Most Important Pitching Stat?

[Bleacherbee]

I believe its WHIP, everything else seems either irrelevant in the grand scheme of things, or too dependent upon the team's defense. I get real tired of seeing ERA+ thrown around as the end-all of statistics when it comes to pitching greatness.

Thoughts?

[leecemark]

--The problem with WHIP is it measures a single as equal to a HR allowed. It isn't exactly a mainstream stat, but I think OPS against is more usefull.

[ElHalo]

Holds, definitely holds.

[Bleacherbee]

--The problem with WHIP is it measures a single as equal to a HR allowed. It isn't exactly a mainstream stat, but I think OPS against is more usefull.

However, wouldn't WHIP be uniform, regardless of team defense or quality of hitter that the pitcher faces?

[ElHalo]

However, wouldn't WHIP be uniform, regardless of team defense or quality of hitter that the pitcher faces?

Uh, of course not. Hits are totally dependant on defense, and how good the hitters are that a pitcher faces.

[Bleacherbee]

Uh, of course not. Hits are totally dependant on defense, and how good the hitters are that a pitcher faces.

But wouldn't quality of hits be more dependant?

[leecemark]

--Pitchers have control over strike outs and walks and alot of control over HRs. Most pitchers have a limited ability to control what happens once a ball is put in play. I'd say K:BB and HR rate are both more important than WHIP.
-- Voted ERA+ for this poll though. If you can only know one stat about a pitcher thats my choice. Of course, its a combination of stats that really needed to evaluate pitchers.

[SABR Matt]

Can I vote for none of the above? they all suck as stats go...sorry.

[leecemark]

--Matt, I see you voted for K:9. Would K:BB have gotten your vote over that given the option?

[Bleacherbee]

Yes, Matt, I am very interested in what your end-all stat for pitchers would be...

[SABR Matt]

yes...K/BB would have been better than K/9

What I use is a variation on the idea of Voros McCracken's defense independent pitching statistics...he turned it into a rough estimate of DIPS ERA...I don't think that works becasue you're making needless assumptions about how defense independent performance turns into runs...but what I do is take an exact count of the defense independent (linear weighted) bases a pitcher allowed and then graft a neutral team defense onto that (the league average rate at which defense dependent events like singles, doubles and triples occur per out) and factor in the impact the pitcher did have on the in play events...and combine that into a defense independent bases per out score. That seems to work pretty darned well and makes logical sense...at least to me anyway. :)

The advantages thta has over other metrics:

1) it's not dependent on any element not in the control of the pitcher himself

2) I makes no assumptions...it's a stats based only on data that actuall exists...there's no straight multipliers like in DIPS ERA.

3) It's as all inclusive as possible for historical data.

Defense Independent Events: HR, BB, K, HBP, WP, and BK. Defense dependent events: Everything else. The HR data is park adjusted.

[wrgptfan]

For a single year for a starting pitcher, I like to use a combination of ERA+ and IP compared to the average IP of the top 10 pitchers (in IP) in the league. The formula is very simple:

Value = ERA+ * (IP / Avg IP of Top 10 Pitchers)

Here are the top 100 seasons since 1893:

Avg IP of
Top 10
Pitcher Year W L IP Pitchers ERA ERA+ Value
--------------------------------------------------------------------------
Walter Johnson 1912 33 12 369.0 265.7 1.39 239.9 333.2
Greg Maddux 1994 16 6 202.0 180.2 1.56 273.3 306.3
Walter Johnson 1913 36 7 346.0 310.7 1.14 258.5 287.9
Bob Gibson 1968 22 9 304.7 295.7 1.12 258.0 265.8
Pedro Martinez 2000 18 6 217.0 238.7 1.74 285.4 259.4
Pete Alexander 1915 31 10 376.3 330.3 1.22 224.9 256.2
Greg Maddux 1995 19 2 209.7 214.0 1.63 259.1 253.9
Roger Clemens 1997 21 7 264.0 249.3 2.05 225.8 239.1
Walter Johnson 1918 23 13 326.0 292.4 1.27 213.6 238.1
Joe Wood 1912 34 5 344.0 265.7 1.91 179.9 232.9
Dwight Gooden 1985 24 4 276.7 271.3 1.53 225.9 230.4
Lefty Grove 1931 31 4 288.7 274.8 2.06 219.2 230.3
Cy Young 1901 33 10 371.3 349.4 1.62 216.5 230.0
Hal Newhouser 1945 25 9 313.3 266.9 1.81 195.2 229.2
Christy Mathewson 1905 31 9 338.7 341.0 1.28 229.8 228.2
Ed Walsh 1912 27 17 393.0 265.7 2.15 148.7 220.0
Pedro Martinez 1999 23 4 213.3 239.7 2.07 245.1 218.1
Ed Walsh 1908 40 15 464.0 350.0 1.42 163.5 216.7
Ed Walsh 1910 18 20 369.7 323.2 1.27 188.8 216.0
Amos Rusie 1894 36 13 444.0 391.4 2.78 189.4 214.9
Pedro Martinez 1997 17 8 241.3 249.3 1.90 221.2 214.2
Mordecai Brown 1906 26 6 277.3 328.6 1.04 252.8 213.3
Sandy Koufax 1966 27 9 323.0 287.7 1.73 189.9 213.2
Walter Johnson 1919 20 14 290.3 292.0 1.49 214.1 212.9
Mordecai Brown 1909 27 9 342.7 311.4 1.31 193.3 212.7
Walter Johnson 1910 25 17 370.0 323.2 1.36 182.9 209.4
Randy Johnson 2002 24 5 260.0 239.3 2.32 189.8 206.3
Dizzy Trout 1944 27 14 352.3 288.6 2.12 168.3 205.4
Dolf Luque 1923 27 8 322.0 315.7 1.93 200.9 204.9
Wilbur Wood 1971 22 13 334.0 307.4 1.91 188.5 204.8
Carl Hubbell 1933 23 12 308.7 290.6 1.66 192.7 204.6
Pete Alexander 1916 33 12 389.0 326.6 1.55 170.2 202.7
Steve Carlton 1972 27 10 346.3 311.1 1.97 182.0 202.6
Randy Johnson 1999 17 9 271.7 239.7 2.48 178.3 202.1
Kevin Brown 1996 17 11 233.0 247.5 1.89 214.2 201.7
Bob Feller 1946 26 15 371.3 279.6 2.18 151.0 200.5
Roger Clemens 1990 21 6 228.3 241.2 1.93 210.9 199.7
Mort Cooper 1942 22 7 278.7 269.0 1.78 192.1 199.0
Jack Coombs 1910 31 9 353.0 323.2 1.30 181.9 198.7
Ron Guidry 1978 25 3 273.7 287.6 1.74 208.1 198.0
Hal Newhouser 1946 26 9 292.7 279.6 1.94 188.1 196.9
Dean Chance 1964 20 9 278.3 283.0 1.65 200.0 196.7
Randy Johnson 1995 18 2 214.3 214.0 2.48 196.3 196.6
Christy Mathewson 1909 25 6 275.3 311.4 1.14 221.9 196.2
Lefty Gomez 1937 21 11 278.3 271.1 2.33 190.6 195.7
Greg Maddux 1998 18 9 251.0 245.5 2.22 190.6 194.8
Robin Roberts 1953 23 16 346.7 271.3 2.75 152.5 194.8
Walter Johnson 1915 27 13 336.7 330.3 1.55 191.0 194.7
Randy Johnson 2001 21 6 249.7 237.4 2.49 184.1 193.6
Bucky Walters 1939 27 11 319.0 279.6 2.29 167.9 191.6
Pete Alexander 1920 27 14 363.3 318.6 1.91 167.5 191.1
Addie Joss 1908 24 11 325.0 350.0 1.16 205.2 190.5
Lefty Grove 1930 28 5 291.0 282.7 2.54 185.1 190.5
Christy Mathewson 1908 37 11 390.7 350.0 1.43 167.5 186.9
Jack Taylor 1902 23 11 324.7 352.1 1.33 202.5 186.7
Thornton Lee 1941 22 11 300.3 280.0 2.37 173.4 186.0
Vida Blue 1971 24 8 312.0 307.4 1.82 183.2 185.9
Dazzy Vance 1928 22 10 280.3 288.3 2.09 190.7 185.4
Dazzy Vance 1924 28 6 308.3 288.6 2.16 173.4 185.2
Gaylord Perry 1972 24 16 342.7 311.1 1.92 168.1 185.1
Christy Mathewson 1912 23 12 310.0 265.7 2.12 158.7 185.1
John Tudor 1985 21 8 275.0 271.3 1.93 182.6 185.1
Kid Nichols 1897 31 11 368.0 337.3 2.64 169.6 185.0
Randy Johnson 2000 19 7 248.7 238.7 2.64 177.3 184.7
Bret Saberhagen 1989 23 6 262.3 253.5 2.16 178.3 184.5
Dennis Eckersley 1990 4 2 73.3 241.2 0.61 606.4 184.4
Kid Nichols 1898 31 12 388.0 362.9 2.13 172.2 184.1
Eddie Cicotte 1917 28 12 346.7 327.3 1.53 173.7 184.0
Eddie Cicotte 1919 29 7 306.7 292.0 1.82 175.1 183.9
Ed Walsh 1907 24 18 422.3 343.6 1.60 149.4 183.6
Warren Spahn 1953 23 7 265.7 271.3 2.10 187.5 183.6
Claude Hendrix 1914 29 10 362.0 344.0 1.69 174.4 183.5
Spud Chandler 1943 20 4 253.0 272.8 1.64 197.4 183.0
Dutch Leonard 1914 19 5 224.7 344.0 0.96 278.7 182.0
Jim Palmer 1975 23 11 323.0 299.4 2.09 168.6 181.9
Pete Alexander 1917 30 13 388.0 327.3 1.83 153.4 181.9
Carl Hubbell 1934 21 12 313.0 288.9 2.30 167.7 181.7
Sandy Koufax 1965 26 8 335.7 295.7 2.04 159.8 181.4
Cy Young 1902 32 11 384.7 352.1 2.15 165.9 181.2
Bob Feller 1940 27 11 320.3 286.3 2.61 160.9 180.1
Tom Seaver 1971 20 10 286.3 307.4 1.76 193.2 179.9
Greg Maddux 1993 20 10 267.0 254.7 2.36 171.5 179.8
Red Faber 1921 25 15 330.7 315.1 2.48 171.2 179.6
Joe McGinnity 1904 35 8 408.0 385.8 1.61 169.3 179.1
Roger Clemens 1991 18 10 271.3 249.1 2.62 164.4 179.1
Randy Johnson 2004 16 14 245.7 234.8 2.60 170.8 178.8
Ed Reulbach 1905 18 14 291.7 341.0 1.42 208.6 178.4
Warren Spahn 1947 21 10 289.7 272.3 2.33 167.5 178.2
Greg Maddux 1997 19 4 232.7 249.3 2.20 190.8 178.1
Steve Carlton 1980 24 9 304.0 277.6 2.34 162.0 177.4
Walter Johnson 1914 28 18 371.7 344.0 1.72 163.8 177.0
Pat Hentgen 1996 20 10 265.7 247.5 3.22 164.7 176.8
Juan Marichal 1966 25 6 307.3 287.7 2.23 165.3 176.5
Carl Hubbell 1936 26 6 304.0 290.6 2.31 168.7 176.5
Johan Santana 2004 20 6 228.0 234.8 2.61 181.5 176.3
Amos Rusie 1893 33 21 482.0 395.2 3.23 144.3 176.0
Denny McLain 1968 31 6 336.0 295.7 1.96 153.9 174.9
Jack Chesbro 1904 41 12 454.7 385.8 1.82 148.3 174.8
Stan Coveleski 1918 22 13 311.0 292.4 1.82 164.0 174.4
Sandy Koufax 1963 25 5 311.0 287.0 1.88 160.8 174.2

[SABR Matt]

Interesting list...

Glad to see you're taking playing time into account....a lot of people look past that.

The main problem I have is that it's based on ERA+ which will bias the list in favor of pitchers who pitched for good teams with good defenses...but as a quick comparison tool, it's very interesting.

[leecemark]

--I think ERA+ and IP give you a very good rough estimate on who the better pitchers are. If one pitcher is clearly better in those departments, thats good enough for me. Once you've got the rough sort from those two figures, you've got a smaller pool from which to examine the numbers closer and see who is better or worse than those numbers suggest. No one stat gives the complete picture.

[frnkbndy1981]

I invented a pitching metric that isn't the best, but helps us understand pitching better.

It is called In Play Hits Prevented.

It turns out that my collegue, SABR Matt, had actually already stumbled upon the logic I used to create this stat before me...so I had this help on this article.

But, the interesting discussion thus far in thread should definitely include this metric.

Check it out:

http://athomeplate.com/sb2305.shtml

[leecemark]

--Frank, that is an interesting analysis, but I'm not sure that looking at only one season really tells you much about a pitchers ability to induce "bad contact" . The career data on Maddux shows pretty clearly that he had that ability. On the other hand, somebody like Jimmy Gobble may be simple lucky. He just happened to have more balls hit at somebody than you would expect.
--The whole premise of the theory that pitchers don't control what happens once a ball is in play is that most pitchers aren't able to repeat low BIP averages. There are some pitchers, such as Maddux, who have demonstrated an ability to deliver lower than normal BIP averages over the course of their careers. In fact, most really good pitchers show that ability to some degree.
--For most guys who parlay a low BIP average into a big season the odds are against repeating it. If you're a GM you want to deal that guy while his value is inflated. If you're a fantasy owner you don't want to pay for that guy the next year. Conversely, bargains are available on guys who had unusally bad luck with BIP average (unless they have proven that is part of their profile over time).

[frnkbndy1981]

Thanks for reading.

I stated in the article that I am going to do many more years of research on this stat then just one.

I actuallyt think that this stat does show that some pitchers, though rare, can prevent In-Play-Hits.

I was doing this analysis hardcore, and will continue to do so, but right now I am sidetracked by this little project of mine that analyzes all the playoff teams of the last forty years to see what wins baseball.

Thanks for reading the article.

[baseball79]

If a guy won 20 games and had a 5.01 ERA and and another guy won 16 games with a 2.98 ERA, who would you pick?

[frnkbndy1981]

The one with the lower ERA 100 times out of 100.

[therealnod]

I just realized that when people make these threads about the most important stats, I often don't think about what I really want to know about a player when I consider what to post. It has occured to me that what I look at when evaluating any player boils down to some very important information that I never see listed among the alternatives: age, league, K/BB ratio, position, handedness, and similar players. Really, I spend a lot of time contemplating those data points. That info is critical in understanding where a player is and where he's likely going. Things like position and handedness aren't very important on their own, but age, league, and K/BB ratio are when you consider the big picture.

For the question posed for this thread, I have this to say:
Out of the five stats to choose from, I'd list ERA+ last; actually, I'd throw ERA+ in the trash can. For a pitcher, in addition to the above data I mentioned I'd want to look at HR/9. I also take a look at the Three True Outcomes (http://www.baseballprospectus.com/article.php?articleid=7240) rate. I take a look at Ks, BBs, and HRs. I give a pitcher a "win" for every TTO confrontation that results in a K, and a "loss" for every BB and HR, then get myself a rate of K/BB+HR. It's a thumbnail stat, and it's fully within the DIPS spirit of pitcher evaluation.

A big welcome to Frank Bundy, III. I suspect maybe Matt dragged you over here? Anyway, I'd like to thank the both of you for choosing to spend some time here at Baseball Fever. It's an exciting time in baseball analysis, and I'd guess there is still a great deal to do yet. So get back to work and stop playing around on message boards! http://www.baseball-fever.com/images/icons/icon12.gif

If you guys aren't careful I'm gonna start plugging At Home Plate (http://www.athomeplate.com) everywhere.

[NickG]

The one with the lower ERA 100 times out of 100.

Absolutely. All we know about the guy that won 20 games is that he had fantastic run support. What does that have to do with him being a good pitcher?

(Unless, of course, the guy with the 5+ ERA played in Coors Field, while the other guy played in... I don't know, Hell's parking lot or something.)

[SABR Matt]

Please do thealrod. :)

AHP could use the traffic...Frank (my front man) is doing an excellent job making Serious Baseball weekly reading material for fans in the know..and I'm sitting here pulling my hair out over how best to study linear weights changes with time to get weights that make sense. :)

PCA is still being tweaked and improved...but...I hope you'll continue to follow my work...I hope I don't annoy every last one of you with my reems of data and tendency to be blunt. :)

[SABR Matt]

BTW Frank...don't send me hate mail...I was kidding about the front man crack. :)

[frnkbndy1981]

Matt...Don't worry about it.

Hey guys...Thank you for the welcome. I will participate in many threads and well create a new one every Wednesday and Sunday which will be a link to my article that was published that day.

Thanks again...and any plugging you do for AtHomePlate.com is more than welcome.

Thanks.

Frank Bundy III

[RuthMayBond]

yes...K/BB would have been better than K/9That's what cracks me up about you. You claim you have the superior answer, then a few minutes later you're changing your mind :crazy :crazy

[leecemark]

--RMB, K:BB wasn't one of the options so Matt didn't exactly change his mind. I was just asking if he would have prefered that over K:9 if it had been an option becasue that would be my own inclination.

[SABR Matt]

That's what cracks me up about you. You claim you have the superior answer, then a few minutes later you're changing your mind :crazy :crazy

I'm confused...what the heck are you talking about?? I didn't change my mind about anything here

[hogwashed]

I would like to get people's opinions of the best pitching stat out there.

[SABR Matt]

Of the ones readily availabe, I recommend you look at Pitching Runs Created...a Hardball Times (www.thehardballtimes.com) invention. It's the most modern and DIPS-compliant (without being pure-DIPS which has flaws) method I know of, and I like the scale it's in.

[Mariano_Rivera]

Idon`t know a lot and I like mine CERA because it's very easy to figure out and understand. I also think it is extremely accurate

[John Beamer]

FIP - a tango creation is pretty useful. It is also "DIPS compliant" and looks like an ERA ..

[SABR Matt]

Where do I find FIP...haven't been able to find a site that lists it...I know THT, BP, and Fangraphs don't list it.

[John Beamer]

THT lists FIP

http://www.hardballtimes.com/thtstats/main/?view=pitching&league_filter[]=2

If you click on the THT pitching under the stats tab you should find it. Hopefully the above link should work. THT also has the formula in its stats glossary.

[SABR Matt]

Well slap me silly...how did I miss that?

THT has a lot of really great info. FIP and PRC are both stats I would sanction using in a debate about a pitcher's mertis.

[adamsowell]

If you were a starting pitcher and you were going to lead the league in one of these 4 stats, what would it be?

-ERA
-Strikeouts
-WHIP
-BA against

[Wade8813]

When evaluating a player's season, I look at WHIP, ERA+, and make sure they at least have a decent amount of IP. I'm not real big on IP like some people are - I just require that they have enough.

[Zagi-CRO]

If you were a starting pitcher and you were going to lead the league in one of these 4 stats, what would it be?

-ERA
-Strikeouts
-WHIP
-BA against

The first of all - WIN.
The 2nd - ERA
The 3rd - strikeouts
and IP > 180.

But, IMO, the best is mix of that, some kind of component index.

[Tyrus4189Cobb]

when you rank a pitcher, what are the best stats to judge him by?

[AutographCollector]

when you rank a pitcher, what are the best stats to judge him by?

Obviously winning % and ERA.

[Tyrus4189Cobb]

Yes, i do go by those. I also go with strikeouts and IP, to make sure that the picther didn't have a low ERA due to few IP.

[Lucifer]

Strikeouts aren't a good measuring stick, as some great pitchers don't strike guys out. Also I wouldn't say wins or losses cause as you see with Johan Santana/Andy Pettitte right now you don't win all or even most the games you pitch well in. I think ERA and hits per inning are the best

[Memphis]

I invented a pitching metric that isn't the best, but helps us understand pitching better.

It is called In Play Hits Prevented.

It turns out that my collegue, SABR Matt, had actually already stumbled upon the logic I used to create this stat before me...so I had this help on this article.

But, the interesting discussion thus far in thread should definitely include this metric.

Check it out:

http://athomeplate.com/sb2305.shtml

I read your article at the above link. Very interesting and I agree with most of it.

However, making "bad contact" could result in a hit; just as making "good contact" can result in an out. So I think the reasoning is somewhat flawed since good-contact-outs are treated equal to bad-contact-outs.

How many times during the course of a single ballgame do we see batters hit screaming line drives for outs? Or, how many seeing eye grounders do we see that in every other respect are routine. There's often a fine line between a routine double play ball and ball that just barely splits the middle infielders for a back-breaking hit.

Further, how many times do we see players fooled on a swing who still manage to get a hit by just sticking their bat out there? I've seen HOME RUNS hit like that! Is that considered "bad contact"?

I don't know. I would still lean HEAVILY to the notion that the pitcher really cannot control what happens to a ball put into play. Not consistently, anyway.

David Emerling
Memphis, TN

[Erik Bedard]

The first of all - WIN.
The 2nd - ERA
The 3rd - strikeouts
and IP > 180.

But, IMO, the best is mix of that, some kind of component index.

Obviously winning % and ERA.

Wins (and W%) are largely useless. Imagine one guy going 5 innings and giving up eight runs while his team scores twelve. He would get the win. Now, imagine someone going nine innings and giving up one unearned run, say on a single and a three-base error. That guy would get the loss. Say they each did that for every one of their starts, making one 30-0 and another 0-30. What makes the guy with a 30-0 record better than the one with an 0-30 record?

W/L used to be relevant when pitchers usually pitched nine innings, but it lost all relevancy somewhere around 1959. Now, it's become something of a joke.

For a basic stat, WHIP, K/BB, H/9, BB/9, HR/9, and K/9 are generally good. But of all those, I'd say K/BB is the best. WHIP is the best of the ones on the poll, though.

[writerhoward]

I want to compare two major league pitching staffs to see which is pitching better. What's the best stat (or stats) to use in that comparison?

Howard

[downstairs]

I want to compare two major league pitching staffs to see which is pitching better. What's the best stat (or stats) to use in that comparison?

Howard

You mean two teams from the same year?

Hoenstly, if you're just looking for "better"... the simple stats work... ERA, runs allowed, etc. The fancier stats will just get you to the same conclusion more times than not.

[writerhoward]

You mean two teams from the same year?

Yes. I want to compare the Mets with both the Phillies and the Braves.

Howard