Slug this
Sometimes the weirdest questions get stuck in my head. Sometimes the most random queries just won’t leave me alone, like an itch that needs to be scratched, or a hungry belly that needs feeding.
So when Paul Konerko steps to the plate, I’m just not that impressed. Oh sure, he hit 40 home runs, good for 5th in the AL. But his SLG of .534 was 9th in the league. Not really impressive in my mind. And I start wondering just how normal it is for a 40-homer guy to slug less than .600. Turns out, once I start digging around, not that out of the ordinary.
Just then Konerko bruises the first offering from Chad Qualls, sending it into the left field seats – a dramatic grand slam that gives the Sox the lead.
So who has the lowest SLG of any hitter with 40 home runs?
The answer to that [ed: insert as-yet-unnamed USSM sponsor here] trivia question would be the Mariners’ own Jay Buhner in 1997.
Buhner hit 40 home runs and slugged .506. His 131 hits are sixth fewest among those with 40 homers, and his 60 XBH are 5th fewest.
Comments
55 Responses to “Slug this”
Was that the year Bone had the foot injury? He led the Majors in K’s that year – 175 K vs only 131 hits. Swinging for the fences in every at-bat, apparently.
Well, he wasn’t swinging at EVERYTHING- he did walk 119 times that year (2nd in the league). His OBP was .388. Not bad for a .243 hitter to have an OPS close to .900.
But he also managed to ground into 23 DP’s that year. Ouch.
Might have been the lowest SLG, but there are 15 guys who slugged at least 40 HR and had lower OPS, “led” by Tony Batista in 2000 (41 HR, 827 OPS).
1997 was also the year that Jay had only 1 Sac Fly the entire season (ironically he had 10 the year before, almost 20% of his career total occuring in one year).
IIRC during that season I believe he was also labelled in some “confidential survey” or leaked scouting report as the “second worst fundamental player in baseball”. Number one supposedly was Manny.
I miss Jay. His 300th HR at Safeco as a PH against Isringhausen in an aborted comeback was a big thrill in 2000. Also the time he got on base in another comeback and Rickey Henderson entered the game as a PR to score the winning run in his first game as an M.
Sexson at times reminded me of what Buhner might have looked like had he played at Safeco in his prime.
I have a question:
So when a batter steps in the batters box. The tv shows his stats which for example read: .270 BA, .370 OBP, .600 SLG, and .970 OPS
So I should expect the batter has a little less then a 1/3 chance of getting a hit, and a 1/10 chance walking or being HBP.
…Ok I get that…
My question is: How do I interpret the SLG and OPS stats in the same manner?
Thanks,
Mark
Sorry if this is too mundane of a question. As my old teacher use to say: you can feel foolish for a moment or you can continue to be ignorant.
#5: If a hitter has a .500 slugging percentage, they’re producing 1 base on their own hit per every 2 at-bats. If they’re batting .333, they’re getting a hit once in every 3 at-bats. So, change the number of at-bats to 6 and the hitter would get 2 hits and 3 bases. Or, a single and a double every 6 at-bats.
You’re on your own with OPS.
#6 Thanks. I have a better understanding of slugging percentage now.
Hopefully someone will help me with understand how the math of OPS, applies to the field.
#7: OPS is “On-base Plus Slugging”, a very quick-and-dirty rate measurement of batting offense. That is, it’s a rate stat (like batting average and on-base average) that attempts to incorporate most of what a player is doing at the plate. There are many more complex rate stats that do this, but all of them are much harder to calculate quickly while looking at the basic stats on TV or in the newspaper. The advantage of OPS is that it’s easy to add up and use, but the disadvantage is that it’s not a very good estimate. One reason is that it overrates the impact of slugging relative to getting on base, simply because slugging tends to be a larger number than on-base. On-base determines the rate at which a player uses up outs, so tends to be a stronger factor than slugging in determining effectiveness; OPS doesn’t reflect that very well. Another problem is that it adds two other rate stats together, but each of them uses a different “per” unit; on-base measures times reaching base per plate appearance while slugging measures total bases per at bat. Generally it’s just a blunt tool and there are much better tools available. Googling “Runs Created” or “Equivalent Average” can point you to some other efforts to encompass offensive performance into one rate stat. They’re worth knowing about, because good measurements hide less of the useful information about players.
Thank you Jim.
I guess the days of looking at a new player on the team and thinking “I’ll just look at the box score to see if he is any good” are basically gone.
From now on I’ll have to check in here and Baseball Prospectus before getting too overly excited.
#8:
While OPS is a blunt tool, it’s a far, far sharper tool than batting average (BA), which has been the traditional measure of offensive output. IMHO, BA is worse than useless because BA actually leads to totally incorrect conclusions about player offensive value.
The average fan is simply looking for something quick and easy to grasp that allows them to generally understand what’s going on. OPS is perfectly fine for that purpose. For those who are so inclined, they can proceed to the further refinements of offensive contributions. But understanding of baseball would be vastly enhanced if newspapers, announcers and fans all routinely used OPS instead of BA as the key offensive stat.
It’s a classic 80:20 situation – 80% of the value derives from the first 20% of effort. OPS is easy to calculate and provides a huge enhancement in combining the value of slugging and on-base percentage. Additional refinements are possible, but consistent with concept of diminishing returns, the improvements offered are proportionally less and require more effort to provide.
My advice to “mark from oly” is to focus on OPS and become familiar with using it. As long as OPS works, fine. If at some point OPS starts to seem limiting or incomplete, then proceed to further refinements.
Why not use OBP, then? It’s far easier to understand, calculate, it’s meaningful where OPS is not (“How often does this player not make an out?” versus “Where does the player fall on an abstract scale made by blending two numbers together?”). It’s 80% of the way there with far, far less work than OPS requires.
Peter – if you’re not that impressed with Konerko, does that mean you’re not impressed with Sexson? Their season lines are almost exactly the same:
Konerko – .283/.375/.534
Sexson – .263/.369/.541
I understand that Sexson is playing in a tougher park, but I gotta think that Sexsons numbers would look a little better than Konerko’s if he played at US Cellular, but not by much.
Well, Sexson’s probably a great indication of what Konerko’s free agent contract is going to look like. Just add a few million for inflation and the fact that Konerko’s not coming off an injury.
I’m with Derek on this one. If I can only have one number, give me OBP over OPS. And personally, I don’t care all that much for SLG… I’d rather look at XBH as a percentage of hits.
Konerko: 39% (24+0+40)/163
Sexson: 52% (36+1+39)/147
But see how close they were in SLG? Yet, the real difference was that Konerko hit for a better batting average. In reality, Sexson was a much better “power hitter” than Konerko this season.
Why XBH as a percentage of hits and not XBH as a percentage of balls in play or plate appearances? It seems like singles are more prone to random fluctuations over short periods of time, and that looking at XBH helps to kill out a bit of the randomness. (It’s tougher to hit a lucky double or a lucky home run than it is to hit a seeing-eye single.) So, why implicitly include singles rate by making it a percentage of hits? Just a thought.
Steve, interesting point about OPS vs. BA. I agree OPS tells us more, but remember it weights BA 2x. Conceptualize it as follows (sorry for going algebraic on you):
OPS = OBP + SLG
OPS = (BA + walks) + (BA + iso slg)
OPS = 2(BA) + walks + iso slg
So while we’re not sure exactly what OPS is telling us, BA must tell us something to get that kind of weight in an improved measure.
I personally think using some kind of Runs Created type stat (in a team context) as the best way to evaluate a hitters offense. I personally like using Base Runs.
Also throw in the extra few million per year that one grand slam (and actually his postseason hitting in general) will earn him. Its the Carlos Beltran/Dexter Jackson effect
#16, the reason is hits advance runners, walks don’t. OPS is still imperfect though. Part of the problem is SLG is per AB as opposed to per PA. This makes OBP and SLG disimaler stats. If you adjust for this you get something like RC/PA. SLG is not supposed to give you the players offensive value to the team, but rather their power per AB. I think isolate powerr is a much better stat to indicate a players “power”, but is still not a good way to judge their offensive value to the team.
It’s just that OPS correlates better with runs produced than does any of BA, OBP, or SLG.
It’s imperfect, it overweights slugging, etc. But it does combine into a single measure a hitters effectiveness at not creating outs plus giving him credit for the number of bases generated when he does get a hit.
If you prohibited me from adding two numbers together to generate a stat, then among AVG, OBP, and SLG I would certainly take OBP. But if I’ve got all three of them in front of me, I’ll add OBP and SLG together and use that.
I’ll know that I’ve got an axe in dire need of sharpening, but at least it will be an axe instead of penknife.
Jason articulates my thoughts rather well in #14.
Comparing their isolated slugging (SLG minus batting average)…
Sexson – .278 (5th in the AL)
Konerko – .250 (9th in the AL)
So who plays the Mets to Konerko’s Beltran this offseason? My money’s on the Red Sox, but there are several other teams looking to upgrade production at 1B (the M’s, fortunately, don’t have to get involved in that scramble).
Peter – Did you do the research on your own, or did you use Lee Sinins’ Sabermetric Baseball Encyclopedia? Does anyone have that? Is it fun and easy to use?
#23, are you sure about that? from http://www.hardballtimes.com/main/article/2005-gold-gloves – “Sexson should be a DH.”
Comparing Sexson and Konerko’s raw slugging, or even raw ISO stats, are misleading. Slugging percentages vary dramatically depending on the ballpark, and the difference between Safeco and US Cellular is very significant. When you compare raw SLG, or raw XBH%, or raw ISO, you’re comparing numbers from one of the least homer-friendly parks in the majors to one of the most homer-friendly parks as if they were equal. If Sexson’s raw numbers are better despite the difference in parks, that indicates that he’s a far better power hitter than Konerko.
#24, a copy of Lahmans Baseball database and access can get you that info too.
The key (I think) to understanding sexsons power (at safeco) is that he BIG flyballs to all fields. This is a huge advantage here. Boone (when he was good) and A-Rod had the same thing going for them. This is why they hit so many home runs despite the park. Konerko is a dead pull hitter. Almost all his home runs were pulled. I think there is a good chance he’d be killed for power at safeco.
Chris,
Don’t take Gassko’s rumblings too seriously. He’s excited about his metric, but it’s wildly unproven and he makes some ridiculous claims based on tiny samples of suspect data. Suggesting that Kelly Johnson could be an “excellent” centerfielder made me laugh out loud, and I’ve been driving the Kelly Johnson bandwagon for a few years.
This may be a dumb question, but would it be possible to combine OBP and SLG into a single stat that actually meant something? By, say, adding together BB + HBP + TB and dividing it all by PA, you’d get the number of bases a batter is likely to reach every time he steps up to the plate, either through a base hit or by a walk/HBP. Sexson in 2005, for example would have a value of (89 + 6 + 302)/656 = .605. Meaning that in ten plate appearances, he would account for 6 bases. It would lump together walks and hits into one figure, but it wouldn’t count BA twice, and it makes more sense conceptually than OPS.
Is there some obvious reason I’m missing why this hypothetical stat wouldn’t work as well as OPS in describing offensive effectiveness?
I think you’ve just described Total Average.
Thanks for pointing that out Dave. There is a serious flaw in that stat. I think basing a defesive range stat on opportunities that may not exist (because they are based on how many opportunites the position SHOULD field) is going to cut some people short and give others better numbers than they should.
I think his idea is okay; I just don’t think he realizes how large the noise can be in one season samples of defensive statistics. Well, he might realize it but not care. Whatever the reason, Gassko’s become famous for a writing style that goes way beyond what his data shows. Sometimes, I feel like I’m reading a parody of a statistical analyst when I read his stuff.
Thanks for the link to Glassko’s “Gold Glove” article. Interesting how Jose Cruz Jr. can be -19 in CF, and yet +14 in RF. I never realized there was that large a skill differential between the two.
“led†by Tony Batista in 2000
Remember that Tony Batista is the guy who managed 30 HR and 100 RBI while still being BELOW replacement level.
So that doesn’t surprise me at all.
#5: If you keep in mind that the AL as a whole is putting up a line of about .270/.340/.430, you can look at a given player’s line and see at a glance if he’s keeping up with the league. (In 2005, the AL numbers were .268/.333/.424, but that’s close enough for government work.) So .770 OPS would be average. Obviously, you can tolerate less than that from a great defensive player up the middle (YuBet), but you need to make it up with much better numbers in your power positions (1B and, uh well, just say 1B for the M’s.)
#16: OBP is not BA + Walks. OBP punishes sacrifice flies (but not, strange but true, sac bunts) and rewards an HBP as much as a BB.
#23: I don’t expect to see the Red Sox spend a lot of money at first base this year unless Manny is gone, and I expect Manny to be back. And Konerko’s been pretty vocal about not wanting to play on the east coast, if I remember correctly.
Yeah. There are three “I don’t know what’s going on here” surprises in the Glassko article. That ought to raise a flag. That, and giving Jeromy Burnitz a gold glove.
We’re all pretty vocal until someone offers us $65M over five years.
#36 Thanks LB.
#38 JMB, if anyone offered me $65m.US over 5 years. I’m sure it wouldn’t be anything close to legal outside of international waters.
#36– these days it’s: “Paul Konerko said Chicago is his first choice but hasn’t ruled out anybody when he enters the free-agent market after the World Series…. Konerko said lifestyle and playing for a team with a chance to win are his primary motivators”
So, correspondingly, who among 40 homer guys has the lowest isolated slugging average?
My quick-and-dirty stat of choice is (OPS minus BABIP). That removes the double-counting of BA; also it penalizes a batter who is lucky with balls in play, and compensates a batter who is unlucky with balls in play.
As for a more comprehensive stat, how about calculating “bases contributed versus outs made”?… adhering to Earl Weaver’s maxim that the team’s 27 outs are its most valuable asset.
For example:
TB + BB + HBP + SH + SF + SB
—————————————-
(AB – H) + SH + SF + CS + GIDP
where (AB – H) is a rough approximation of non-sacrifice outs made at the plate.
There’s no reason to believe that a hitter is “lucky” or “unlucky” on his batting average on balls in play. Ichiro is always going to outperform Yorvit Torrealba on BABIP. It’s not luck. It’s skill.
The BABIP theory works for pitchers. It does not apply to hitters.
Well… while that’s true, batting average is subject to a lot of random variance from season-to-season even for hitters. I wouldn’t throw it out, though, but it’s worth noting that a year’s AVG can be a fluke.
Sure. Ichiro, obviously, is going to have a lot of variation in his BABIP. But Ichiro’s BABIP is always going to be higher than almost every catcher in baseball.
The whole point of the DIPS theory that brought batting average on balls in play to light is that the pitcher does not have significant control over the result of a batted ball. The hitter, clearly, is the one who does have significant control over the result when he strikes the ball.
BABIP for hitters most likely isn’t totally useless. Much like riding a donkey to work, you’ll get there eventually. I just have no idea why, in the presence of bicycles, horses, cars, and even the monorail, you’d choose to ride a donkey.
43: If BABIP is a skill, it should be repeatable from one season to the next, right? I’m not sure how repeatable BABIP is.
Leaving aside the debate on BABIP for a moment, i.e. just using plain old BA instead of BABIP: it could still be argued that (OPS minus BA) is a slight improvement over OPS, by avoiding double-weighting BA.
This would apply especially to a batter who gets lots of singles with only a few extra-base hits or walks (Ichiro! for instance). Not that I’m trying to slag Ichiro!… but double-counting all those singles does overstate his contributions at the plate a little.
I think OPS is fine as a quick and dirty stat, but tweaking it to be more accurate is pointless because that defeats its primary utility, which is ease of calculation. If you want a single stat that is very accurate, just look up MLVr or EQA. The latter is especially fast to find, so speed shouldn’t be an issue. For me, I’ve completely stopped using OPS and instead use either those two advance metrics or simply write out BA/OBP/SA. OPS is just too blunt for my tastes and it doesn’t really have an advantage over the other stats I listed.
I think even to a certain extent, though, Ichiro’s BABIP over the course of a season will be a much better indication of his talent for getting base hits than a high-K, high-BB type hitter. That is, I think a lot of this has to do with sample size. Ichiro leads off, plays a lot of games, and doesn’t walk or strike out much. So, he’s probably at the top of the league in balls in play, while someone like Adam Dunn, in just as many games, doesn’t give us as many opportunities to measure his ability to hit singles.
Jose Cruz, Jr. this year was kind of an example of this. His career BABIP is about .298. With Arizona the first half of the season, he only had 137 balls-in-play, and hit for a miserable, and fluky, .227 BABIP. Then, with the Dodgers, he only had 107 balls-in-play, and hit for an equally fluky .383 BABIP.
If Ichiro’s BABIP is going to fluctuate a lot, and he gets 550-600 balls-in-play a season, it’s easy to see how it can fluctuate a lot more for part-time players and three true outcomes-type players.
How about this: who has the lowest SLG for a guy who hit 50 or more HR in a season? I’m pretty sure Andruw Jones is at or near the top of that list.
How about this: who has the lowest SLG for a guy who hit 50 or more HR in a season? I’m pretty sure Andruw Jones is at or near the top of that list.
You’d be right.
Dave, perhaps the most interesting thing about his idea of moving Johnson to CF is that Ryan Langerhans and Jeff Francoeur(2/3 of his “dream” lineup) could both handle CF at a better rate than Kelly could.
As to where he think’s we are getting the offense to replace Andruw…uhh…at least its cheaper?
The only SLG lower than Andruw Jones’s .575 I could find with a comparable number of HR was in 1964 – Harmon Killebrew hit 49 HR with a SLG of .548
#42: Bill James’s (please pardon any misattribution) “RC/27″ (Runs Created per 27 Outs) was a pioneering attempt at such an Offense Per Out type of stat. Advantages are that it addresses the critical issue in offense, that the task is to score as much as possible in as few outs as possible; also the resulting number looks just like a team Runs/Game number, so it provides a guess as to what a lineup of nine identical players would score in a 9-inning game. Disadvantages are a lack of park adjustments, although that’s since been added to other similar stats, and the RC formula has been improved upon by many other approaches in the last 20 years or so.
You’re right to emphasize outs. Probably the best thing for someone with a lot of exposure to traditional batting stats to grasp is the out rates for hitters. Since it’s the denominator for the Offense/Out rate, it has a large impact on performance. Unfortunately, broadcasters still don’t show OBP enough to really demonstrate the out rates of different players to the typical fan.
Wow, Killebrew in 1964…49 homers, but just 11 doubles and a triple (to go with 95 singles). That was the year *before* the Twins won their first pennant since they were the Senators (when they rode Killer, Carew, Jim Kaat and MVP Zoilo Versalles to the World Series). Couldn’t get past Koufax and Claude Osteen in the WS though…
But OBP/SLG tells a million times more than OPS in my opinion. OPS isn’t even a real number; OBP and SLG have different denominators, so adding them doesn’t yield anything meaningful. OBP and SLG are much better.