OPS/3

by SLOtown on Dec 29, 2005 1:24 PM PST Comment 52 comments

So I was reading through Moneyball the other day and I came across this passage:

An extra point of on-base percentage was clearly more valuable than an extra point of slugging percentage - but by how much? He (Paul DePodesta) proceeded to tinker with his own version of Bill James' "Runs Created" formula. When he finished, he had a model for predicting run production that was more accurate than any he knew of. In his model, an extra point of on-base percentage was worth three times an extra point of slugging percentage.(p.128)

DePo saw that adding OBP and SLG together meant they were of equal importance. He knew this couldn't be true, that's why he created this formula. So, I have calculated this value for a handful of A's players. I call this value OPS/3 since the formula is a player's OBP plus 1/3 his SLG.

Player OBP SLG OPS OPS/3
Kendall 345 321 660 452
Johnson 355 451 806 505
Swisher 322 446 768 470
Ellis 384 477 861 543
Kotsay 325 421 746 465
Crosby 346 456 802 498
Bradley 350 484 834 511
Chavez 329 466 795 484
Payton 306 444 750 454
Kielty 350 395 745 482

Looking at these numbers, a few things jump out at me. Kendall was widely considered to have had a terrible 2005 at the plate while JayPay was praised for picking up the offense. Looking at their OPS, one could easily agree that Payton was a better hitter than Kendall this year, but if you look at their OPS/3, you see that their run production and, in turn value as a hitter is about equal. The reason for this is that OPS/3 focuses much more on OBP, and Kendall is a much more patient hitter than Payton.
I wondered what the corellation between OPS and OPS/3 was, if a player had a high OPS it meant he had a high OPS/3. I figured if the corelation was high, then this whole experiment was a waste of time. I found the r^2 to be .787 which is a poor correlation. I know, I know, small sample size, but I still think there's something to this OPS/3 stat and we SABRANers should use it along side OPS not only because it's in Moneyball, our bible, but also cuz we should use the most accurate stat possible to calculate offensive production. If it's good enough for DePo, it's good enough for me.

0 recs | Comment 52 comments

Story-email Email Printer Print

Comments

Interesting...

One number that really jumps out is Ellis .543 OPS3 value. Alomst 40 points higher than anyone else who was on the team! And how many months did it take Macha to move him to the top spot in the lineup?

"Macha, Howe...whatever." --salb918

by BruceBochte on Dec 29, 2005 1:43 PM PST reply actions 0 recs

When I see

that a stat suggests Kendall and Payton were comparable offensively, it doesn't make me think "Hey, Kendall did better than I thought" or "Hey, Payton did worse than I thought." It makes me question the formula.

You have one guy who gets hit by a lot of pitches and never gets XBHs, and one guy who gets a lot of XBHs and never walks. You watched them both last year--who do YOU think had a better year?

My formula: OBP without XBH = LOB.

by Nico on Dec 29, 2005 1:46 PM PST reply actions 0 recs

What you have to remember, Nico,

is that OBP is the single-most important determinant of scoring runs. Jason Kendall derives almost all of his value from his OBP, whereas Payton derives all of his value from his power. OBP > Power, and it's not particularly close.

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 29, 2005 1:53 PM PST up reply actions 0 recs

It's not that I've forgotten

this, Sal, I just question it to some degree--certainly to the degree of giving it thrice the weight. OBP is, first of all, at least a slightly imperfect calculation in that it gives 0/1 weight to GIDPs instead of 0/2 weight, whereas in fact, every time Kendall killed a rally with a DP, the A's were substantially inconvenienced, and I became extremely agitated,and somewhere a puppy had its front left leg run over by a bus. That's really, really bad.

But more significantly, walks and HBPs simply don't advance runners like singles do, or like doubles and HRs really do, big time el grande style. That is their weakness, one which is emphasized if you have too many high OBP, low SLG guys together.

Bottom line: a lineup of 9 Kendalls would be very unproductive, as would a lineup of 9 Paytons. But would I prefer to have 7 Kendalls and 2 Paytons, or 5 Kendalls and 4 Paytons? Probably the latter, personally.

by Nico on Dec 29, 2005 4:05 PM PST up reply actions 0 recs

DPs

Nico, I think you've pointed to one of the few real weaknesses of switching to OPS/3: OBP, for some reason, does not count a DP as an 0-for-2, while SLG does count it that way. Kendall's SLG was severely depressed by all his GIDP (almost as much as you were, Nico). So if you cut the impact of SLG by 2/3, you cut the impact of all those GIDPs by 2/3.

Another way of putting it is this: if you had 2 players whose stats were identical except that Kendall had 27 GIDP and Kendoll had 7, there would be a 20-out difference in the calculation of their SLGs, which would show up as a big difference in a standard OPS. Under the OPS/3 formula, that difference is much, much smaller.

"And Julio Franco is batting right-handed!" -- Wayne Hagin, A's radio play-by-play, mid-80s

by Nick on Dec 29, 2005 4:22 PM PST up reply actions 0 recs

I do not think

that standard SLG calculations reflect DPs: SLG = TB/PA. An out-based metric, however, makes a heck of a lot more sense - as you so eloquently point out.

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 29, 2005 4:25 PM PST up reply actions 0 recs

As if Kendall

needed something to depress his slugging percentage. It's like chopping off all of someone's fingers except for the left thumb--and then passing a law against hitchhiking.

I'm interested in this Kendoll--if we sign him, is he willing to dive to tag Michael Yaung?

by Nico on Dec 29, 2005 4:31 PM PST up reply actions 0 recs

No, but...

...his wife, Barbie, is a heck of a looker.

And he won't require much in the way of parking space.

by Ozzz on Dec 29, 2005 7:45 PM PST up reply actions 0 recs

Walking the middle road.

It's a good philosophy if you don't want to offend anyone. And OBP = 3*SLG is probably overstating the case, as you correctly point out. And the simple stats do not account for DPs either, although I submit that they are a minor affect except for outliers like our own Jason Kendall.

What's eating you is the frustration of watching our offense in 2005. But if you were a Cubs fan in 2004, you would be frustrated with the low-OBP, high-SLG offense, too.

Our eyes deceive us constantly, and as frustrating as our DP/HBP/LOB offense is, it is the byproduct of OBP. I know that you value OBP, and you make a good point that you need to have some heavy hitters, too. But if you have to have one or the other, give me OBP every day of the week and twice on Sundays. Payton's .302 OBP is awful and hurts the offense on balance, and as terrible as Kendall was last year (and it was bad) we have to recognize that the .345 OBP contributed to the offense in a positive way.

And you know, I never liked puppies anyway.

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 29, 2005 4:24 PM PST up reply actions 0 recs

You and Poochini's vet,

Katia Bolsov.

by Nico on Dec 29, 2005 4:33 PM PST up reply actions 0 recs

OBP without SLG equals LOBs in extraordinary

quantities.

The A's have fallen into the LOB rut too many times to downplay the importance of power in the mix. The first two months of last season the only one hitting with power was Kotsay. Everyone else was doing the "Kendall". Walks, singles, HIDP.

The Holy Flails

Let's NOT go back to that sickening vacuum of runs.

"...It might have been a great year with a real DH."

by A s Eh on Dec 29, 2005 6:56 PM PST up reply actions 0 recs

Exactly, A'sEh.

Sometimes I think the A's boast their high LOB totals like an achievement: proof that they really are getting lots of guys on base, yessiree!

Get 'em on, get 'em over, get 'em in. It's not a "two outta three ain't bad" proposition. You really need to do all three; you really do.

by Nico on Dec 29, 2005 7:01 PM PST up reply actions 0 recs

Tell that to MeatLoaf

by LowcountryJoe on Dec 29, 2005 7:23 PM PST up reply actions 0 recs

LOB a good sign?

Nico wrote: Sometimes I think the A's boast their high LOB totals like an achievement: proof that they really are getting lots of guys on base, yessiree!

They wouldn't be wrong to do so. There is a strong positive correlation between team LOB and team runs scored. For 2005 AL teams, here are some correlation coefficients:

Runs Scored to LOB: +0.78
Runs Scored to Total Base Runners(): +0.74
Runs Scored to Total Bases: +0.82
Total Base Runners to LOB: +0.93
Runs Scored to LOB%(): +.47

In other words, team LOB numbers correlate almost perfectly with the number of base runners, which in turn correlates very strongly with scoring runs. And there is a significant positive correlation between the percent of runners stranded and the total number of runs. In other words, yes, leaving a lot of runners is a good sign on the whole, and as good a predictor of run-scoring prowess as anything.

For those who prefer anecdotes to stats, the team that left the most runners on base, and had the highest LOB percentage, was the Yankees, who scored the second most runs. The team that left the fewest on base was the White Sox, who were ninth in runs scored. The team that stranded the lowest percentage of base runners was the Devil Rays, about whom the less said the better.

(Def. of non-standard stats: () Total Base Runners = Hits - HR + BB + IBB + HBP). (**)LOB% = LOB / TBR)

by matthias on Dec 29, 2005 10:24 PM PST up reply actions 0 recs

sorry about the formatting there

something about how I wrote the data was interpreted as HTML formatting instructions by something.

by matthias on Dec 29, 2005 10:25 PM PST up reply actions 0 recs

I know, matthias--

Whenever the A's leave runners at 2nd and 3rd, you can hear me cursing, "Damn, if only we had left the bases loaded!"

by Nico on Dec 29, 2005 10:27 PM PST up reply actions 0 recs

I'm sorry

that statistics are sometimes counter-intuitive.

Well no, I'm not - they wouldn't be useful if all they ever did was confirm your gut feelings.

by matthias on Dec 29, 2005 10:32 PM PST up reply actions 0 recs

I've never intuited

anything about my counter, other than the fact that it was probably Formica.

No, it's not surprising, because teams that score more and teams that get more people on base (or make fewer outs) naturally go together.

It's when you observe your team doing one of the things well, but not the other, that you need to put the formula to the side for a moment and look at why. And if a lot of your getting on base isn't moving other people to some of the other bases, particularly the flatter 5-sided one, well...

by Nico on Dec 29, 2005 10:43 PM PST up reply actions 0 recs

What aren't they doing well?

What exactly have you observed them "not doing well"? What's the evidence?

Here's another little stat I cooked up for this discussion: Run Efficiency = (R - HR) / TBR. Roughly the percent of base runners that score. In the 2005 AL, Boston ranked 1st in this category with 33.6%, and Minnesota ranked last at 29.2%. Oakland ranked 4th with 31.9%. Fourth is also where they ranked in total base runners.

So no, I do not see how the A's are particularly bad at getting runners around to score.

What the A's really were bad at last year was hitting home runs. So they got most of their runs by clogging up the base paths with hits and walks and HBPs, and they were about as good at bringing those runners around as anybody.

by matthias on Dec 29, 2005 10:56 PM PST up reply actions 0 recs

You're looking strictly

at "run totals," which were inflated by a flurry of 9-15 run outbursts on days when 5 runs would have done just as well. The number of games the A's had in which they scored 0-2 runs (when 4-5 would have done the trick) tells the story of how well they could "get 'em in" on an as-needed basis.

But yes, nobody could run it up on Jose Lima or Jason Christiansen (remember that fun 16-0 nailbiter?) like the 2005 Oakland A's.

by Nico on Dec 29, 2005 11:01 PM PST up reply actions 0 recs

Yes, my stats are imperfect...

have you got better ones?

Again, what exactly is it you think the A's did consistenly poorly at compared to the rest of the league?

Just say what it is; I'll be happy to run the numbers and see if your claim is true.

by matthias on Dec 29, 2005 11:04 PM PST up reply actions 0 recs

I'm not sure

how it can be looked up, but here's how I would articulate it: The A's had (have) a sound offensive philosophy of emphasizing high OBP, but weren't (aren't) good at strategically straying from it. It's a lack of versatility/balance/creativity within a basically good system.

I was pleased at the Payton acquisition because Payton is kind of an "anti-A's" hitter. But he was also "what the doctor ordered" for that lineup at that time, and I was heartened to see the A's react to a need, not a philosophy. And .302 OBP and all, Payton helped. But--in fairness--only because there were other guys who had high OBPs, drove pitch counts up, etc. Balance gooooood, one way to attack baa-a-a-aaad.

by Nico on Dec 29, 2005 11:18 PM PST up reply actions 0 recs

Yeah

I think there's something to that, but I don't know how to state in a testable way.

The weird thing about the A's was that most of their offensive stats were nearly identical to those of the Angels. The big, striking difference was that the Angels' run distribution was smoother. Why was that? Luck, or something they can change? I just don't know.

As to Payton - he helped a lot for about a month. Then he, like everybody else, just disappeared. Again, was the good month just luck? Or did people figure him out?

And you don't mention one other player: Bobby Crosby. When he was in the lineup, the team was good. When he was out, it wasn't. Too simple?

by matthias on Dec 30, 2005 9:57 AM PST up reply actions 0 recs

Woot!

Run Distribution!

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 30, 2005 10:11 AM PST up reply actions 0 recs

what about ...

... Nacho Runs distribution?

@('.')@

by monkeyball on Dec 30, 2005 5:12 PM PST up reply actions 0 recs

Gotta go!

"...It might have been a great year with a real DH."

by A s Eh on Dec 31, 2005 11:10 AM PST up reply actions 0 recs

You missed the point

Why OBP is more important than SLG is because runs get scored when clusters of players get on base in sequence. If player A gets on base, he gives up to three guys or more the chance to move him around, and he moves the guys in front of him around, as well. Now, if he does this with extra base hits, so much the better, but even singles, walks, HBP, reached on error (didn't Kendall lead the league in this?), etc. are also ways that men get on base without increasing their slugging percentage, yet contribute to rallies.

That's why OBP matters more, if you have to choose. Better that you don;t, but if you do....

Kendall is also a ground ball hitter, in a big way, which is one reason he grounds into so many double plays, but also why he reaches so much on errors. But to my way of thinking, its a further measure of his offensive contribution, because in the midst of a rally, a ground ball out is far more likely to move players along than a fly ball, foul out, pop fly, or line-out.

by dingerpower on Dec 30, 2005 10:12 AM PST up reply actions 0 recs

Yes, dingerpower,

but here's the rub. When the sequence, in a given inning, is three "walks or singles" and the last one is a walk, you have done something unusual--managed to have 3 of 5 hitters not make an out--and you have still not yet scored.

I suspect the reason the Angels' run distribution was so much more even than the A's is that the two teams operate in opposite extremes. The Angels play for one run to obsession while the A's play for the big inning to obsession. As a result, the Angels score "here and there" constantly while the A's go dormant, dormant, erupt! In the end they score about the same number of total runs--with the A's even coming out ahead here--but the Angels are better suited to winning low scoring games.

The A's offensive distribution would be preferable behind the Rangers' pitching staff, but the Angels' offense would be preferable behind a solid rotation--which is what both the A's and Angels had.

by Nico on Dec 30, 2005 10:23 AM PST up reply actions 0 recs

Nicely put

I'll add this:
A's hitters tend to be "clonis" due to the A's hitting philosphy coupled to a youthful lineup & roster.

Lots of sliders away and forkballs (Basicly Lowball/Groundball pitchers) that limit BBs to 1 or 2 per game shut down the A's like a "Frog's Ass",
... and that is water-tight!

Good teams percieve our weakness and use it against us.

How do you improve "Runs Distribution"?

If I was BB I'd try to have at least two veteran bats that are lowball predators with power.

Huff & Gomes are two such professional hitters. The effect two such examples could have on our youth should not be marginalized. Like must professional athletes hitting is a "see & do". No examples = can't do. ("Naturals" excepted).

If Saarloos pitched against the A's he would own us by not walking anyone. The A's need a hitter or two that tattoo Saarloos like a drunken sailor in a 1 week coma in a tattoo shop run by dykes...

...or close would do.

"...It might have been a great year with a real DH."

by A s Eh on Dec 31, 2005 11:29 AM PST up reply actions 0 recs

clonis = "clonish" <sorry>

Because our hitters are "clonish" also explains why we "went off" in certain games for 6 to 20-something runs during the same week as being totally shut down.

When one likes the opposing pitcher, almost all like the opposing pitcher.

"...It might have been a great year with a real DH."

by A s Eh on Dec 31, 2005 11:41 AM PST up reply actions 0 recs

Bingo, A'sEh.

It's what I call the "Carlos Silva problem"--we can't hit guys like Silva, because we wait for walks and try to run up the pitch count, and then when the count is 0-2 we reach for a slider to protect the plate, and ground to shortstop.

The basic philosophy is great and works more often than not; the lack of versatility (making adjustments to fit the opponent) is the problem.

by Nico on Dec 31, 2005 11:54 AM PST up reply actions 0 recs

Context is everything

Sometimes we forget context. If the A's could've afforded Vlad, the story would be different. They can't, so are relegated to finding those hitters at the second level who contribute most by dint of small things, such as the ways they get on base or keep rallies going.

As for your example, any season-long trend is susceptible to being found a failure in certain instances. Since it is impossible to determine in just which innings and situations the rallies and runs come, I don't think your point is persuasive.

And finally, take one of those charts that give a run value to different on-base setups (Man on 1st, one out -- you know the sort I mean) and apply it to the Angels' or White Sox' small ball shenanigans (SB, CS, S) and I think you'll find that, net-net, the impact on overall run production runs from zero to negative.

In other words, a lot of effort to achieve nothing.

by dingerpower on Dec 31, 2005 2:12 PM PST up reply actions 0 recs

I agree to a point, dinger,

unless you have the resources and just don't choose to use them. For example, all the times Kendall--an exceptional hit-and-run guy--was allowed to GIDP. (Late in the year the A's finally began putting on the hit-and-run with Kendall in those situations, but it took 4 months). Or acquiring Ray Durham but giving him the red light on the bases.

You don't need to build a team around stealing and hit-and-run (like the Angels do to a fault), but it's a good idea to use what you have when the situation--e.g., the opposing pitcher--suggests that your "tried and true" approach probably won't work as well today.

In other words, trying to steal against Buehrle, trying to hit-and-run against K-Rod, trying to walk against Silva, and trying to go deep against Hudson--not the best approaches if other options are there. And they usually are, somewhere, on any roster.

by Nico on Dec 31, 2005 4:48 PM PST up reply actions 0 recs

Golf has clubs to suit the situation

NFL offenses have variations of Running, passing and special teams

In MLB getting on base is good for starters but scoring is where it's at!

"...It might have been a great year with a real DH."

by A s Eh on Jan 1, 2006 3:41 PM PST up reply actions 0 recs

Many have tried to

put proper weights on gross stats (OBP, SLG, etc.) or events (1B, BB, HR, etc.)

My contribution is simple (2.2*OBP+ISO), but a lacking in accuracy in outlier cases such as Jason Kendall or Mark Bellhorn. To get high accuracy, you need to hit up stats like GPA (via Harball Times) or EqA (via Baseball Prospectus). These stats use some form of linear weights on individual events (here's a useful discussion on linear weights).

A commonly accepted weighting system is as follows:
1.6*BB + 2.6*1B + 3.6*2B + 4.6*3B + 5.6*HR

The comments in this thread are also very useful, since a novice (like me) is having lin weights explained to him.

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 29, 2005 1:52 PM PST reply actions 0 recs

OPS, GPA, EqA, LW, etc.

As Sal says, many people have tried something similar to this. Weighting OBP 3 times relative to slugging is probably too much.

Aaron Gleeman's GPA, used by the hardball times is (1.8 * OBP + SLG)/4. (He divides by 4 to put it on a scale similar to batting average).
The linear weights formula in Sal's post is (almost) equivalent to 1.6*OBP + SLG.
Baseball Prospectus' EqA is a little more complicated, but similar.
A quick check I just did of last year's runs scored, OBP, and SLG by team suggests a weighting of 2.1 for OBP.

But the real point is that picking the "right" weight there just isn't that important. A good reference is this article from BPro about EqA. According to them, historically run scoring correlates by about .922 with OPS, and by about .930 with EqA or other improved stats. That means you're only getting about a 1% improvement by giving OBP and SLG their proper weights in a formula like this.

Bright moments!

by andeux on Dec 29, 2005 2:24 PM PST up reply actions 0 recs

The last issue you bring up

is one which I was only superficially aware. Thanks for putting some numbers to it...

I do wonder, has the relative importance of power and on-base abilities changed over the years? Were home runs more important in the deadball era than they are today?

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 29, 2005 2:39 PM PST up reply actions 0 recs

Good question

I'm pretty sure the relative value of different events will change depending on the offense context, but I'm not sure exactly how. There may be some answers in the "How are runs really created" articles at tangotiger.net, but I haven't read them in detail.

You might also want to look at David Grabiner's Sabermetric Manifesto from about 10 years ago. He suggests that the relative weight given to OBP and SLG should be the inverse of the ratio of the league averages of those quantities. The basic idea seems to be this: on a team level, runs are correlated with OBP*SLG. The marginal effect of one player is then of the form (OBP)*(team SLG) + SLG * (team OBP) - more or less the same as the derivation of the product rule for derivatives. I think there are some holes in this argument, but it's a nice heuristic.

Bright moments!

by andeux on Dec 29, 2005 3:12 PM PST up reply actions 0 recs

Anyone count cards?

This sort of issue--the relative value of an enhanced methodology in practical terms--comes up all the time in the blackjack card counting discussions I occasionally follow. There are all sorts of methods and levels of detail a would-be expert can absorb, and of course it's all about enhancing your advantage and leveraging bet levels at the corresponding right moment. But in the final analysis learning and using a four-level counting system with 200 play indices doesn't necessarily add all that much win expectation over a much simpler approach.

Anyway, andeaux' note at the end again reminded me how much these statistical worlds overlap. And then there's monkeyball's "standard deviation."

Costly Seat Downgrade

by FreeSeatUpgrade on Dec 29, 2005 3:26 PM PST up reply actions 0 recs

I'm sorry, but ...

... all of my deviations are nonstandard. (I picked them up from an itinerant Romanian tinker who had Gerry-rigged a set of tools to fit slipshod Soviet-built devices.)

@('.')@

by monkeyball on Dec 30, 2005 8:44 AM PST up reply actions 0 recs

RC

Slg% and OBP% are both flawed if you really want to do well.

Not all outs are equal, and not all non-outs are equal.

A home run is worth only 3x as much as a single, though in Slg% it is worth 4x as much. A single is worth more than a walk.

THT Annual has the values for each listed. I'd post them but my copy is out in my car and I'm lazy. That is the way to do a RC formula. Check Kendall vs Payton that way.

by Yanny on Dec 29, 2005 9:35 PM PST reply actions 0 recs

SalB or Someone?

Can you post numbers or a link or something that shows that OBP is more important to run-scoring than SLG, because I happen to agree with Nico on this one. In fact, I remember reading that passage in Moneyball and thinking to myself: "Man, DePo has really overvalued OBP here."

Take this extremely simplistic example: Would you rather have Team 1 or Team 2 below?

Team 1: 350 OBP, 420 SLG, 770 OPS, 490 OPS/3
Team 2: 320 OBP, 500 SLG, 820 OPS, 487 OPS/3

According to OPS/3, we'd rather have a team of patient hitters with little power than a team of slightly less patient hitters with significantly more power. I realize that you don't score unless you get guys on base, but you also don't score if you can't drive guys in... Maybe it's just a matter of weighting it correctly, but if there is some article out there that shows the correlation between OBP and R, and the correlation between SLG and R, I'd love to read it. Anybody? Anybody? Bueller?

by Uncle Charlie on Dec 30, 2005 1:44 PM PST reply actions 0 recs

You can start here

But I will warn you, this is shameless self-promotion on my part.

http://www.beyondtheboxscore.com/story/2005/11/30/225611/12

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 30, 2005 1:53 PM PST up reply actions 0 recs

That's good stuff.

I think the discussion afterwords really hits at the big value problem BB> or <1B. I was particularly intrigued by the removal of IBB (sorry, I forgot who mentioned that) and the cost of a single to ISO. I like the idea of the "Dashboard Metric" and I hope there is some agreement on weights and simplification of league norms.
Damn, now I'm going to have to read that blog too. I take it all back, you suck.

Barry and the "Intangibles"

by Duke of left field on Dec 30, 2005 10:13 PM PST up reply actions 0 recs

By the way...

... I think it also important to evaluate the philosophies of the Angels and the A's in two different contexts: regular season and playoffs.

In the regular season, I think the A's philosophy of "nothing, nothing, nothing, petience, erupt" is OK because over the course of nearly 1,500 innings you are going to have enough eruptions to balance all of the "nothing" innings.

However, come playoff time, when an entire series can be made up of as few as 27 or 36 innings, I think the "nothing, nothing, nothing, erupt" offensive philosophy is a liability. The team with a philosophy geared toward bettering their odds of scoring one run in an inning is better-suited for the playoffs than the team that gears up for one big inning, because (a) they keep pressure on the opponent and (b) you never know when that eruption is going to come.

by Uncle Charlie on Dec 30, 2005 1:51 PM PST reply actions 0 recs

Plus it comes less often

against really good pitching, which you see more of in the playoffs.

One of the reasons I can't get more amped about OBP is the knowledge that .320 is considered a poor OBP and .350 a solid OBP, and yet the difference is all of one plate appearance every 33, aka one plate appearance a week.

Jason Kendall gets hit by a pitch about once a week, but that alone doesn't elevate him--in my mind--from a poor offensive player to a good one, in one swell foop. The fact that almost every single success gets him only one base, and that so many of those don't advance anyone anywhere (certainly not two bases at a time), seems as--if not more--significant.

So a 3:1 weighting in favor of OBP just doesn't seem right to me.

by Nico on Dec 30, 2005 2:08 PM PST up reply actions 0 recs

adding up the numbers

.320 is considered a poor OBP and .350 a solid OBP, and yet the difference is all of one plate appearance every 33, aka one plate appearance a week.

Let's go with that thought. How do other stats compare? A .240 batting average (poor) and .270 batting average (solid) amount to one extra hit every 33 ABs. Does that defy logic? I don't think so; it's just a reflection of the fact that very small differences in skill between players add up to big differences in wins over the long haul - which is exactly why you need good stats to evaluate players accurately.

Now let's compare to SLG. What does a 0.030 improvement in SLG amount to? It means, roughly, once every 33 ABs getting a double instead of a single. Those who weight OBP more heavily than SLG are saying, the difference between a double and a single is not nearly as big a difference in value as the difference between getting on base and making an out.

If a weighting of 3x is used, then the analyst is saying that an 0.030 improvement in OBP is equal in value to a 0.090 improvement in SLG - for example the difference between 0.400 (poor) and 0.490 (solid). Does that make sense? Well it amounts to once every eleven at bats (in which on average there are a few hits) getting a double instead of a single.

Is getting a double instead of a single every 11 at bats the same value as getting a walk instead of an out every 33 plate appearances? My gut feel, like yours, is that that seems a little bit too much weight on OBP. Most sabermetricians do put the weight closer to 2x, and that feels about right.

by matthias on Dec 30, 2005 3:57 PM PST reply actions 0 recs

Maybe we don't

really disagree (how dull!), because a 2:1 weighting seems fine to me. 3:1 just seems like too much.

Anyway, my issue is more with versatility and balance--of any preferred approach. It's something I think plagues the Angels too, because they only know how to give up too many outs and not be patient enough, while the A's only know how to take the opposite approach. Both approaches may work for their respective teams a good 75% of the time or so, but 40 "other games" is a lot...

by Nico on Dec 30, 2005 4:48 PM PST reply actions 0 recs

yeah... math...

sounds great. i like when the baseball men hit the baseball, then run real fast.

by kvn on Dec 30, 2005 5:29 PM PST reply actions 0 recs

current thinking

Current thinking is that OBP is about 1.8 times as valuable as slugging.
THT does a stat called GPA with the formula (1.8*OBP + SLG)/3. It's essentially OPS with more weight on OBP reweighted so that it looks more like batting average.

by Ellis on Dec 30, 2005 8:12 PM PST reply actions 0 recs

Now 1.8

I can get jiggy with. OBP is definitely important, and was undervalued circa Moneyball, so props to the A's for recognizing this. I'm not anti-OBP, just not convinced it's anywhere near thrice as important as the almighty slug.

by Nico on Dec 30, 2005 8:56 PM PST up reply actions 0 recs

Funny you should mention it

Everybody who was concerned that Moneyball would reveal our deep secrets was a few years too late. There is significant evidence that the market corrected itself w.r.t. OBP in the late nineties. Beane allowed Lewis to talk up OBP only after the fact.

Copernicus felt the same way about the geocentric crew.

by salb918 on Dec 30, 2005 10:30 PM PST up reply actions 0 recs