Some Curious Things About BABIP

[RonCo]

Knock...there's also no reason you need to manually pick through 1974 merely because you are familiar with that year. Just go to the Lahman database, download the excel/csv version, and you'll have every player throughout the history of the MLB available to you in moments. Just open in Excel, and play away.

[RonCo]

Originally Posted by jbergey22

Perhaps being traded to good teams helped the pitchers BABIP in your example Ron. Teams in a pennant race tend to make deals to improve at some point in the season. Athough Im not sure how common that was from 60-85.

Teams have been making trade deadline deals to contend since the dawn of time.

Your guess is as good as any. I would be reluctant to draw any inference from my study as it stands other than the obvious "pitchers BABIP are unpredictable."

[knockahoma]

Thank you, Ron. And thank you Curtis for the nice thought.

I'm one of those guys who loves reading philosophy. I've read a fair amount of that stuff. Also dig the quantum physics, what I can grasp of it. So, I come to sabermetrics looking through those lenses. I'd love to launch off into how I see those disciplines affect sabermetrics. I fear I'd bore the hell out of everybody.

I'll try to follow through on this studies, if I can find the adjusted BABIPs.

Someone asked my background in studies. Mostly looking at media ratings and marketing data, which also contains a fair amount of white noise. If you got through my example of the "Jukebox Effect" in radio, you know what I mean. You approach statistics like you do a snake when farming its venom. Very good things and very bad things can happen. Careful!

So, I'm trying to articulate my concern with chaos theory and the butterfly effect in baseball. My concern with mass player studies is that each player has a butterfly, or two, in his stats. And so, there's a danger that what we call one big butterfly is really thousands of little butterflies. I know you guys who do the math already know this.

But, eventually, you have to check on Messersmith and see what specific butterflies he had around him. Sometimes, in MY experience with media stuff, mass studies are treated as if they are fractals-- A geometric pattern that is repeated at ever smaller scales. Too often, I see studies stop at the mass scale and template that onto all individuals.

Someone asked why former baseball players struggle with sabermetrics. I'm not sure they struggle. I think they're close enough to see the swarm of butterflies that math struggles to see. Yogi's 90% butterfly is especially worth noting. The human eye makes a lot of mistakes. But, even the deepest math owes its birth to human observation.

I was encouraged by Ron's numbers. I'll explain why on my next post, probably tomorrow. In the meantime, I've taken the ideas given to my engineer who has a very good math background, and together we'll try to get those numbers requested.

Couple of quick "Traditionalist" thoughts on why pitchers changing teams in mid-season did well.

*better defense
*change of scenery, better mental make-up
* A coach who found a way to make the player's pitches effective again.

But, when it's all said and done, I feel like I'll be right back where I started, with Messersmith. What butterflies were floating around him?

I probably come off as sounding pompous, or something with all that philosophy yada yada. In fact, the reverse is true. It's my deep respect for James' "unseen Monster" butterflies that humbles me as I try to examine all this.

[Curtis]

Originally Posted by knockahoma

It's my deep respect for James' "unseen Monster" butterflies that humbles me as I try to examine all this.

Mothra lives!

[knockahoma]

Damn it, Curtis

I was really hoping to erase that morass of yada yada before anyone posted.

Earlier yesterday, I had reminded myself of my original goal. I don't really want to get into what a pitcher does, or doesn't control on BABIP. That's following the swarm of butterflies into a pretty deep fog bank, at this point.

I'd like to find out how much BABIPs change from year to year, or from team to team (era dependent, I sense) and see if OOTP can model that. Ron's numbers were right along the lines of the goal. That aim seems do-able, especially as I seem to recall someone writing that "God" does code in some pitcher influence on BABIP. Just jump it up a bit, or move it down.

And damn the butterflies.

[knockahoma]

In either case, if I were a GM looking to make a deal I would look at BABIP only at the very end of a much longer assessment trail.

Same here.

Instead, I suggest designing an experiment that studies a topic, gathering enough data to make the experiment hold its weight, and let it fly. I also suggest presenting the data in a format that makes a case. Given Knock's professional status, I assume he knows all this, but it's not evident from what he's doing here...maybe it's just me...I don't know

I am familiar with that method. You kind of get familiar with that with physics and philosophy, etc. It's how Schrodinger's cat has made Einstein (at least for now) look wrong when he said, "God doesn't play with dice!" It's the scientific method that is utilized in philosophical arguments (such as testing Determinism against Free Will, the nature of language, etc.). Testing the hypothesis is game-time in the Scientific Method.

Not so much from work. We examine and evaluate studies dealing with broadcasting, advertising, marketing, ratings, etc. But, scientific studies don't originate here.

I'm still trying to get a handle on keeping all things consistent in these studies. I was aware of the adjusted BABIP issue. I'd worried about that several pages back, but I simply haven't found a page that lays that out for more than a single pitcher at a time. I'm looking for at least Team pages so I don't have to mouse click thousands of times to get what I want. I'm guessing you're looking, too?

I'm also musing over various OOTP settings to give the game its fairest opportunity, whether fictional, or historical.

So, a night largely made up of insomnia left me with a few hours. Did some looking at BABIP... to follow shortly.

[knockahoma]

Alright. A few studies made from a sleep-deprived guy. Before I comment, I'll just lay this stuff out. And just a heads-up. I may be off on one or two pitchers due to bleary eyes and mind, but it's 95% correct, at least.

I wanted to get a bit closer than Ron's groupings. This is a 74 replay through the 75 season, three year calc, injury and trade off, player develop off. I counted all pitchers who had at least 150 IP for both years. Then I looked at the variance between their babip from 74 to 75. The following shows how many pitchers fell into each grouping:

POINT VARIANCE
Between 74 and 75

1-10 33 Pitchers

11-20 21 Pitchers

21-30 15 pitchers

31-40 4 pitchers

41-50 3 pitchers

51-60 2 pitchers

61-70 2 pitchers

70+ 0 pitchers.

Next, I went to the real numbers for the 72,73 season. I did only the NL on this one. I had opened the 74 league and just looked back on the previous two seasons in the NL. Then I went to the AL and within about ten pitchers I said, " What's going on? BABIP is being blown out of the water, here!" And then I remembered: The DH rule hit baseball in 73. So, I finished the NL, for now. Same rules apply:

POINT VARIANCE between 72 and 73
(real life, BABIP not adjusted).

1-10 12 pitchers

11-20 7 pitchers

21-30 8 pitchers

31-40 6 pitchers

41-50 2 pitchers

51-60 1 pitcher

I was curious if variances had changed much in 30 odd years. So, the numbers from 2006 and 2007, REAL LIFE NUMBERS here.

1-10 20 pitchers

11-20 16 pitchers

21-30 8 pitchers

31-40 6 pitchers

41-50 2 pitchers

51-60 2 pitchers

61-70 2 pitchers

70 + 0 pitchers

And finally, for now, a 16 team fictional set in the modern day, no trades, no injuries.

1-10 18 pitchers

11-20 17 pitchers

21-30 13 pitchers

31-40 2 pitchers

41-50 6 pitchers

51-60 1 pitchers

61-70 0 pitchers

70 + 2 pitchers

By the way, 4 of the top group were on the same team. Four others were on two other teams. So, 3 teams had 8 of the 18 pitchers in the top group. I will check the defensive abilities of those teams when I get a chance, today.


Overall, I like those numbers quite a bit. I have at least one more test I want to do, and that's a historical replay with trades on. I mean, there are so many potential contaminants in all this, but I guess it's a start. Actually, as I finished up, I wondered if there was too much consistency for some. But, what would be considered a consistent "babip" variation? I've asked before. I don't think I've ever seen a comment on that.

[RonCo]

Originally Posted by knockahoma

Same here. This is one of the weirder aspects of the entire internet-wide discussion. BABIP isn't a "traditionalist's" stat. It originated with fantasy baseball, I believe I was told. So... the stat guys developed a stat which was later attacked as largely unimportant by other stat guys.

You're thinking of WHIP, which was made popular by inclusion in the original Rotisserie League. And WHIP has not been "attacked" by anyone, it's just not a highly useful metric.

[knockahoma]

I can't believe someone is up.

No, I wasn't clear. The article I read was talking simply about hits versus innings pitched. A baseball GM, or manager, or coach was talking about it. Even that was considered innovative at the time-- at least by that guy.

I was just trying to show by that "innovation" how traditionalists had paid far more attention to HRs, BBs, K's when judging pitchers before something like BABIP came along.

So, BABIP was developed (presumeably) as a useful tool in evaluating pitchers by stat guys, then Voros came along and said, basically, this stat isn't that useful. It would be better to pay more attention to HRs, BBs, K's when judging a pitcher.

And many in the traditional pews nodded their heads and said, " Amen!"

I guess I just don't remember the traditional guys in baseball ever giving more concern over balls in play than they did the big three.

But, now I am doubting my memory. I thought in one of the BABIP threads, someone had written that BABIP came out of rotisserie ball, not sabermetrics. I've never discussed WHIP on this site. What's the final answer?

[RonCo]

Originally Posted by knockahoma

By the way, 4 of the top group were on the same team. Four others were on two other teams. So, 3 teams had 8 of the 18 pitchers in the top group. I will check the defensive abilities of those teams when I get a chance, today.

Overall, I like those numbers quite a bit. I have at least one more test I want to do, and that's a historical replay with trades on. I mean, there are so many potential contaminants in all this, but I guess it's a start. Actually, as I finished up, I wondered if there was too much consistency for some. But, what would be considered a consistent "babip" variation? I've asked before. I don't think I've ever seen a comment on that.

For one run, it's not too bad. But yes, it appears from this one pass that OOTP pitchers are more consistent than real life pitchers were/are. This is possibly due to defense being a little too good in OOTP, or possibly because Markus has coded some small BABIP control into his game that is stronger than what exists in real life. I'm guessing a bit of both.

You're answering your own question, BTW. The answer to "what is a reasonable variation of BABIP?" is a complex discussion.

Note -- when trying to determine the overall quality of the sim, using unadjusted numbers is sometimes valuable. This is because the question at hand is "how are all the elements working together?" But when comparing the sim to real life to compare individual elements of the design, the process needs to get a lot better constrained.

[knockahoma]

You're answering your own question, BTW. The answer to "what is a reasonable variation of BABIP?" is a complex discussion.

I can understand how it would be. But the fellas better get on it. It sounds kind of weird when someone says, " pitcher's have too much variation on their BABIP to be in good control of it."

Q: What's a reasonable variation?

A: We're not sure, yet.

Note -- when trying to determine the overall quality of the sim, using unadjusted numbers is sometimes valuable. This is because the question at hand is "how are all the elements working together?" But when comparing the sim to real life to compare individual elements of the design, the process needs to get a lot better constrained.

I understand. I had posted my concern about that several pages ago and on one of my last posts. I would use those adjusted BABIPs, but I can't find such a page, except one that loads players one by one and I think I'll skip that thousand click journey. I've googled my heart out. If anyone knows of a site with adjusted BABIP, shown at least by team...

[knockahoma]

For one run, it's not too bad. But yes, it appears from this one pass that OOTP pitchers are more consistent than real life pitchers were/are. This is possibly due to defense being a little too good in OOTP, or possibly because Markus has coded some small BABIP control into his game that is stronger than what exists in real life. I'm guessing a bit of both.

Might be.

If he can adjust by gathering real-life BABIP variances (in all their shapes and forms) and get to a place where the game apes those results.... could be something special. Defense and pitching balanced. I think I'm pretty pleased with hitters in X.. at least in the early going.

[RonCo]

There are other things that can affect this -- imbalances in player ratings, league totals, park factors, in particular. Also, if a single team amasses a collection of players on the far end of a scale, the numbers _could_ sway one way or another.

Based on a few things Markus has said, I'm thinking he's double-counted a couple things people are using as "possible BABIP control factors." Hence his pitchers control BABIP a little bit more than real pitchers do. Regardless, from your numbers above, assuming the additional consistency is true, whatever influence is happening is not influencing things _too_ badly.

Either way, though, while OOTP is going to create several consistent pitchers, it can't today guarantee that the right guys (Andy Messersmith?) are going to be the consistent ones in any particular year. To do this, random nature of the game of baseball would need to be coded out of the game.

[knockahoma]

A note on the fictional. It was a no-trade league. The pitchers and defenses were pretty static. I'm curious to see what trades might do to those BABIP groups.

Yeah, I can accept Messersmith bombing out once in a while. I actually like that possibility being in the game. The thing that seemed unreasonable ---if so much was due to chance--- is that he had consistently good years with the Dodgers and consistently bad years with the Padres. It was as if both teams were weighted for pitchers. That signaled to me that something might be slightly out of whack.

If Messersmith imports well enough to be good with the Dodgers 5 out of 6 seasons, you'd think he'd get at least 2 or 3 good years out of 6 with the Padres. If some thought I was just bothered seeing Messersmith fail, that wasn't it at all.

[RonCo]

Originally Posted by knockahoma

I can't believe someone is up.

No, I wasn't clear. The article I read was talking simply about hits versus innings pitched. A baseball GM, or manager, or coach was talking about it. Even that was considered innovative at the time-- at least by that guy.

I was just trying to show by that "innovation" how traditionalists had paid far more attention to HRs, BBs, K's when judging pitchers before something like BABIP came along.

So, BABIP was developed (presumeably) as a useful tool in evaluating pitchers by stat guys, then Voros came along and said, basically, this stat isn't that useful. It would be better to pay more attention to HRs, BBs, K's when judging a pitcher.

And many in the traditional pews nodded their heads and said, " Amen!"

I guess I just don't remember the traditional guys in baseball ever giving more concern over balls in play than they did the big three.

But, now I am doubting my memory. I thought in one of the BABIP threads, someone had written that BABIP came out of rotisserie ball, not sabermetrics. I've never discussed WHIP on this site. What's the final answer?

Geez.

Baseball guys didn't call it BABIP. They called it "hits." And they felt a pitcher _strongly_ controlled the hits he gives up--many still do. Perhaps they are right, despite a lot of performance data that refutes this "strong" control. At the end of the day, though, a lot of sabermetrics supports the old-school thoughts...it's just where sabermetrics studies clash that gets all the ink.

I'm re-reading Bill James's old Baseball Abstracts right now, and its incredible how coarse and rudimentary baseball knowledge really was when I was a kid (in the mid-80s). Earl Weaver was chided because he kept information on 3x5 cards. Davey Johnson was considered an outcast because he started using computers to keep stats. James's running commentary in these books is like pure revolutionary poetry. His praise of forward thinking guys like Bobby Cox, Earl Weaver, and Whitey Herzog is golden and bears up through history...his ripping of old-school baseball establishment is equally brilliant. I can name a bunch of big-name baseball guys that wold chuckle at the use of Bill James and the word "humility" in the same breath.

Anyway...

WHIP was in Rotisserie Ball. BABIP was not.

[RonCo]

Originally Posted by knockahoma

Yeah, I can accept Messersmith bombing out once in a while. I actually like that possibility being in the game. The thing that seemed unreasonable ---if so much was due to chance--- is that he had consistently good years with the Dodgers and consistently bad years with the Padres. It was as if both teams were weighted for pitchers. That signaled to me that something might be slightly out of whack.

If Messersmith imports well enough to be good with the Dodgers 5 out of 6 seasons, you'd think he'd get at least 2 or 3 good years out of 6 with the Padres. If some thought I was just bothered seeing Messersmith fail, that wasn't it at all.

Your data said that in real life (72-73) only 12/36 pitchers (33%) stayed within 10 points. Similar numbers in (2005-06) -- 20/56 (35%).

So this would say that if Markus has it right, Messersmith could be expected to have a BABIP within 10 points of his real value only 3.5 out of every 10 sims.

[knockahoma]

Geez,

Sometimes I wonder why we miss each other so much when we probably agree on so much. Even now, I'm not sure what accent to put on your "Geez".

Geez.

Baseball guys didn't call it BABIP. They called it "hits." And they felt a pitcher _strongly_ controlled the hits he gives up--many still do. Perhaps they are right, despite a lot of performance data that refutes this "strong" control. At the end of the day, though, a lot of sabermetrics supports the old-school thoughts...it's just where sabermetrics studies clash that gets all the ink.

I know they called them "hits". What I was saying was that, even back in the day, managers were more concerned with pitchers giving up homers, a lot of walks, or getting enough strikeouts than they were just with hits.

In other words, BABIP came along and said, " Look at me! I'm a new stat! Look at me!" Several years later, Voros comes along and says, " That ain't much of a stat. Look at HRs, BBs, and Ks first."

But that's what baseball people had been doing a long time... and never really stopped doing, in my opinion.

[lynchjm24]

Originally Posted by knockahoma

I can't believe someone is up.

No, I wasn't clear. The article I read was talking simply about hits versus innings pitched. A baseball GM, or manager, or coach was talking about it. Even that was considered innovative at the time-- at least by that guy.

I was just trying to show by that "innovation" how traditionalists had paid far more attention to HRs, BBs, K's when judging pitchers before something like BABIP came along.

So, BABIP was developed (presumeably) as a useful tool in evaluating pitchers by stat guys, then Voros came along and said, basically, this stat isn't that useful. It would be better to pay more attention to HRs, BBs, K's when judging a pitcher.

And many in the traditional pews nodded their heads and said, " Amen!"

I guess I just don't remember the traditional guys in baseball ever giving more concern over balls in play than they did the big three.

But, now I am doubting my memory. I thought in one of the BABIP threads, someone had written that BABIP came out of rotisserie ball, not sabermetrics. I've never discussed WHIP on this site. What's the final answer?

I don't remember BABIP even existing before Voros. BABIP has never had anything to do with Roto.

[andymac]

Originally Posted by knockahoma

But that's what baseball people had been doing a long time... and never really stopped doing, in my opinion.

I think this is where you are off. I absolutely think that "baseball people" used to look at "hits" as being just as important as HR's, BB's & K's. In fact, they didn't even look at HR's, BB's & K's all that much. They looked at IP, Hits & E.R.A.

[knockahoma]

dp