B-Log. The Ziegler Numbers
47So, while I was working on my Brad Ziegler story, I thought a lot about a theory that Brian Bannister has been working up lately. Banny’s feeling is that, in a very general sense, there are only three types of successful pitchers in the Big Leagues.
Type A: A power, strikeout pitcher who keeps walks down and doesn’t give up too many home runs.
Type B: An extreme ground ball pitcher who keeps walks down.
Type C: A blend of Type A and Type B.
Josh Beckett is a great example of a Type A. In 2007, the year he finished second in the Cy Young voting, his strikeout-to-walk ratio was 4.85 to 1, and he only gave up 17 homers all year. In 2006, on the other hand, Beckett was hammered pretty good. THAT year his strikeout to walk was a much more pedestrian 2.14 to 1, and he gave up 36 homers. And that was the whole difference … his BABIP (batting average on balls hit in play) was actually SIGNIFICANTLY HIGHER in 2007 (.316 in 2007 vs. .270 in 2006) but adding strikeouts, cutting down walks and keeping the ball in the park dropped his ERA almost two full runs.
Brandon Webb is a good example of Type B. Well, he’s actually a freak. Webb is a strikeout pitcher too, which makes him (as his Ziegler Numbers will show) perhaps the best pitcher in baseball. But his unique dominance comes from an incredible lifetime 4.20 to 1 groundball-to-flyball ratio. As Banny says, it sometimes seems like Webb is playing a different game. A better example of the Type B, at least this year, is Mike Mussina, who for years was a Type A guy, a power pitcher, but this year he has not walked anybody (1.21 walks per nine) and he’s getting more ground balls than ever before.
And then a blend example might be someone like James Shields, a guy who gets a good number of strikeouts, a good number of ground balls. really keeps the walks down and beats you without the trumpets blaring in the background. Paul Maholm of the Pirates is another good example — good strikeout to walk, gets more than his share of ground balls.
I was thinking about this as I considered why Brad Ziegler has been so successful. I mean, it doesn’t take a detailed calculation to figure it out — his ground ball to fly ball is 5.18, which is simply off the charts, it’s better than Brandon Webb (who is the gold standard). Because of that sick sink on his pitches, he has not given up a home run, majors or minors, since (get this) 2006. Anyway, he was the one who got me thinking: One way to measure a pitcher’s dominance might be simply to look at two ratios — ground-ball-to-fly-ball and strikeouts-two-walks. Would that give us a list of the best pitchers? So, I multiplied those two numbers, multiplied the whole thing by 10 to give it a nice big feel, and I like the list. The statistic definitely needs some tinkering, but this is a good starting point.
Brad Ziegler, in case you are wondering, scored a 103.6 Ziegler number, which is fabulous. Here, though, are the best pitchers in baseball by the Zieglers:
1. Brandon Webb
Ziegler number: 112.6
Core numbers: 15-4, 3.04 ERA.
Comment: It’s risky, of course, to give any pitcher a long-term contract. But I’m willing to wager that with the way Webb sinks the ball, he will be a dominant pitcher for a long time. At this moment, I’d pick him No. 1 overall.
2. Roy Halladay
Ziegler number: 106.8
Core numbers: 12-8, 2.85 ERA.
Comment: Halladay’s strikeout are up this year, he hardly walks anybody as usual, and he still has that 2 to 1 ground ball ratio. Halladay has been a terrific and hardworking pitcher (his 36 complete games are the most for any pitcher since 2000) but because of injuries and such he has 123 wins, which puts him on pace for that “I’m not sure if he’s a Hall of Famer” purgatory that David Cone, Curt Schilling, Kevin Brown and Luis Tiant probably will share someday.
3. Cliff Lee
Ziegler number: 85.6
Core numbers: 14-2, 2.58 ERA.
Comment: His 6.3 strikeout-to-walk ratio is beyond incredible — it’s hard to believe that Lee’s first full year he walked 81 in 179 innings. I recall that his reputation then was a wild, overpowering lefty. Now, he’s still getting his strikeouts but he has become crafty* — he hardly walks anybody.
*We’ve mentioned here before that only lefties are ever called crafty. It doesn’t seem fair. So I would like to nominate the aged Greg Maddux as the first crafty righty in baseball history. I don’t know, it just seems like it would give meaning to the somewhat unfulfilling final stage of his career.
4. Derek Lowe
Ziegler number: 79.4
Core numbers: 8-9, 3.70 ERA.
Comment: Obviously, a 3.70 ERA is nothing to mock, but I have to tell you … looking at Lowe’s numbers I have no idea why that ERA isn’t even better. He’s an extreme ground ball pitcher, his strikeout to walk is 3.1 to 1, he has given up only 11 home runs all year. The key seems to be that with runners in scoring position, batters are hitting .331 against him. Lowe, like Banny, is just not nearly as good out of the stretch as he is out of the windup.*
*I wonder if anyone has done a study on how much less effective pitchers are out of the stretch than out of the windup. You have to figure that in general they are somewhat less effective or else they would pitch out of the stretch all the time. Right? Zack Greinke, for one, has been considerably BETTER out of the stretch this year. Anyway, I’m sure someone has studied this …
5. Mike Mussina
Ziegler number: 76.5
Core numbers: 13-7, 3.56 ERA.
Comment: Give it up to Moose, who appeared to be absolutely done as an effective pitcher and instead started getting more ground balls and now seems to have a couple more good years left in him, if he wants. He might even have an outside shot at 300 victories.
6. Dan Haren
Ziegler number: 70.9
Core numbers: 11-5, 2.62 ERA
Comment: He’s not really a big ground ball pitcher (1.27 to 1 ratio) but he gets enough sink to keep the ball in the ballpark, he gets a lot of strikeouts, he’s stopped walking anybody, he’s awfully good. The Diamondbacks are 26-37 when either Webb or Haren does not start.
7. Tim Hudson
Ziegler number: 57.5
Core numbers: 11-7, 3.17 ERA
Comment: He’s an extreme ground ball pitcher — not quite in Webb’s league but the next best thing. That’s there for him every year (except 2006, when his ground ball numbers dropped, his homers went up, and his ERA skyrocketed to 4.86). If he stays healthy, I think he will be effective for a few more years.
8. James Shields
Ziegler number: 56.0
Core numbers: 9-7, 3.66 ERA
Comment: One of my favorite pitchers to watch; I’ve been on Tampa all year so obviously I’m not backing off now. I still think with Kazmir, Garza and Shields they’ve got enough pitching to be a real factor in this race until the very end.
9. Andy Pettitte
Ziegler number: 55.4
Core numbers: 12-8, 4.18 ERA
Comment: That’s an inflated ERA for someone with such a good Ziegler … I think it has to do with the .323 BABIP. That is a bit high. In other words, I think Pettitte has been a bit unlucky this year, though of course he has been quite lucky in that the Yankees are scoring 5 runs per game for him. See, it all evens out.
10. Jesse Litsch
Ziegler number: 52.2
Core numbers: 8-7, 4.46 ERA.
Comment: Not sure how Litsch got on this list — he does get ground balls, which is good. And his strikeout to walk is good, which is amazing because he doesn’t strike out anybody (he doesn’t walk anobdy either). This is a glitch in the Ziegler Formula — I need to incorporate strikeouts per nine, I think. Plus Litsch does give up home runs. I’ll work on it. In the meantime, here are three other pitchers who might interest you.
C.C. Sabathia
Ziegler number: 47.5
Comment: He’s not really a ground ball pitcher — a pure Type A guy who is probably the most important player in the NL right now.
Johan Santana
Ziegler number: 38.2
Comment: Lots of people seem to be arguing about whether Santana is REALLY have a disappointing year or if it’s just hype — after all the guy is seventh in the league in ERA. But here’s something that is true: His strikeout to walk numbers are way, way down — that seems to be what is keeping him from being dominant.
Brian Bannister
Ziegler number: 20.5
Comment: Yeah, that’s not a great Ziegler. But things are looking up. Really. In July, he struck out about a batter per inning, he got a few more ground balls, from the rose-tinted view of Banny Log, a lot of stuff started falling into place as he tries to reinivent himself a little bit. True, what most will notice is that he did have a 7.20 ERA for the month, but that was because he had a ridiculous .392 BABIP … that won’t last. Things will get better in August. I feel it. Although it might be tough to do Banny Logs from Beijing. I’m not sure if they even allow that over there. We’ll have to see how that goes.
Loading ...
Are Banny Logs more subsersive to the revolutionary authoritarian slave labor Marxists than joechats?
Very interesting post. I’m curious though. Where did Chien-Ming Wang rank on this list for the last two years? He doesn’t strike out many, but he’s a ground ball machine with a 2.15 BB/9 in 2006, and a 2.66 BB/9 in 2007.
Also, where do you find GB/FB ratios?
Next step, divide by HR/9IP (not in order):
Webb: 228
Halladay: 165.5
Lee 199.5
Lowe 110
Mussina 76
Haren 105
Hudson 82
Shields 53.3
Pettitte 61.5
Litsch 39.2
Santana 43.6
Bannister 18
Rich,
G/F ratio can be found on ESPN’s pitching stat page under “expanded II”
Joe,
I think Greg Maddux comes out ahead of Litsch (and for that matter Pettitte and Shields) with a 56.8 “Ziegler”
Wang has a Ziegler number 33.7, but he almost never gives up a HR, so if divided by HR/9 his number is 89.
Kevin and Wally - Thank you. Much obliged.
If there’s no study, at least you could use PitchFX data to see if a pitcher’s stuff alters from the stretch or not.
Hmm.
Well here is how the Cubs Aces fare..
Zambrano- 32.86 which seems low…
Harden- 23.79 which seems REALLY low. He has a nice 3.66 K/BB ratio, but really hurts in the G/F with a .65.
Some interesting results
Tim Hudson 53.25
Greg maddux 56.8
Relievers seem to do very well..
Rivera has a Zeigler number of 172.42
Other interesting results..
Jon Rauch- 46.07
Trevor Hoffman 46.27
A couple of tweaks:
1) We need to play with the multiplier, but we need to do it last. What do we want the outcome to be? I think that keeping 100 a mark of greatness works well. So, in it the current form, it should stay ten. But if we Wally’s idea of incorporating HR/9, we’d need to change it. (When I was teaching, and figuring out how I wanted to adjust for test difficulty, I’d focus on the second highest score. The highest
2) Perhaps the multiplier should different for starters and relievers? We’d have to look at raw Ziegler numbers for large groups of both to figure this out. I think that we should consider, though. I understand that this is a rate stat, and we’re not going to turn it into a value stat stat. But there’s such a difference between starters and relievers.
I also heard Mussina being referred to as “crafty” as well. That makes 2 righties as the all non-lefty crafty team.
Interesting, although it seems to me that Banny’s theory leaves out a group of quite successful historical pitchers, what Bill James once called the “Robin Roberts family” of pitchers — moderate strikeout levels, very low walk rates, give up a lot of home runs. In a big park (Oakland, Petco, Shea) I would htink this type of pitcher could still be successful, although certainly harder than it was back when only 3 or 4 people in most lineups could reach the seats very often.
2008 Lowe from stretch:
5.52 K/9, 2.65 BB/9, .82 HR/9, 4.04 FIP
2008 Lowe from windup:
7.24/1.95/.65, 3.18 FIP
2008 NL from stretch
6.30/3.46/1.02, 4.42
2008 NL from windup
7.38/3.48/1.07, 4.26
Lowe’s difference is pretty far from the average.
his career numbers however (3.91 FIP from stretch, 3.74 FIP from windup), dont suggest that this should continue
On crafty righties, I took all qualified starters from 08 and added up their fastball percentage and cutter percentage (via Fangraphs), then multiplied that by their average speed for those pitches to create “Craftyness Index” (CI). I took Wakefield out of the equation
League Leaders in CI, 2008
Matt Morris, R, 41.2
Bronson Arroyo, R, 41.3
John Smoltz, R, 43.0
Tom Glavine, L, 43.1
Mike Mussina, R, 44.7
Barry Zito, L, 46.8
Gil Meche, R, 48.0
Ted Lilly, L, 48.1
Paul Byrd, R, 49.0
Jamie Moyer, L, 49.5
hmmm, 4 Lefties and 6 Righties.
So it seems like Banny’s saying (with my own touch) that something like this is the way to guarantee success:
Type A: Flyball pitchers (>40%) with high K rates ( > 8 K/9, 2.0 K/BB Min)
Type B: Groundball pitchers (>50%) with low BB rates ( 1.25 GB/FB, 3.0 K/BB Min)
Interesting. I’d be curious to see how many teams have guys that meet this criteria and what their standings are. Maybe you could come up with a list to see which teams have the most on their staff.
My last post didn’t go through right. Here’s what I came up with:
Type A: Flyball pitchers (>40%) with high K rates ( > 8 K/9, 2.0 K/BB Min)
Type B: Groundball pitchers (>50%) with low BB rates ( 1.25 GB/FB, 3.0 K/BB Min)
Let’s try this again
Type B: Groundball pitchers (>50%) with low BB rates ( 1.25 GB/FB, 3.0 K/BB Min)
Type C: GB/FB mix with high K/BB ratios ( >1.25 GB/FB, 3.0 K/BB Min)
Type B: Groundball pitchers (>50%) with low BB rates ( < 3.0 BB/9, 2.0 K/BB Min)
Sorry. The comments wasn’t going through right.
About the stretch/wind-up question. Dusty Baker agrees with you about there being something about it- http://www.firejoemorgan.com/2008/07/dusty-baker-loves-obp.html
FJM isn’t sure what to think.
You know, Brian Bannister has to be the most frustrating pitcher. Just when I think he’s a very good starter… he pulls off a string of bad starts. Then, just when I think he’s overrated… he pulls off 2-3 surprisingly good starts that give me hope…. before pulling off some sting of surprisingly bad starts again. Now he’s just thrown consecutive decent starts again. I’m ready to pull my hair out. I just don’t know what to think of this guy anymore.
You know Devon,
I believe that is what is eventually called “league average”
Pitching from the stretch:
Andy Dolphin found in The Book that pitchers tend to do worse out of the stretch to the tune of .008 points of OBP. It’s unclear how much of that is due to fielder positioning (for example, the wider hole on the right side when the first baseman is holding the runner). That translates into something like 0.23 points of ERA. It will vary from pitcher to pitcher, but you need a huge sample size before concluding that any particular guy is far off from the league average stretch penalty.
Well, Hudson is getting TJ surgery, so so much for that…
What if we tinker with the equation so that league-average is closer to 100? I just think that’s a benchmark that people are comfortable with. For example, how about:
(K/BB ratio) x (GB %) x 10
Using league averages to date that would be:
2.03 x .44 x 10 = 89
Using that metric, 100 is a little more than 10% above league average; a good, solid season and an easy-to-grasp benchmark. Here’s this season’s top ten under this equation:
Lee - 301
Halladay - 279
Haren - 254
Mussina - 238
Shields - 213
Webb - 212
Beckett - 186
Lowe - 182
Oswalt - 162
Pettitte - 158
8 of the 10 are also on Joe’s list. Hudson and Litsch fall out, Beckett and Oswalt climb in.
Of the 96 qualified starters in both leagues, 47 score 100 or better (league average falls to 89 by including non-qualified pitchers). If half are over 100 and half are below that feels pretty good to me.
6 pitchers scored 200+, 13 scored 150+. The lowest-rated qualified pitcher in baseball by this equation is Barry Zito at 41.
Some other guys:
Sabathia - 152
Maddux - 140
Lincecum - 138
J. Santana - 136
Volquez - 115
Zambrano - 101
Duchsherer - 100
Joe Blanton - 83
Bannister - 80
Verlander - 79
Could someone do the numbers for Felix Hernandez?
My home/road splits:
7-1 home record
2-6 road record
2.16 home ERA
5.93 away ERA
0.88 home WHIP
1.40 away WHIP
So should Ziegler move into the rotation?
Mr. Posnanski,
I am a charter member of THE JOSÉ GUILLEN SOCIETY, a group of modest notoriety on Geoff Baker’s Seattle Mariners blog in the Seattle Times. Our SOCIETY is very aware of Señor Guillen’s temperamental nature, and believe that this same trait that has become a detriment for the Royals, was an integral part of the M’s success in 2007, one team’s junk being another team’s treasure). Consequently, THE JOSÉ GUILLEN SOCIETY is lobbying the front offices of the Mariners and Royals through these public blogs to enact a simple trade.
THE JOSÉ GUILLEN SOCIETY believes that it would be in both organizations best interest if:
1). The M’s re-acquire José Guillen by assuming his ‘09 and 2010 contract, OR
2). The M’s trade workhorse pitcher, Carlos Silva, straight up for José Guillen. The next three years of Silva at $12m per year for the next two years of Guillen at $10m per year.
Win/Win
Respectfully,
THE JOSÉ GUILLEN SOCIETY
Two points:
1. Please, please, please come up with some word other than ‘dominant’ to describe good but not overwhelming performance. It’s really both tiresome and wasteful of an otherwise good word.
2. Right-handers have been known as ‘crafty’ for years, but only if they throw a knuckleball — I know I’ve heard both Tim Wakefield and Charlie Hough referred to as ‘crafty’, and I’d be surprised if Phil Niekro and Wilbur Wood didn’t get their fair share of the term as well in their day.
Jose Guillen to Seattle for Silva. Make it happen!
i really like mikey’s metric up there. it seems to accurately portray the best pitchers in the league quite well. i’d be curious to see how some of the all-time greatest pitching seasons fare using that formula
Speaking of pitching out of the stretch (off topic I know), an interesting point I’d like to make is that during the late 90’s and into 2001 when Joe Kerrigan was the pitching coach for the Red Sox, Sox pitchers rarely, if ever, pitched out of the stretch. There was no attempt to hold runners. I’ve always wondered how much that contributed to the staff’s production. I recall that they had the best pitching staff in the league during that period of time. Just a little anecdotal evidence regarding the efficacy of not pitching out of the stretch.
I like this stat. It’s simple and all of the numbers are in the pitcher’s control and largely independent of park, teamates, etc.
I don’t like multiplying the two, however. When you multiply two ratios you give a bonus to the guys who are average in both instead of really good at one and not so good at the other, i.e. Type C guys end up ranked higher.
For instance, two pitchers A and B. Let’s say pitcher A is a real strikeout pitcher with perfect control, but gives up flyballs and ground balls at about the same rate, while pitcher B is about average at both.
K/BB GB/FB
A 9 1
B 5 5
If we multiply the ratios, the “Ziegler Numbers” look like this:
A 9
B 25
Do we really think pitcher B is almost 3X better than pitcher A?
This number is more like OPS. We appreciate OBP as much as we do slugging and so we add them together.
Adding the two ratios gives both pitchers a score of 10 and says they’re roughly equal. Call this the Z+
The only problem with adding is that the two ratios are on slightly different scales, e.g the average GB/FB ratio is higher than the average K/BB ratio and so our score would be biased toward groundball pitchers over strikeout pitchers. (though it doesn’t seem to bother us that this same thing happens in OPS).
To account for the bias we could normalize our ratios before we start, by dividing all the K/BB ratios by the highest K/BB ratio and all the GB/FB ratios by the highest GB/FB ratio (looking at only pitchers with at least 30 ip in order to eliminate statistical outliers). Call this the Z+adj.
So the top 10 (w/ >30 ip, closers included) looks like this:
Ziegler# Z+ Z+adj
Mariano Rivera 150.8 129.0 128.6
Jonathan Papelbon 113.6 94.2 100.2
Brandon Webb 112.6 67.1 103.2
Roy Halladay 106.8 71.6 90.1
Matt Thornton 106.5 66.9 92.7
Cliff Lee 85.6 76.2 84.1
Brad Ziegler 85.4 64.2 116.2
Hong-Chih Kuo 83.4 68.1 80.2
Chad Bradford 81.8 62.6 113.0
Derek Lowe 79.4 56.6 83.2
And here’s what happens to each pitchers ranking using the two modified stats.
Mariano Rivera 00
Jonathan Papelbon 0 -3
Brandon Webb -5 -1
Roy Halladay 0 -6
Matt Thornton -4 -2
Cliff Lee +3 -7
Brad Ziegler -7 +5
Hong-Chih Kuo +1 -10
Chad Bradford -7 +6
Derek Lowe -13 -4
BaH….who knew “tab” + “space” would post the comment.
I wasn’t quite finished.
The new top ten by Z+adj:
Mariano Rivera 128.6
Brad Ziegler 116.2
Chad Bradford 113.0
Brandon Webb 103.2
Jonathan Papelbon 100.2
Sean Green 98.7
Matt Thornton 92.7
Joe Smith 91.6
Roy Corcoran 91.1
Roy Halladay 90.1
I have to admit, I’m not particularly fond of this list as it seems to have provided too much benefit to extreme groundball pitchers. Perhaps I should have normalized by league average ratios rather than maximums in each category. (The problem with normalizing by averages is that it spreads out the data and really accentuates the differences between the leaders and everyone else, so you end up with the best having crazy high numbers, while there’s little distinction between all the average guys.)
Okay…this one combines normalization by league average with Mikey’s idea, so league average is by definition 100:
{([K/BB]-[League Avg K/BB])/[League Avg K/BB]+([GB/FB]-[League Avg GB/FB])/[League Avg GB/FB]}*10+100
……………………………..K/BB…GB/FB….Ziegler#….Normalized
Mariano Rivera…………..11.6…….1.3…….150.8…….140.5
Jonathan Papelbon………8.0…….1.42……113.6…….125.6
Brandon Webb……………3.33…..3.38……112.6…….119.7
Roy Halladay……………..5.04…..2.12……106.8…….117.8
Matt Thornton…………….4.08…..2.61……106.5…….117.2
Cliff Lee ………………6.25…..1.37……..85.6…….117.5
Brad Ziegler……………….1.88…..4.54……..85.4…….121.9
Hong-Chih Kuo…………..5.21……1.6………83.4…….114.7
Chad Bradford……………1.86…….4.4……..81.8…….120.8
Derek Lowe……………….3.09…..2.57…….79.4……..112.6
Not a lot of movement in the top 10. Ziegler and Bradford move up a few notches. Halladay, Thorton and Lee drop down a few notches (though Lee is now much much closer to Halladay and Thornton instead of 20 points back.)
Some of the more interesting effects occur outside the top ten. Fausto Carmona moves up 168 places (from #202 to #34) with my formula, Bob Howry moves up 76 and Rich Harden moves up 62 to just below league average while Scot Shields and Andy Pettitte each fall about 15 places, Joakim Soria moves up 11 from #30 to #19.
I don’t like the idea of using HR/9 because GB/FB already partially accounts for tendency to give up home runs and the other part is mostly park and opponent effects. (Unless…does anyone know if there’s such a thing as an extreme pop-up pitcher? I suspect these are the same guys that are good at getting ground balls, i.e. pitchers with a lot of movement that prevent batters from squaring up on the pitch.)
I also don’t like the idea of using HR/9 for just the reasons listed by Rocketman.
What I find really intriguing about this exercise is that we may be hitting on a formula that’s predictive of major league success.
Maybe I’m being totally naive, but it seems like K/B ratio and GB % or GB/FB ratio in AA or AAA could be fairly good predictors.
I mean, we don’t really need a new stat to tell us that Webb and Halladay are two of the best pitchers in baseball. But if you have something that separates two prospects whose core numbers are similar that could be really interesting.
One reason Derek Lowe’s ERA is not lower may well be having having the likes of Nomar at ss and Kent at 2b.
Another great column in the Star today, by the way.
Re: Derek Lowe. I have watched or listened to almost every game Lowe has pitched since coming to the Dodgers, and here’s what’s wrong with him: he holds grudges.
Faster than any other pitcher I’ve ever seen, as soon as Lowe has some kind of bad play behind him, no matter the inning, he’s likely to lose it. If it happens with nobody out, then he gives up (say) one unearned run and some earned runs. If it happens with two outs, he only gives up unearned runs. It affects his W/L, his runs allowed, his earned runs allowed.
In Bob Hamman’s auotbiography (he is arguably the best bridge player alive, winner of multiple world championships) he wrote that his bridge game really improved when he stopped worrying about what went wrong on a hand until after the session of bridge was over, and just focused on the next hand. In baseball this is even easier; every game is taped, there are coaches to remind you what you did wrong, and you get to regroup between innings. Lowe just lacks that ability. I bet if you did a study showing the number of runs scored in innings with a defensive error, Lowe would come out as one of the worst pitchers in baseball. And if I were Russell Martin (or even Joe Torre) after each error I would go out to the mound, tell Lowe to regroup, to pretend that error was a bloop single, and get the next guy out. Because when Lowe doesn’t have errors behind him, he’s pretty darned good.
“Not sure how Litsch got on this list — he does get ground balls, which is good. And his strikeout to walk is good, which is amazing because he doesn’t strike out anybody (he doesn’t walk anobdy either). This is a glitch in the Ziegler Formula — I need to incorporate strikeouts per nine, I think.”
Instead of multiplying by 10 to get nice big numbers, how does it look if you multiply by K/9?
No relievers? I think Mo’s Ziegler number would be ridiculous this year. I’d calculate it if I could find his gb/fb ratio somewhere. His K/BB ratio is a videogame’esque 58:5.
And no, no, HELL NO on the crafty righty. That’s a lefty honor. Righties are “warriors.”
Lefties are “crafty.”
“So the top 10 (w/ >30 ip, closers included) looks like this:
Ziegler# Z+ Z+adj”
Rocketman, you are a brilliant, brilliant individual. bravo!
I found this site via Baseball Musings, and thought I’d contribute.
Rocketman: I like the standardization approach, because we’re unsure of the scales of K/BB and GF/FB, and you’re explanation of HR/9 is spot on. However, it should be noted that OBP*SLG is generally considered a better measure than OPS, and is the basis for Bill James’ original runs created formula (OBP*SLG*AB, since revised).
As an alternative to standardization, you might consider treating K/9 and BB/9 separately. Walks may have a greater effect on pitchers who have lower GB/FB ratios, so treating individuals with 12 K/9 and 4 BB/9 as equal to those with 6 K/9 and 2 BB/9 may not be appropriate.
Lastly, simply subtracting then dividing by the league average value may not give you the results you’re looking for if there are large differences in the variance of the two stats. If you’re looking for standardization in the statistical sense, subtract the mean and divide by the standard deviation to equate the scales.
Good work from everyone involved.
Mikey,
I like the idea. However, if “89″ is the league avg., and you’ve got all these guys up there in the 100s and 200s, and Barry Zito is down there only, something is wrong with your math if you think “100″ is 10% above avg. Pitching stats–like pretty much everything else in baseball–are, however you choose to mix, match, and interpret them, still normally distributed, which is pretty beautiful if you think about it. Stats like OPS and ERA (+) work because, if you look at the data points, they are essentially symmetrical, all the way out to the outliers. Your graph would belie the distributive consistency of baseball, which leaves me somewhat skeptical. I suggest you give some more thought to your scale and recalibrate.
Rocketman,
Nice work. I think an extreme pop-up pitcher would be a power pitcher. Like how people say a fastball “rises”. To generate a pop-up you have to do the opposite of a ground ball, get a pitch to “rise”. I think I would be in favor of incorporating HR/FB and IFFB%. The problem is that HR/FB Does not vary all that much so it really would not have that great of an impact on the standings. The problem with IFFB% is that it dramatically favors relievers.
I think the numbers need to bias towards the groundball pitchers. Think about it, lets say a pitcher averages 7 innings (21 outs). The best pitcher will only have 7.8 strikeouts (average is 5). The rest are either line outs, ground balls, or fly balls. Line drives go for hits an absurd percent of the time so lets forget about that. A fly ball can leave the park so favor the most common out, ground balls. On average 16 “other” outs need to be made, that is more than three times the number of Ks. Of course this would not apply to relievers, as a strikeout is often needed by them.
I’m done rambling, back to work.
Re: Pitching out of the stretch
I think it would be interesting to correlate SB, CS, CS% and handidness with how well someone pitched out of the stretch. For instance Maddux notoriously holds runners poorly. This is largely a choice he has made. I suspect his performance pitching from the stretch is closer to his performance pitching from the windup than most pitchers. Some righty who holds runners well might have a larger deviation because he pays so much attention to the runners. I think leftys and righties ought to be seperated from each other fof the purpose of such a study.
It’s not the same but Baseball Prospectus has a metric that uses a pitcher’s K-rate, BB-rate, and GB/FB ratio to come up with what they call “QERA,” which estimates what a pitcher’s ERA should be based on those three things (things that a pitcher has the most control over).
From BP’s glossary:
QuikERA (QERA), which estimates what a pitcher’s ERA should be based solely on his strikeout rate, walk rate, and GB/FB ratio. These three components–K rate, BB rate, GB/FB–stabilize very quickly, and they have the strongest predictive relationship with a pitcher’s ERA going forward. What’s more, they are not very dependent on park effects, allowing us to make reasonable comparisons of pitchers across different teams.
The formula for QERA is as follows:
QERA =(2.69+K%*(-3.4)+BB%*3.88+GB%*(-0.66))^2
Note that everything ends up expressed in terms of percentages: strikeouts per opponent plate appearance, walks per opponent plate appearance, and groundballs as a percentage of all balls hit into play. Andy Pettitte, for example, has a 19.6% K rate, a 7.9% BB rate, and a 62.7% GB rate, giving him a QERA of 3.68. Note further that QERA is exponential, which is appropriate since run scoring is not linear.
Dru, when Kerrigan was with the Red Sox, he wanted his pitchers to not use the slide step– they always pitched from the stretch with runners on base, just didn’t do the slide-step motion.
It was a good idea because the AL at that time was in a huge offensive era, so the other team risking outs by trying to steal bases might help you in the long run. In one-run games, it might not be such a great idea. And in a lower run scoring environment it might not work so well.
I wonder if there would be certain situtations where a pitcher might be better off pitching from the windup, even with a runner on first– say a pitcher who is just way better from the windup has two out and a very slow runner on first.
Some runners are so slow and seem to pay so little attention, you might be able to get away with it without giving up a steal every time. If a pitcher’s much worse in the stretch, maybe it’d be worth it– a pitcher like, say Clay Buchholz lately?