Some Curious Things About BABIP

[knockahoma]

The numbers you posted last night looked like they lined up pretty well with real life, am I missing something in those numbers? In your own post you say they look pretty good.

I thought so, too.

Messersmith could be a red flag, or he could be one pitcher in a handful of sims. Based on everything else you've posted I don't know why you'd still assume he is a sign of a larger problem.

The results I got made me wonder how strongly defense ties to pitcher performance. If Messersmith started off as the "same" rated pitcher with San Diego and LA, would his career be vastly different?

In LA, Messersmith was consistently one of the top pitchers in the league with a sparkling BABIP. Not always. But, five out of six imports.

In San Diego, his fortunes were reversed; his HRs, BBs, K's remained constant, but his BABIP skyrocketed on average 50-80 points. Messersmith fell into the bottom 20% of pitcher BABIP.

That strikes me as more than just chance. To me, a "red-flag" is not an indictment. It's just signals a look-see.

I haven't had a chance to sit down and do percentages. These are the real life variances of pitchers with 150 innings in both 1972 and 1973 (national league only).

POINT VAR.
1-10 12 pitchers

11-20 7 pitchers

21-30 8 pitchers

31-40 6 pitchers

41-50 2 pitchers

51-60 1 pitcher

The Real Life numbers of pitchers from 06 to 07 were:
[QUOTE]1-10 20 pitchers

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

Then, a 16 team fictional.

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

Does the fictional more closely resemble the 1970s numbers?



Looking at so many historical BABIP over the past week, I get the image of pitchers as captains at stormy sea. No one can keep a perfect rudder. But some keep the boat closer to course than others.

There are several questions that come to mind.

* What will happen to the fictional numbers once pitchers and fielders are traded?

* As League BABIP has come up 20-25 points the past 30 years, how has that affected pitcher distribution in those categories?

* I'd like to take the same study and do 3 years in a row this time for real life BABIP variance. I'm curious how many real life pitchers could stay within a 10 or 15 point variance over three seasons.

* I'd like to take the same study and do 4 years in a row.

* I could live with the spread I see right now in OOTP.

* I wonder if OOTP often turns a pitcher the quality of a Messersmith into a lower tier pitcher year after year, sim after sim, simply by the quality of his team's defense-- not through chance-- but through strongly influencing machinations. Such a tendency would also affect fictional leagues.

* How often does that kind of thing happen with veteran quality pitchers in OOTP?

* What would the variance breakdown look like if we could categorize by IPs, maybe by increemnts of 50? Using only pitchers with 150 innings in both seasons erases a lot of other stuff that might be diagnosed at least in a game like OOTP.

* If it happens frequently, then armed with that knowledge, can we too often trade for pitchers with average to poor HR, K, BB totals, plug them into the "super BABIP defense" and turn them into stars?

* Where the hell can somebody find adjusted BABIP in a format that doesn't require 10,000 clicks?

Some of those questions apply to the game and some eminate simply from pure curiousity.

[knockahoma]

Lynch,

Are you seeing my question about which defensive stats you think are worthwhile?

[jbergey22]

Knock,

Did you see Rons reply above. Is that what you are suggesting taking a variable BABIP off the pitchers stats over that variable off the leagues BABIP?

[knockahoma]

Knock,

Did you see Rons reply above. Is that what you are suggesting taking a variable BABIP off the pitchers stats over that variable off the leagues BABIP?

Not yet. I'd like to do a lot more tests. I'm not attacking OOTP. I'm just asking questions, ya know? I don't think I know enough about OOTP's BABIP program, yet.

I'd like to know that OOTP does two things:

1. Does a pretty good job of modeling league wide BABIP variance.

2. Gives BABIP defenses the right balance of influence over pitchers.

What if the situation is this? If you're San Diego with a bad defense, is it pointless to go after even league aces like Messersmith until you build a defense that can give him a good BABIP, thus elevating him to a good pitcher. It seems that was the case in my sims.


San Diego troubled me because Messersmith's other numbers (HRs, Ks, BBs) were pretty spot on. . So, apparently, the randomness didn't affect those numbers. Then, is it defense, or randomness, that blew up Messersmith's BABIP numbers over and over again in San Diego?

Again, this is an issue that would affect fictionals, too.

I probably won't be posting for a few days, just as a heads-up.

[jbergey22]

Originally Posted by knockahoma

Not yet. I'd like to do a lot more tests. I'm not attacking OOTP. I'm just asking questions, ya know? I don't think I know enough about OOTP's BABIP program, yet.

I'd like to know that OOTP does two things:

1. Does a pretty good job of modeling league wide BABIP variance.

2. Give BABIP defenses the right balance of influence over pitchers. What if the situation is this? If you're San Diego with a bad defense, is it pointless to go after even league aces like Messersmith until you build a defense that can give him a good BABIP, thus elevating him to a good pitcher. San Diego troubled me because Messersmith's other numbers (HRs, Ks, BBs) were pretty spot on. . So, apparently, the randomness didn't affect those numbers. Then, is it defense, or randomness, that blew up Messersmith's BABIP numbers in San Diego?

Good question. He wouldnt put up the Messersmith numbers you were used to seeing but just imagine how the poor pitchers would do behind that defense? If you think of it that way he is likely worth getting but wont be good enough to dominate until they get the great defense behind them. Have you tried taking a poor pitcher with great defense and moving him to a poor defensive team?(maybe another test?) Instead of an ERA change of 1-1.5 the poor pitchers ERA might change 2-3 runs per game. Just a guess.

As far as your last question. It would probably be some combination of defense, random(less likely because you ran a few tests) and possibly a park that allowed more hits. I wouldnt think a difference of that much could just be 1 thing.

[lynchjm24]

Originally Posted by knockahoma

Lynch,

Are you seeing my question about which defensive stats you think are worthwhile?

I saw it, forgot to answer it.

There are certainly none from history that are of any value. Maybe if you were looking at double play totals for middle infielders.

Today, I think they are getting better. I guess UZR is probably the best one. I had a long drawnout discussion with MGL back in 2002 or 2003 when he was first making that data public on Baseball Primer (now Baseball Think Factory). I still like a combination of scouting with some of those stats.

Probably in a few years we'll have numbers that are much better then we do today. Pitch f/x/Hit f/x or whatever they are called... that technology is going to advance pretty quickly I think.

[lynchjm24]

Originally Posted by knockahoma

Not yet. I'd like to do a lot more tests. I'm not attacking OOTP. I'm just asking questions, ya know? I don't think I know enough about OOTP's BABIP program, yet.

I'd like to know that OOTP does two things:

1. Does a pretty good job of modeling league wide BABIP variance.

2. Gives BABIP defenses the right balance of influence over pitchers.

What if the situation is this? If you're San Diego with a bad defense, is it pointless to go after even league aces like Messersmith until you build a defense that can give him a good BABIP, thus elevating him to a good pitcher. It seems that was the case in my sims.


San Diego troubled me because Messersmith's other numbers (HRs, Ks, BBs) were pretty spot on. . So, apparently, the randomness didn't affect those numbers. Then, is it defense, or randomness, that blew up Messersmith's BABIP numbers over and over again in San Diego?

Again, this is an issue that would affect fictionals, too.

I probably won't be posting for a few days, just as a heads-up.

I think you need to get a much bigger sample then 1 pitcher and 6 sims before you come to any kind of conclusion.

You'd also need to run some sort of control tests where you give everyone the same defense.

That would be a pretty cool feature in OOTP actually, if Markus built in the ability to test and in broad strokes make all defensive players exactly average for example.

[lynchjm24]

Originally Posted by knockahoma

So, apparently, the randomness didn't affect those numbers. Then, is it defense, or randomness, that blew up Messersmith's BABIP numbers over and over again in San Diego?
.

I think you need to consider park factors on top of defense and randomness. Park factors in OOTP have always done some pretty wonky things.

[Curtis]

Originally Posted by lynchjm24

You'd also need to run some sort of control tests where you give everyone the same defense.

That would be a pretty cool feature in OOTP actually, if Markus built in the ability to test and in broad strokes make all defensive players exactly average for example.

It (along with a bunch of other things) would be pretty easy to do, if Markus coded a 'mass select' feature into the game.

Just sayin'.

[RonCo]

It's easy to do even without the mass select. Just export the rosters in csv, change them all in Excel with a few clicks of the mouse, and import.

[Curtis]

It may be easy for you Ron, but I don't have Excel, and I don't understand the whole export/import thing. I'm not even certain what it means, and I sure don't know what it accomplishes.

[RonCo]

Any of the spreadsheet programs will work, I'm sure. Excel is just the most ubiquitous.

If you go into one of the set-up screens, you're able to export roster data in txt or csv format. Once you've done that you can mass-edit lots of ratings (but not all), and once that's done you can to a import on the same file. It accomplishes editing lots of players at the same time in whatever way you want them to be. I use it to set up test leagues with all players of one falvor or another quite often.

Spend 4-5 minutes trying it sometime, and see what you can do. It may resolve a portion of your angst. Just sayin'

[StyxNCa]

Originally Posted by Curtis

It may be easy for you Ron, but I don't have Excel, and I don't understand the whole export/import thing. I'm not even certain what it means, and I sure don't know what it accomplishes.

One thing it does accomplish is it sets all minor league players to having a major league contract, $0 but still a major league contract.

It also causes your players to disappear from your roster, lineup and pitching staff screens even though they are still on your team and one person said the aren't showing up on 22 teams in his league.

It also eliminates the R/L splits from the player's profile page. Who knows what else it accomplishes. Those are just the things I know about.

You don't even have to edit anything before import for all of this to happen.

I do not recommend usage of the import/export feature right now.

[Curtis]

Ronco, thanks for explaining.

I can't read a spreadsheet. I might have a spreadsheet program somewhere in my Applications folder (probably as part of the Appleworks suite), but I wouldn't know how to use it. I do my editing using anyof three or four plaintext type apps (textedit being my favorite, and I can just about struggle through TextWrangler). I've tried using OpenOffice, but I couldn't make heads nor tails of it and had to give up.

EDIT: Oh, and I'm off-topic, so don't feel like you need to continue to respond to me.

[professor ape]

With all of the talk about BABIP over the past few years I am curious why I haven't seen discussions about SLGIP. Are doubles and triples as random as hits in general? Do good defenses have a bigger impact on extra base hits? Are threads like this the reason for the 2012 doomsday myths?

[Le Grande Orange]

Originally Posted by Curtis

I can't read a spreadsheet. I might have a spreadsheet program somewhere in my Applications folder (probably as part of the Appleworks suite), but I wouldn't know how to use it. I do my editing using anyof three or four plaintext type apps (textedit being my favorite, and I can just about struggle through TextWrangler). I've tried using OpenOffice, but I couldn't make heads nor tails of it and had to give up.

I would suggest hanging in there and trying to understand whichever spreadsheet program you have.

Some years back, I didn't know anything at all about Excel. Then, I slowly started learning it. I'm far, far from being an expert on the program, but I know enough now to do some basic formulas in it. It comes in very handy for all sorts of things beyond just number-crunching.

[hitmanwa]

Where might 1 find how RTZ, FRAR and such are calculated? I've looked on Baseball-reference.com and haven't had any luck.

[StyxNCa]

I have always valued OB% considerably more than either AVG or BABIP. Even 40 years ago I didn't pay nearly as much attention to AVG as I did OB%. Even when I played the game on the field it was my OB% that I concentrated on.

For how I evaluate players, BABIP serves no purpose at all.