- 100 being average). The Killebrew example should be the one that forever converts Twins fans from BA to OPS.

One more comment about Harmon. The league batting average during his career was .259. Thus, if batting average is your primary metric, then you'd have to argue that the Killer was a below average hitter or at best, average.

Further, you'd have to argue that his 1969 MVP was not deserved -- he hit only .276. But, he had the BEST OBP in the league (he walked 145 times) and second best OPS (1.011). Oh, and he hit 49 HRs.

Gosh, you go away for a little while and the discussion balloons. On a less technical note than the one I sounded on the use of batting average yesterday, what it comes down to for me is that batting average is a stat that has less value than was once thought, but isn't to be totally discarded like RBI or pitcher wins. Unlike those two things it is indicative, in some broad way, of a skill applied effectively (or not).

I have always viewed baseball as essentially two things: Probabalistic, and binary. Sure, things like AVG, OBP and SLG are measurements and not effectively probabilities, they give us an indication of what the probabilities are, and every move made on a macro level in baseball -- that is, above the level of physically swinging the bat -- is made with an eye toward increasing the probability of scoring a run, or preventing one. This is taken to absurd lengths by the Tony Larussas of the world, who feel they must carry seventeen left-handed relievers to get one out at a time, but it is, I believe, a truth of the game. And the game is binary in that almost every action that occurs can be catalogued as a "yes" or a "no" response to a question: Did the pitcher throw a strike? Did the batter make contact? Was it a ground ball? Was the fielder able to reach and properly field it? Was his throw accurate? Did it beat the runner? The binary nature of the game lends greatly to its appreciation through statistics -- more traditional ones like batting average and fielding percentage, and more recent innovations such as zone rating or EqA. Other sports are not so easily collated and codified. It's one of the things I like about baseball: There's a sense that everything is knowable, if you just think hard enough.

Okay, I'm done.

i'm a stathead, sure. i like to tinker with runs created formulas. but you gotta ask yourself: what's fun to root for? what's fun to watch?

the only guy who i ever saw make drawing walks incredibly fun to watch was ricky henderson. when he was in the mood, he'd just slap foul balls until the pitcher threw ball four. then alfonzo would smack one down the third-base line and olerud would draw a full-count walk, o the '99 mets.

anyway. it's fun to root for david wright to hit .300 and drive in 100; it's fun to look on the back of a baseball card; it's fun to see nice round numbers. it's fun to think of the twins as batgirl characters!

"My own theory is that statheads don't like it because it's too easy to calculate and everyone can do it once you've mastered long division."

My shot at brevity: They don't like it because it's limited both descriptively (a single counts the same as a home run) and predictively (as has been noted, for most players BA bounces around year-to-year).

This is why statheads burn with righteous indignation when, for example, earlier in the season Dick Bremer cited Jacque Jones's improved batting average against lefties and jumped to the conclusion Jones is becoming better against southpaws. It's distorting the issue. It's hiding the fact that Jones's "improvement" was really only a matter of picking up a few more scratch singles here and there (cue the Crash Davis pool hall scene). It's also ignoring that Jones was still failing too often. It's also denying that Jones really isn't close to the hitter facing lefties as he is righties.

I still think batting average is very useful, however, for several reasons:

1) As much as statheads have fetishes for guys who hit only .260 but post .400 obp's, there really aren't a whole lot of beasts like that. For most players, batting average is going to be a huge component of their getting on base.

2) If you can find a guy who can consistently post outstanding BA's, that's a pretty good weapon. Nobody's complaining when they have Ichiro on their team, or Tony Gwynn, or Kirby Puckett.

3) Many statheads, particularly prospect hounds, categorize players according to "old player skills" (lots of walks and power, bad defense, poor footspeed) and "young player skills" (good batting average, good defense, good speed). Very often players from the "young player skill" group go on to have more productive and longer careers than guys from the "old" group, who peak a little earlier but then fall off the cliff (like Ben Grieve). This isn't always true, of course, but it is a pattern for a lot of players.

4) In general, hits are more productive than walks.

Dang! The "streakiness" question brings up one of my favorite questions: are monthly averages meaningful? I always see them posted and pretty much disregard them. I'm curious what folks think about the guy who "warms up when the weather warms up", etc.

Maudeleyn: It's hard to deny that some players are "slow starters," or that so-and-so has been ridiculously good in June but awful in July for 8 years in a row. But with so many players in the league, it could just be chance that a few guys' hot streaks line up from year to year. I generally dismiss these stats, too. Mostly...

My favorite question is: When a batter is "seeing the ball well," is it really him, or is it just that he's been getting a little luckier than usual lately, soon to be followed by "regression toward the mean," that is, a descent back to Earth?

I wouldn't call it being lucky so much as being in a groove where your pitch selection and timing are vastly improved... but yeah, I think it's a streak that will come to an end... because if it was really something the batter was doing, those streaks would eventually start lasting longer and longer as guys figured out how they were able to "see the ball" so well.

Some of my pet peeve stats are those "lifetime vs. " numbers... for instance, let's say the Mets and Astros are playing and Jeff Don'tfeelwell comes up against Pedro. The graphic on the TV says "Bagwell has a .300 lifetime average vs. the Mets". So what? The 2005 Mets are not the 2004 Mets, are not the 1991 Mets.

Yeah, or "Jim Thome always hit well in the Metrodome." Could that be that the only team he ever played in the Metrodome was the Twins, mostly during the sucking years? Lemme guess, he also hit much better at Tropicana Field than in Yankee Stadium. Yes, a stadium's dimensions can better "fit" a certain player, but come on... you can't just ignore other reasons for a player's historical success.

A commenter suggests that a reason that batting average is important is because:

4) In general, hits are more productive than walks.

I repectfully point out that OPS accounts for hits in both slugging and OBP, whereas walks only show up in OBP, thus hits are given more weight than walks. I also respectfully point out that doubles, triples and home runs are more productive than singles. This is reflected in OPS, but not at all in average.

"...doubles, triples and home runs are more productive than singles."

That's kind of tricky, too... bases-loaded single that drives in two or three vs. bases-empty triple. See? ;-)

Sacky said: "My favorite question is: When a batter is "seeing the ball well," is it really him, or is it just that he's been getting a little luckier than usual lately, soon to be followed by "regression toward the mean," that is, a descent back to Earth?"

Example: Jacque Jones

I think batting average is still useful at comparing players

Under .200 = very bad (unless pitcher)
.200-.250 = bad
.250-.275 = okay
.275-.300 = good
.300-.350 = great
.350-.399 = MVP
Over .400 = Ted Williams

however, if you're comparing players within a group (ie .275-.300) then you need more than BA.