Best lefty off all time? My vote is Koufax!

[Peter_Spaeth]

Originally Posted by Snowman

My instincts about Maddux's BABIP were wrong. You're right, Maddux did beat the league average BABIP, particularly between 1992-1998 (see plot below). But you appear to be misunderstanding what I'm saying. I'm not saying Greg Maddux was just a "lucky" pitcher. He was an excellent pitcher. I'm saying that people conflate his remarkable ability to control the ball with him having the ability to also control where the ball goes after someone puts it into play. The extent to which pitchers actually have this ability is miniscule at best. It's probably at least an order of magnitude less than people are thinking of when they make that claim. Maddux rarely walked hitters. He led the league in BB/9 9 times, and was probably in the top 3 15 times or more. This was his superpower. As I mentioned earlier, there is some research (which I'd have to read again, it's been a while) that suggests a really strong pitcher may have a small, but measurable effect on their BABIP, but that estimate is only something like 5 points worth of BABIP, which is to say out of every 1,000 times a ball gets put into play, an elite pitcher is able to prevent an additional 5 of those into becoming hits than his peers (hence I said it's a tough sell). However, a pitcher's BABIP can often fluctuate 70 or 80 points from one season to the next. Even if 5 of those points are within their control, that still leaves 65 to 75 points worth of variance or "luck" which is completely outside of their control.

BABIP is a very useful statistic for putting other stats into context. It is influenced primarily by luck, but also by the defensive talent of the players on the field, the skill of the batters, and by the ballpark. Hitters have a fair amount of control over their BABIP numbers (though they are also very much subject to luck in the short term) as exit velocity is highly correlated to BABIP values. The harder you hit the ball, the more likely it is to drop in for a hit. But pitchers face an approximately uniform (top of the order inflated) distribution of batters, so hitting talent mostly evens out for them with some minor exceptions (e.g., pitching in the NL yields a slightly lower BABIP than the AL because of the DH spot, and pitching in a division that is stacked with good hitters can deflate your BABIP if you have a higher than average number of starts against strong offensive teams than your peers. But these effects are fairly small. The overwhelming majority of the variance in BABIP values is simply due to random chance. And this variance is actually pretty wide from season to season, and it correlates highly with the fluctuations you see with other stats that are highly subject to luck as well (like ERA and WHIP).

A pitcher like Maddux had a few things going for him which should have helped him outperform the league average BABIP numbers. He pitched in the NL, was in a pitcher's park, and had Andruw Jones chasing down balls for him in CF. I'm not sure exactly how much each of those factors weighs in exactly off the top of my head, but they do have a measurable impact. But even if it is true that a pitcher as great as Maddux is capable of "beating" the BABIP line, the evidence shows that it would only be to the tune of a few balls out of 1,000. That's certainly not what people who promote the idea that he can control ball flights with his pitching style mean when they make such claims. If playing in a pitcher's park is worth 1 or 2 balls per 1,000, and having Andruw Jones running down fly balls is worth 1 or 2 per 1,000, and pitching in the NL is worth 2-3 balls per 1,000 and having god-like control is worth 3-5 balls per 1,000, that would add up to someone like Maddux beating the BABIP line by 9 points.

If you haven't read it before, this is worth a read. It has a pretty good explanation of BABIP and why it's important.

https://library.fangraphs.com/pitching/babip/


And since I was wrong and am happy to admit when I'm wrong, here's a plot of Maddux vs the league average BABIP showing that he did in fact beat the league for a good several-year run in the 90s (note the blue line is MLB average, not NL average, which would be slightly lower).

Two points: one, you at first presented your "instincts" about Maddux as if they were facts you already knew. Read the language of your post.

Two, Kershaw's BABIP is 27 points below the ML average for his career. What's your take on that which obviously can't be explained by NL alone?

[Carter08]

Many probably know this but I’ll repeat it here. Simply amazing:

Maddux faced 20,421 batters during his time in the league. In those 20,421 at-bats, only 310 hitters saw a 3-0 count. Out of those 310, 3-0 counts, 177 of them were intentional walks.

[Peter_Spaeth]

Until he ended it with an intentional pass, Maddux once went 72 straight innings without a walk.

[Snowman]

Originally Posted by Peter_Spaeth

Until he ended it with an intentional pass, Maddux once went 72 straight innings without a walk.

And when asked about his scoreless innings steak, his response was "honestly, it was mostly luck"

[Carter08]

Originally Posted by Snowman

And when asked about his scoreless innings steak, his response was "honestly, it was mostly luck"

Humble guy. An admirable quality usually displayed by people confident they are good.

[Snowman]

Originally Posted by Carter08

Humble guy. An admirable quality usually displayed by people confident they are good.

Or perhaps more likely, he was simply a realist.

[Snowman]

Originally Posted by Peter_Spaeth

Two points: one, you at first presented your "instincts" about Maddux as if they were facts you already knew. Read the language of your post.

Two, Kershaw's BABIP is 27 points below the ML average for his career. What's your take on that which obviously can't be explained by NL alone?

Again, you conflating. The facts I already know are that a pitcher's BABIP regresses to the mean and each pitcher has little to no control over their values. What I was wrong about was that Maddux's values were 9 points below league average. But that still doesn't mean he is able to control his BABIP. If you look up his teammates, they too all beat the league average BABIP. In other words, the ballpark, pitching in the NL, and the defense behind him was responsible for most, if not all, of his ability to beat it.

As far as Kershaw goes, it appears to be the same thing. I just looked up 5 or 6 of his teammates over the years in LA to check their BABIP values. Grienke, Urias, Buehler, Jansen, Baez, all of them are 20 to 40 points below league average BABIP. Again, this means it is their defense, the fact that they all pitch in the NL, and the ballpark that account for the differences, not some magical ability that Kershaw possesses.

[Peter_Spaeth]

Originally Posted by Snowman

Again, you conflating. The facts I already know are that a pitcher's BABIP regresses to the mean and each pitcher has little to no control over their values. What I was wrong about was that Maddux's values were 9 points below league average. But that still doesn't mean he is able to control his BABIP. If you look up his teammates, they too all beat the league average BABIP. In other words, the ballpark, pitching in the NL, and the defense behind him was responsible for most, if not all, of his ability to beat it.

As far as Kershaw goes, it appears to be the same thing. I just looked up 5 or 6 of his teammates over the years in LA to check their BABIP values. Grienke, Urias, Buehler, Jansen, Baez, all of them are 20 to 40 points below league average BABIP. Again, this means it is their defense, the fact that they all pitch in the NL, and the ballpark that account for the differences, not some magical ability that Kershaw possesses.

Is it also the case that pitchers regress to the mean in extra base hits and home runs against?

[Snowman]

Originally Posted by Peter_Spaeth

Is it also the case that pitchers regress to the mean in extra base hits and home runs against?

No. They will regress to their own individual expected means, but not to the league averages. Bad pitchers serve up more meatballs than good pitchers. This is not contradictory to the discussion above.

[Peter_Spaeth]

Originally Posted by Snowman

No. They will regress to their own individual expected means, but not to the league averages. Bad pitchers serve up more meatballs than good pitchers. This is not contradictory to the discussion above.

If a pitcher like Maddux was better at keeping the ball in the park, and/or could limit extra base hits better, then that seems at least some evidence he could in fact control where/how hard the ball was hit against him, even if not reflected in batting average itself. Do you agree?

Take a hypothetical at bat, a bad pitcher hangs a curve and the batter hits it over the wall. Maddux paints the corner with a slider and the batter gets a bloop single off the end of the bat. Same BABIP but different (in most cases) outcome.

[Bigdaddy]

In Sandy's own words:

"I became a good pitcher when I stopped trying to make them miss the ball and started trying to make them hit it."


And the whole idea of 'weak contact' is within the pitcher's control - Are they consistently ahead or behind in the count: are they grooving a ball down the middle of the plate, or painting the corners; are they disrupting a batter's timing??? Great pitchers consistently pitch ahead in the count, paint the corners and keep batters off balance - and induce weak contact.

[Snowman]

Originally Posted by Bigdaddy

In Sandy's own words:

"I became a good pitcher when I stopped trying to make them miss the ball and started trying to make them hit it."

That's also when his BB/9 rate fell though. And when his strike zone grew.

[Snowman]

Originally Posted by Peter_Spaeth

If a pitcher like Maddux was better at keeping the ball in the park, and/or could limit extra base hits better, then that seems at least some evidence he could in fact control where/how hard the ball was hit against him, even if not reflected in batting average itself. Do you agree?

Take a hypothetical at bat, a bad pitcher hangs a curve and the batter hits it over the wall. Maddux paints the corner with a slider and the batter gets a bloop single off the end of the bat. Same BABIP but different (in most cases) outcome.

I don't know about doubles and triples. I missed that part of your question. I'd have to look at that. My gut would tell me that they likely regress. But HR rates definitely do not regress to league averages.

[Peter_Spaeth]

Originally Posted by Snowman

I don't know about doubles and triples. I missed that part of your question. I'd have to look at that. My gut would tell me that they likely regress. But HR rates definitely do not regress to league averages.

What we need, if it doesn't already exist, is a slugging average for balls in play stat.

On plain old SLG against, Maddux over his career was some 55 points below the average. That sounds meaningful? Especially since his BA against was 14 points better than average. A non-statistician would conclude from that he was limiting extra base hits pretty well.

[earlywynnfan]

Originally Posted by Snowman

Again, you conflating. The facts I already know are that a pitcher's BABIP regresses to the mean and each pitcher has little to no control over their values. What I was wrong about was that Maddux's values were 9 points below league average. But that still doesn't mean he is able to control his BABIP. If you look up his teammates, they too all beat the league average BABIP. In other words, the ballpark, pitching in the NL, and the defense behind him was responsible for most, if not all, of his ability to beat it.

As far as Kershaw goes, it appears to be the same thing. I just looked up 5 or 6 of his teammates over the years in LA to check their BABIP values. Grienke, Urias, Buehler, Jansen, Baez, all of them are 20 to 40 points below league average BABIP. Again, this means it is their defense, the fact that they all pitch in the NL, and the ballpark that account for the differences, not some magical ability that Kershaw possesses.

Back to the original Grove vs. Koufax line, can you please use your statistics to explain Koufax's widely disparate home vs. away records??

[Snowman]

Originally Posted by earlywynnfan

Back to the original Grove vs. Koufax line, can you please use your statistics to explain Koufax's widely disparate home vs. away records??

His away numbers are worse, for sure, but I don't know that I'd call them "widely disparate". There's an expected disparity for all pitchers when pitching at home and on the road. Part of the "home field advantage" in baseball comes from an umpire's subconscious bias in calling balls and strikes, just like in basketball with fouls. Even when they are trying their best to be neutral, it is somehow still human nature to call the games more favorably for the home team than the away team. The effect is small, but measurable over the course of a career. When you look at Koufax's career Home vs Away numbers, they don't really look all that out of line to me when you consider the fact that he pitched in a pitcher's park. Here's what I see. Note he had almost identical IPs for both. Also, ERA values are much more reliable over the course of a career with 1,000+ IP, so it's fair to look at those in the context of a career of this length, whereas it wouldn't be from season to season.

Home IP: 1158.0
Away IP: 1166.1

ERA Home: 2.48
ERA Away: 3.04

BB/9 Home: 2.9
BB/9 Away: 3.4

K/9 Home: 9.5
K/9 Away: 9.1

WHIP Home: 1.045
WHIP Away: 1.167

HR% Home: 2.2%
HR% Away: 2.1%

BABIP Home: 0.252
BABIP Away: 0.266

When I look at those numbers, the most interesting difference to me is the BB/9 rate. That's a significant gap, and one that definitely has an impact on his WHIP delta. Why was he walking more batters outside of LA? That's not a park effect. Some small disparity exists from umpire subconscious bias as I mentioned, but not that much, I wouldn't think. The differences in BABIP are probalby entirely explainable through park differences and his BB/9 & K/9 rates. I don't think there's much delta attributable to luck over that sample size, and the delta is narrow enough that it is within expectation. There is an expectation also though of a player's general discomfort level when on the road. People just perform better at home. I definitely acknowledge he was better at home than on the road, but I don't see anything that looks wildly out of line with expectations. The BB/9 rate is the most interesting part to me though. Pitching in Dodger stadium definitely helped too.

[Aquarian Sports Cards]

Originally Posted by Snowman

His away numbers are worse, for sure, but I don't know that I'd call them "widely disparate". There's an expected disparity for all pitchers when pitching at home and on the road. Part of the "home field advantage" in baseball comes from an umpire's subconscious bias in calling balls and strikes, just like in basketball with fouls. Even when they are trying their best to be neutral, it is somehow still human nature to call the games more favorably for the home team than the away team. The effect is small, but measurable over the course of a career. When you look at Koufax's career Home vs Away numbers, they don't really look all that out of line to me when you consider the fact that he pitched in a pitcher's park. Here's what I see. Note he had almost identical IPs for both. Also, ERA values are much more reliable over the course of a career with 1,000+ IP, so it's fair to look at those in the context of a career of this length, whereas it wouldn't be from season to season.

Home IP: 1158.0
Away IP: 1166.1

ERA Home: 2.48
ERA Away: 3.04

BB/9 Home: 2.9
BB/9 Away: 3.4

K/9 Home: 9.5
K/9 Away: 9.1

WHIP Home: 1.045
WHIP Away: 1.167

HR% Home: 2.2%
HR% Away: 2.1%

BABIP Home: 0.252
BABIP Away: 0.266

When I look at those numbers, the most interesting difference to me is the BB/9 rate. That's a significant gap, and one that definitely has an impact on his WHIP delta. Why was he walking more batters outside of LA? That's not a park effect. Some small disparity exists from umpire subconscious bias as I mentioned, but not that much, I wouldn't think. The differences in BABIP are probalby entirely explainable through park differences and his BB/9 & K/9 rates. I don't think there's much delta attributable to luck over that sample size, and the delta is narrow enough that it is within expectation. There is an expectation also though of a player's general discomfort level when on the road. People just perform better at home. I definitely acknowledge he was better at home than on the road, but I don't see anything that looks wildly out of line with expectations. The BB/9 rate is the most interesting part to me though. Pitching in Dodger stadium definitely helped too.

If you know you are a little more hittable away from your home park, might you not be inclined to nibble a bit more?

[Peter_Spaeth]

Still interested in your take on Maddux having a slugging average against 55 points below average.

[Snowman]

Originally Posted by Peter_Spaeth

Still interested in your take on Maddux having a slugging average against 55 points below average.

Don't all good pitchers have below average slugging against though?

[Lorewalker]

Originally Posted by Snowman

Part of the "home field advantage" in baseball comes from an umpire's subconscious bias in calling balls and strikes, just like in basketball with fouls. Even when they are trying their best to be neutral, it is somehow still human nature to call the games more favorably for the home team than the away team. The effect is small, but measurable over the course of a career.

I would think an ump's bias would be such a small part, at best, of the home field advantage. There are many factors which would make a player better at home than on the road. Such as not being tired from traveling or from time change, being able to be at their houses and having those comforts of friends, family, pets, familiar surroundings, being before the home crowd and playing at a stadium they spent 1/2 of the season. To only mention ump bias, when there are other factors which obviously influence home field advantage more, is misleading.

[BobC]

Originally Posted by Lorewalker

I would think an ump's bias would be such a small part, at best, of the home field advantage. There are many factors which would make a player better at home than on the road. Such as not being tired from traveling or from time change, being able to be at their houses and having those comforts of friends, family, pets, familiar surroundings, being before the home crowd and playing at a stadium they spent 1/2 of the season. To only mention ump bias, when there are other factors which obviously influence home field advantage more, is misleading.

And if the talk of potentially one day having the balls and strikes called by machine ever comes to pass, that should significantly reduce the effect of whatever umpire bias there might actually be.

[Peter_Spaeth]

Originally Posted by BobC

And if the talk of potentially one day having the balls and strikes called by machine ever comes to pass, that should significantly reduce the effect of whatever umpire bias there might actually be.

The home team that controlled the machine would just program the bias in.

[Tabe]

Originally Posted by Snowman

His away numbers are worse, for sure, but I don't know that I'd call them "widely disparate". There's an expected disparity for all pitchers when pitching at home and on the road. Part of the "home field advantage" in baseball comes from an umpire's subconscious bias in calling balls and strikes, just like in basketball with fouls. Even when they are trying their best to be neutral, it is somehow still human nature to call the games more favorably for the home team than the away team. The effect is small, but measurable over the course of a career. When you look at Koufax's career Home vs Away numbers, they don't really look all that out of line to me when you consider the fact that he pitched in a pitcher's park. Here's what I see. Note he had almost identical IPs for both. Also, ERA values are much more reliable over the course of a career with 1,000+ IP, so it's fair to look at those in the context of a career of this length, whereas it wouldn't be from season to season.

Home IP: 1158.0
Away IP: 1166.1

ERA Home: 2.48
ERA Away: 3.04

BB/9 Home: 2.9
BB/9 Away: 3.4

K/9 Home: 9.5
K/9 Away: 9.1

WHIP Home: 1.045
WHIP Away: 1.167

HR% Home: 2.2%
HR% Away: 2.1%

BABIP Home: 0.252
BABIP Away: 0.266

When I look at those numbers, the most interesting difference to me is the BB/9 rate. That's a significant gap, and one that definitely has an impact on his WHIP delta. Why was he walking more batters outside of LA? That's not a park effect. Some small disparity exists from umpire subconscious bias as I mentioned, but not that much, I wouldn't think. The differences in BABIP are probalby entirely explainable through park differences and his BB/9 & K/9 rates. I don't think there's much delta attributable to luck over that sample size, and the delta is narrow enough that it is within expectation. There is an expectation also though of a player's general discomfort level when on the road. People just perform better at home. I definitely acknowledge he was better at home than on the road, but I don't see anything that looks wildly out of line with expectations. The BB/9 rate is the most interesting part to me though. Pitching in Dodger stadium definitely helped too.

In general, when referring to Koufax's disparate splits, people are usually referring to his numbers from 1962-66, the years that got him in the Hall and in which he pitched in Dodger Stadium. And that's because his splits those years (as a whole - there was one where they were pretty equal) are so different.

However, there's more to it than even that. And that's because in the LA years prior (1958-1961), he pitched in a home park that was absolutely horrendous for lefties. That's why he had a 4.33 ERA there. He also had a 4.04 ERA in Brooklyn. However, it was in those two ballparks where his lack of control was also prominently on display - 1.95 K/BB in LA at the Coliseum and 2.20 at Ebbets.

So what does all that mean? Well, it means that Sandy's unreal numbers at Dodger Stadium overwhelm the 7 years of mediocre (or worse) numbers in his other two home stadiums but raising them up enough to make it LOOK like there wasn't huge splits for him.

[Peter_Spaeth]

Originally Posted by Tabe

In general, when referring to Koufax's disparate splits, people are usually referring to his numbers from 1962-66, the years that got him in the Hall and in which he pitched in Dodger Stadium. And that's because his splits those years (as a whole - there was one where they were pretty equal) are so different.

However, there's more to it than even that. And that's because in the LA years prior (1958-1961), he pitched in a home park that was absolutely horrendous for lefties. That's why he had a 4.33 ERA there. He also had a 4.04 ERA in Brooklyn. However, it was in those two ballparks where his lack of control was also prominently on display - 1.95 K/BB in LA at the Coliseum and 2.20 at Ebbets.

So what does all that mean? Well, it means that Sandy's unreal numbers at Dodger Stadium overwhelm the 7 years of mediocre (or worse) numbers in his other two home stadiums but raising them up enough to make it LOOK like there wasn't huge splits for him.

I believe there's a name for this phenomenon, when aggregate data shows one thing but breaking it down into distinct constituent parts (here, the Coliseum vs. the other home parks) shows something else. Simpson's Paradox maybe? Or it's something like it anyhow.