From the "What Goes Round Comes Round Dept":

Mike Scioscia has rediscovered the joys of Whiteyball.

Take a bunch of speedy hitters with a little pop who don't walk as much as the manager likes, but create mayhem once aboard.

Anchored with one or more genuine power threats, (for Jack The Ripper, read Bad Vlad)

Take the extra base when it's there; don't swing for the fences when there are men in scoring positions; have a flexible bench and a bullpen that can get you the best matchup you can get, but don't micromanage.

This is all in Herzog's book, mostly in the section : "You gotta have a plan"

Mike Scioscia has rediscovered the joys of Whiteyball.

This is only about 15 years to late for the Angels from when Whtey first tried it...

1992 Angels

13 years to be exact. That was a disaster. They won 72 games but to look at the roster you'd think they should lose 110.

Whitey left his best stuff in St Louis. His idea of building the Angels was this: "Hubie Brooks and Von Hayes played pretty well against me in the mid 80's. I think I'll sign them"

Brooks was a DH with a .247 on base average.

results in an offense that while less effective overall, produces a higher rate of two-out runs relative to the league average and a better offensive team such as the Yankees

This makes no sense -- do two-out runs count more than other runs? Does having an offense that's "less effective overall" make you "a better offensive team"?

Jeff often writes some pretty good articles and I generally respect his sabermetric knowledge, but the premise in this one is over the top. Since we have no evidence (to put it mildly) that any player or team has more or less of an ability to hit much better or worse than any other player or team, we have to conclude that the Angels have a mediocre offense (which they do) and that they just got a little lucky this year as far as hitting with RISP. If someone wants to give accolades to the team or to Scoscia for that, that's fine by me. Let's not, however, call this an organizational philosophy, or something like that. That's silly. In fact, since Jeff's webiste is all about giving advice to companies and corporations, how about this: If you give your organization credit for something that worked in the past by sheer luck, but is not an optimal strategy, you are dooming yourself in the future. Many times you have to look at the efficacy of strategies independent of the results, when there is much "noise" in those results, compared to the subtle difference between good and bad strategies in the short run. That should be the lesson. Exactly that scenario happened with the 2002 Angels. They had a medicore team and won the WS, presumably with the help of a lot of luck. They chose to do nothing with their roster (they thought they were already a great team - after all, they just won the WS) and the next year they finished almost exactly where they should have (3rd place with a sub-.500 record).

BTW, how does Jeff know how many of their stolen bases were on busted hit and runs, and I have never heard of a batter deliberately not swinging when the hit and run is on because he thinks the runner is going to be safe!

Oh, and even the "gaudy" stolen base totals and success rate for the Angels doesn't add a whole lot in terms of marginal wins to a team. And yes, fast runners usually means better baserunning, which translates to a few extra runs a year, again, not much, but maybe .5 to 1 extra win a year for a particularly fast team overall, although with Anderson and Molina and even Vladdy with his oft-injured legs, I'm not sure that the Angels are as fast overall as they used to be...

Second paragraph above should be "evidence that no one player or team...with RISP"

Exactly that scenario happened with the 2002 Angels. They had a medicore team and won the WS, presumably with the help of a lot of luck.

Luck, as in Barry Lamar in LF and a rock star wannabe at the plate.

The Angels are here (in the playoffs) because of their pitching/defense, and because the rest of the AL west sucked again (yes, even Oakland sucked for 2/3 of the season). They are the White Sox with even less power in the lineup, and a manager who's not as prolific in the use of obscenities but who calls for just as many dubious plays on the field. And it pains me to say this, as I really would like to see the Yankees gone, but the Halos will be truly lucky to play in game 5...

Let's not, however, call this an organizational philosophy, or something like that. That's silly.

Not if that is their organizational philosophy. If that's the case, maybe you'd rather argue that the philosophy itself is silly....

Also, certainly some of the value of Scioscia's use of the SB comes in suppressed GIDP totals. With a team full of GIDPs waiting to happen (Molinas, Vlad, G.A., Guillen & Glaus last year), Scioscia has kept DP numbers lower than expected, I think.

BPro's latest [url=http://www.baseballprospectus.com/statistics/postseasonodds.php]postseason odds{/url]:

Angels and Yankees are at 50-50
White Sox and Cardinals are 9 to 1 favorites
Astros with a slight advantage

Some totals vs. RISP2 numbers in the Scioscia era....

<u>Year  BA    Rk    RISP2 Rk    OPS   Rk    RISP2 Rk</u>
2000  .280  5     .263  4     .825  3     .808  1
2001  .261  11    .220  14    .732  10    .677  13
2002  .282  1     .257  5     .773  5     .759  5
2003  .268  7     .244  9     .743  10    .694  12
2004  .282  1     .273  1     .770  9     .786  5
2005  .270  6     .279  1     .773  11    .768  3

that link, again

A good discussion on "luck" and "Noise" is being talked about on the Voros thread of Playoff probabilities. I agree that it is far too easy to just dismiss hitting with men on as 'luck' or to claim it as 'noise', or to claim that the same results could happen with 'coin flips'.

I use to argue with guys that players like Eddie Murray were better than what their sabermetric values said they were, because they provided much value with men on hitting. Sabermatricians always referred to standard deviation etc., and refused to include it in their player values.

Well, I told them that Sparky Anderson must really be happy now that the sabermatricians don't give Murray credit for the men on hitting. Because now all of those grand slams and three run homers that Murray hit are no longer worth anything more than the average home run according to the sabermetric formulas. So I guess Baltimore should have some of the wins they got from those performances erased.

1992 Angels

What's crazy about that team, that I just noticed, is that Rene Gonzales was their best offensive player.

Again, for emphasis:
Rene Gonzales was their best offensive player.

He may or may not have led them in VORP or WARP, but he did play in 104 games and led the team in OBP and SLG. Wow.

"Vladdy with his oft-injured legs" was 13/14 in SBs this year, and famously scored one game-winning run from second base on a bunt. I don't know if I'd call him a particularly good base-runner, though....

As for the busted H&R plays, I have my doubts. Teams that H&R a lot tend to have crappy base-stealers get lots of CS. With the exception of the terrible Juan Rivera (who Scioscia mistook for a base-stealer early in the season, and compled a 1/10 record), there were very few attempts by the clubfoots. There were only 15 total attempts by non-Rivera guys you wouldn't consider base-stealers; 9 if you consider McPherson a viable threat (and I do).

GIDP in the Scioscia era, along with OBP & SB/CS:

<u>Year   OBP  Rk  GDP  Rk   SB/CS    %</u>
2000  .352  6   126  11   93/52  64% 
2001  .327  9   109  14  116/52  69%
2002  .341  4   105  14  117/51  70%
2003  .330  8   125  9   129/61  68%
2004  .341  6   123  9   143/46  76%
2005  .325  9   123  11  159/56  74%

Seems the running game has really taken the bite out of GIDP. The eye-opener is that 2002 line -- 4th in OBP, dead last in double plays.

Man those '92 Angels had a putrid offense. Only 2 regulars with an OBP greater than .300? Decent pitching staff, I guess.

There's also a potentially great handle in there--any suggestions on puns involving Chuck Crim?

Chuck Crim Dog?

Exactly that scenario happened with the 2002 Angels. They had a medicore team and won the WS, presumably with the help of a lot of luck. They chose to do nothing with their roster (they thought they were already a great team - after all, they just won the WS) and the next year they finished almost exactly where they should have (3rd place with a sub-.500 record).

Player    At Bats in 2002       At Bats in 2003
Glaus           569                   319
Eckstein        608                   452
Erstad          625                   258
Fullmer         429                   206
Figgins          12                   240
Riggs             0                    61
DaVanon          30                   330
Owens           ---                   241

Can we please stop pretending that the 2003 team has anything to do with the 2002 team? Glaus, Erstad, and Fullmer combined for a 111 OPS+ in 2002 and a 110 OPS+ in 2003 -- the only difference was that those three guys had a total of 927 less plate appearances in 2003 (not to mention the defensive dropoff in center when Erstad went down). Another factor was Eckstein's slump -- his OPS+ dropped from 103 to 79, and Benji Gil being forced into a regular role for three weeks was awful (his OPS+ was 31 in 2003 after he had posted a 96 in a platoon role the year before). And Eric Owens was never supposed to start 62 games.

2003 reflects nothing on 2002, and the only thing it has to do with 2004 is that it allowed Legs Figgins and Robb Quinlan an opportunity to show they could play a little.

Also, certainly some of the value of Scioscia's use of the SB comes in suppressed GIDP totals. With a team full of GIDPs waiting to happen (Molinas, Vlad, G.A., Guillen & Glaus last year), Scioscia has kept DP numbers lower than expected, I think.



Overall they hit into a DP 13.3% of all opportunities and the AL average is 13.2%.

Of course they may have hit into more DPs if they hadn't been hitting and running, especially with some of the players you listed...

Oh, and the other shortstop that played when Eckstein went down was Alfredo Amezaga, who was terrible, and also had nothing to do with the 2002 team.

How bad were the injuries? Shawn Wooten played 133 innings at third base.

...and I have no idea how to properly format a table...obviously.

Let me see if I can fix that...

Matt,
very nice job on the data.

However, I'd cross-check that data against GB/FB data.

GDPs are a function of GBs more than RO1 (i'd guess).

Looking at post 15:

So, the Angels in 2005 had fewer baserunners than in 2004, yet they grounded into the same number of DPs? And that is supposed to show that Scioscia is really a genuis this year? And I fail to see how "the running game has taken a bite out of the GIDP"; the Angels are running more than ever before, yet are hitting in pretty much same number of double plays, and they've done so with even FEWER runners on base to begin with. I'm not drinking this kool-aid...

As teh Devil notes, teh LAA are about league average wrt DP OPPS - because a DP OPP is going to be defined as hitting a GB w RO1<2O.

So, the Angels in 2005 had fewer baserunners than in 2004, yet they grounded into the same number of DPs? And that is supposed to show that Scioscia is really a genuis this year?

Yeah, Levski, that's exactly what I said.

As teh Devil notes, teh LAA are about league average wrt DP OPPS - because a DP OPP is going to be defined as hitting a GB w RO1<2O.

Surprsingly, BP doesn't count only GB but include all times with a runner on 1st and < 2 outs.

The Angels are basically league average on G/F hit, though (1.19 to 1.26 league average).

Sorry, Matt, that was a bit caustic on my part. But you did write that

Seems the running game has really taken the bite out of GIDP.

Which I read as a sign of approval of Scioscia's managing, yet I fail to see the evidence for that.

But what is forgotten in all this talk of RISP (less so RISP2) is the fact that this is one of the key times when a strikeout is worse than other outs. Anyone have the percentages for the number of Angels that got to third with fewer than two outs who got home, as opposed to the league average? Those happy slappy contact hitters may miss out on the two run bombs, but you'll get some of those runs back with RBI groundouts and sacrifice flies...

This Devil's Workday -- use pre and /pre tags.

And can you or Chris please define those terms?

Waterloo, the bottom line is that the Angels did not have a very good team in 2002 (sabermetric talent-wise) and won the WS. The GM did almost nothing in the off-season and AFAICR, their initial starting lineup in 03 was almost exactly the same as it was in 02. I am presuming that they assumed that they had a great team since they had just won the WS. Unfortunately (for them, and most other teams), that is not the proper way to evaluate how good a team you have (your w/l or playoff record). They were not likely to do well in 03 going in because they did not have a very good team. How the team and distriution of PA's played out that year, I don't know, and is not important to my argument. Had each player had exactly the same PA's that they had in 02, the team likley would have been around a .500 or slightly better team, also from what I recall of their true talent at the end of the 02 season. That is my argument. If I mischaracterized it before by suggesting that the 03 team was like the 02 team (at the end of the 03 season), then accept my sincerist of apologies....

Of course they may have hit into more DPs if they hadn't been hitting and running, especially with some of the players you listed...

Well, when a guy steals second base, you no longer have a DP opportunity. So there might also be an effect of reducing opportunities. Is there any data that might relate to that?

mgl, If the Angels change their approach with RISP at the plate and on the base paths from their general approach, then isn't their overall batting #'s not applicable to RISP situations?

To make an extreme example, if a batter hits with his eyes closed with RISP and with them open w/o, and we saw difference in those averages..wouldn't it be pretty clear why?

This is the problem with taking macro numbers and applying them across all different situations evenly.

Levski -- Unless I'm even dumber than I think, the fact that the Angels *always* rank lower in GIDP than OBP suggests to me that Scioscia is successfully using the running game to suppress GIDP. The value of that activity is another question, but as a basic premise I don't see the argument against it.

Waterloo, the bottom line is that the Angels did not have a very good team in 2002 (sabermetric talent-wise) and won the WS.

The Angels had the best Pythagorean record in baseball in 2002, and also won the World Series. Which sabermetric talents did they lack?

Matt, you are looking at the ranks, I am looking at the raw numbers (team OBP, GIDP, total SB attempts). The way I look at it, the Angels can't control how other team do in the GIDP/SB attempt department; they can only control how the Angels do it. And the Angels of 2005, comprared to the Angels of 2000, have grounded into the same number of DPs, despite running more and having lower OBP. IMO, if I were assembling the Angels team, I'd look at the 2000 team and forget the 2005 version. And tell the 2000 guys to stop running...

It may not have been a great team, but as a neutral baseball observer from England, that was the most memorable team I've seen in 10 years of following the sport.

Not that this observation is worth anything at all, but I felt like making it.

(1.19 to 1.26 league average).


What is their number?

the fact that the Angels *always* rank lower in GIDP than OBP suggests to me that Scioscia is successfully using the running game to suppress GIDP

It suggests to me their BIP distribution (there are what- 3000 BIP on a team in a season) is the issue.

if I were assembling the Angels team, I'd look at the 2000 team and forget the 2005 version.

Well, I wouldn't want to go 82-80 again....

I see your point about the raw numbers comparisons, thanks. Two X factors about that -- 1) The numbers are also influenced heavily by personnel. I would have expected the 2004 squad especially to have led the league in GIDP, based on previous experience. And some of the guys who were on both the 2000 and 2005 teams have gotten a lot slower; and 2) Much less relevantly, I suppose there might be some small variation based on comparative league quality of defenders....

It suggests to me their BIP distribution (there are what- 3000 BIP on a team in a season) is the issue.

Keep in mind too that the Angels strike out less than anyone else.

Waterloo, the bottom line is that the Angels did not have a very good team in 2002 (sabermetric talent-wise) and won the WS. The GM did almost nothing in the off-season and AFAICR, their initial starting lineup in 03 was almost exactly the same as it was in 02. I am presuming that they assumed that they had a great team since they had just won the WS. Unfortunately (for them, and most other teams), that is not the proper way to evaluate how good a team you have (your w/l or playoff record). They were not likely to do well in 03 going in because they did not have a very good team. How the team and distriution of PA's played out that year, I don't know, and is not important to my argument. Had each player had exactly the same PA's that they had in 02, the team likley would have been around a .500 or slightly better team, also from what I recall of their true talent at the end of the 02 season. That is my argument. If I mischaracterized it before by suggesting that the 03 team was like the 02 team (at the end of the 03 season), then accept my sincerist of apologies....

I agree that Stoneman erred in not trying to improve the team between 2002 and 2003. He made some particularly poor decisions on the bench and ignored the fact that Aaron Sele was a total waste in the rotation.

I do not agree that the 2003 team should have been expected to regress quite as far as they did, or that their true talent level was just barely on the good side of .500.

Let's get EqA involved instead of OPS+, as it's better ...

              2002            2003
Player     EqA    Outs     EqA    Outs
BMolina   .206     335    .253     301   
Spiezio   .277     367    .266     395
Kennedy   .274     339    .263     344
Glaus     .277     438    .273     243
Eckstein  .264     465    .238     257
Anderson  .286     457    .293     444
Erstad    .247     460    .233     198
Salmon    .298     355    .289     390
Fullmer   .295     311    .298     148
Total     .269    3527    .269    2720

As a whole, and on a per-out basis, the starting lineup performed exactly as well in 2003 as they did in 2002. This was largely due to Bengie Molina making up for the regression of almost everyone else, however. But as his EqA has been .257 over 2004 and 2005, it sure seems like his 2003 is closer to his true talent level than his 2002.

The problem in 2003 was that 800 outs were given to lousy backups like Eric Owens and Alfredo Amezaga.

Levski -- Unless I'm even dumber than I think, the fact that the Angels *always* rank lower in GIDP than OBP suggests to me that Scioscia is successfully using the running game to suppress GIDP.

Well, another factor to consider is that, outside of 2000, the Angel OBP has been light on HR, so there have been more guys on first base than the average team with the same OBP.

Well, I wouldn't want to go 82-80 again....

That 2000 team had some lousy pitching...

Matt, rather than looking merely at GDP, I think you would want to look at GDP + CS. If you are avoiding double plays because the lead runner is getting thrown out on the previous play, the total effect is still the same.

I think you would want to look at GDP + CS. If you are avoiding double plays because the lead runner is getting thrown out on the previous play, the total effect is still the same.

To be most specific, I suppose it'd be best to add only CS to second base with 0 or 1 outs....

#  NAME              TEAM  LG  YEAR  OPPS  DP  DP%   NETDP
1. Adam Kennedy      ANA   AL  2005  72    4   5.6%  -5.51
2. Steve Finley      ANA   AL  2005  88    7   8.0%  -4.62
3. Darin Erstad      ANA   AL  2005  112   12 10.7%  -2.79
4. Chone Figgins     ANA   AL  2005  108   13 12.0%  -1.26
5. Casey Kotchman    ANA   AL  2005  30    3  10.0%  -0.96
6. Orlando Cabrera   ANA   AL  2005  88    11 12.5%  -0.62
7. Lou Merloni       ANA   AL  2005  3     0   0.0%  -0.40
8. Vladimir Guerrero ANA   AL  2005  131   17 13.0%  -0.30
9. Curtis Pride      ANA   AL  2005  2     0   0.0%  -0.26
10. Zach Sorensen    ANA   AL  2005  2      0  0.0%  -0.26
11. Josh Paul        ANA   AL  2005  9     1  11.1%  -0.19
12. Chris Prieto     ANA   AL  2005  1     0   0.0%  -0.13
13. Kelvim Escobar   ANA   AL  2005  1     0   0.0%  -0.13
14. Dallas McPherson ANA   AL  2005  38    5  13.2%  -0.02
15. Robb Quinlan     ANA   AL  2005  26    4  15.4%   0.57
16. Jose Molina      ANA   AL  2005  29    5  17.2%   1.17
17. Jeff DaVanon     ANA   AL  2005  42    7  16.7%   1.45
18. Garret Anderson  ANA   AL  2005  124   18 14.5%   1.63
19. Ben Molina       ANA   AL  2005  80    14 17.5%   3.44
20. Maicer Izturis   ANA   AL  2005  32    8  25.0%   3.77
21. Juan Rivera      ANA   AL  2005  68    15 22.1%   6.02

Okay, finally done.

his Devil's Workday -- use pre and /pre tags.

And can you or Chris please define those terms?

Thanks Matt!

OPPS = number of PA with less than 2 outs and a runner on 1B.

DP = obvious

DP% = DP / OPPS

NET DP = number of DPs they hit less than or more than what an average player would have with the same amount of OPPS.

Since we have no evidence (to put it mildly) that any player or team has more or less of an ability to hit much better or worse with RISP than any other player or team

This has been a topic of discussion on the SABR statistical analysis E-group the last few days (actually, the more general discussion has been about "clutch hitting", which brings a definitional problem into play as well).

The studies that have been done to date are based on the premise that a player's expected level of performance in the situation - his baseline, if you will - is his overall performance in all situations. A "good" hitter with RISP, by this definition, is one who performs better in RISP than he does overall; a "poor" hitter is one who performs worse with RISP than he does overall. But I think this may be the wrong baseline to use.

We know, from looking at the data, that there are situational variations in performance across groups of players; we know thanks to Bill James, for example, that players as a group tend to hit better with a runner on first base than they do with the bases empty. So what we need to do in evaluating situational performance - and what has yet to be done in any study that I've seen - is to set an appropriate expected performance baseline for the specific situation *first* (which is probably not the expected performance of that group of players in all situations), then compare the performance of players against that expectation. IOW, if the expectation is for a player to lose 10% of his value when hitting with RISP, for example, then a player who loses only 5% of his value when hitting w/ RISP is a good hitter in those situations.

I do not know whether such a study would find evidence of ability or not, and I'm not making a statement one way or the other. My only point here is that we should first be sure that we've set the correct baseline for comparison, because if we're making comparisons against the wrong baseline we can't be sure we've gotten the correct answer.

-- MWE

RE: mgl's #5

A lot of provocative & interesting expressions here. No time to deal with them all now, but I'll start with one.

mgl said:
how does Jeff know how many of their stolen bases were on busted hit and runs, and I have never heard of a batter deliberately not swinging when the hit and run is on because he thinks the runner is going to be safe!

I think you might have mis-read part of my assertion, when you suggested I suggested I knew how many. Let me quote myself:

A number of the Angel stolen bases occur when the manager has called a run-and-hit play where the batter either didn't make contact or allowed the ball to go by based on the belief the runner had the base stolen. <u>I can't tell you how many</u>, but trust me, the Angels know the number and success rate.

and I, for one, agree with me.

<u>ALSO</u>
My info may be dated, and it may no longer happen, when I played (obviously, not pro), I was instructed and trained to do just that. In the majors, there were plays in 1962 when Maurice "The Gangster of Glove" Wills had those 104 SBs and was followed in the line-up by Jim Gilliam. There was discussion in the papers about situations both where Gilliam batting behind him would put the ball in play if he thought Wills would get nailed and also let balls go by on run and hit calls if he thought it was an easy steal for Wills and a pitch he didn't think he could do much with. I think that's where coaches got the idea they needed to teach that Gilliam skill.

They did it back then; I saw a small handful of Angels plays this year where it appeared just those events happened. And as far as the Angels tracking everything that affects runners moving around the bases, I've seen the mass of print-outs that cover multiple data slices. Scioscia shared some of his numbers with me during the conversation, and they sent me more later. They are passionate (in the clubhouse and dugout...I don't know about upstairs) about numbers and reading them to support feedback and tactics.

To be most specific, I suppose it'd be best to add only CS to second base with 0 or 1 outs....

Is there anywhere you can get that info. without going through game logs?

Or at the very least, is there a site that lists SB and CS by base (like stole 2nd, stole 3rd, cs 2nd, etc.)?

This Devil -- Thanks. I think the main point to be made about your chart is that when Scioscia uses the running game, he's effectively removing the DP-OPP. (I'm guessing here that the runners-situation vis-a-vis batters is measured at the point when the batter *does* something, rather than where he started the count.)

Matt, I would certainly think (or hope) so. I should also add that OPPS also include situations with runners on 1st and 2nd and less than 2 outs and bases loaded with less than 2 outs, just so everybody's clear.

Mike, that's a good poing re: clutch. You said nobody's done that when looking at "clutch," but has anybody done something similar when looking at playoff numbers, since hitters are likely to be facing better pitching? Although sample size is obviously a big issue here...

A point about the '02/'03 squads, that's actually relevant to '05 -- the '02 team might have been the best in baseball, but it wasn't particularly deep, except in the post-season (due the addition of K-Rod and Figgins). What Stoneman learned in '03, besides the folly of standing pat, was that you need to have a better bench than Adam Riggs, Shawn Wooten and Bengie Gil; a better 5th starter than Aaron Sele, and even duplication at some spots, in order to absorb unexpected injuries. Refusing to give Figgins a starting job anywhere is definitely part of that....

we know thanks to Bill James, for example, that players as a group tend to hit better with a runner on first base than they do with the bases empty

What about the performance of a pitcher with runners on versus bases empty? If James looked at hitters he must have looked at pitchers. Seems that a pitcher will contribute to the success of hitters with RISP, because my guess would be that pitchers become worse with RISP, and for a number of reasons-- increased pressure of having runners on/RISP (both from threat of SB and of the run scoring), impact of pitching from the stretch, likelihood of BB or hit increasing after giving up BB or hit, etc.

I guess some of the effect of the above on a hitter's stats w/RISP could be offset by the likelihood that he'll be facing a fresh reliever or L/ROOGY type, and you have to give due credit to a hitter's ability to take advantage of RISP situations, but there are a lot of variables in play here.

Plus I maintain that some guys both try harder with RISP and do not try as hard with the bases empty. And by some guys I mean Garret Anderson.

Some AB/GIDP stats of guys who played with & without the Angels over the last four seasons. Bold indicates their results while with Los Anaheim.

<u>Player     2002 2003 2004 2005</u>
Guerrero   30.7 21.9 32.2 32.5
Guillen    18.5 30.3 40.4 39.4
Rivera     20.8 21.6 35.5 23.3
Cabrera    35.2 34.8 38.6 54.0
Finley     50.5 86.0 44.9 67.7
Eckstein   86.9 50.2 51.5 48.5
Glaus      47.4 39.9 34.5 76.9

With the exception of Glaus, that data seems to be pointing in the same direction....

MWE, great post in #47. I would only add that you should also look at the specific batter-pitcher matchup.

What you propose is also very important for a stat like Leverage Index, and I am not sure Tango did it properly. He writes that the highest leverage situation is bottom of the 9th, home team trails by 1, 1 out, runners on 2nd and 3rd. But in that situation, a walk is very likely, which would dilute the effect of a great player.

If James looked at hitters he must have looked at pitchers.

Actually, he did not. He tossed this off as a comment in his large rookie study, when comparing the performances of rookie players to matched veterans. And he only mentioned it when dealing with one specific situation - bases empty vs. runner on 1B.

Tom Ruane expands on this in his study of clutch hitting.

-- MWE

mgl said in #5

If you give your organization credit for something that worked in the past by sheer luck, but is <u>not an optimal strategy, you are dooming yourself in the future</u>. Many times you have to look at the efficacy of strategies independent of the results, when there is much "noise" in those results, compared to the subtle difference between good and bad strategies in the short run. That should be the lesson. Exactly that scenario happened with the 2002 Angels. They had a medicore team and won the WS, presumably with the help of a lot of luck. They chose to do nothing with their roster (they thought they were already a great team - after all, they just won the WS) and the next year they finished almost exactly where they should have (3rd place with a sub-.500 record).

I think the part I underlined seems like a seam in your overarching argument. I had suggested that for a strategy to work, it needed to be well-executed in a thorough and structured way, and that it needed to be a viable strategy, but didn't have to be THE optimal strategy. Your point is not pursuing the optimal strategy dooms you in the future. We disagree on this one.

EXCEPT...
it seems to me some of your thinking agrees with my side, at least that sub-optimal strategies can still lead to success. My interpretation of the overarching direction of your work is that change imposes vast variation on outcomes, that, for example, the 2002 Angels won the Series more from luck than deservedness/skill/accomplishment/true winning behavior. If I accept that point, it's hard for me also to accept at the same time that any strategy that's not the optimal one dooms you to failure. I think most of your work would support my side of the argument.

ON 2003
I'll argue 2003 was as much an outlier as 2002 for the Angels. "Whatever doesn't make you stronger, kills you", as Norma Jean Mortensen once said. You noted already they made no refinements, improvements in the roster, perhaps based on the delusion that they might get away with that and still be competitive. But that was an extraordinarily suicidal act for them, and then they got some pretty serious injuries, as Matt Welch already noted. If you buy into the idea 2002 is a fluke for Angel team performance because of exceptional factors, I think one has to buy into the idea that 2003, equally, is a fluke in the trend.

As I said in the weblog essay, I'm not sure it's the most durable approach, but I'm sure it's working. I think the approach is intrinsically less durable than a traditional OBA-based offense, but I think it's viable. In an extreme all-Darin-All-the-Time Model or without Vlad, I don't think it's a good one. When it comes together, and, as we can see, it does periodically, it's fierce National League baseball in an environment not used to facing it, another slight advantage. Like Paul Byrd's "Hurry-Up" uncommon six-second snippet of time between pitches, it undermines the concentration and usual skills of some opponents some of the time, while others have no problems w/it.

The Angels have all this intent. They pursue this viable-not-optimal (I believe) strategy based on a set of tactics they drill into players up and down the organization. Matt Welch's GIDP table in #54 is one example of what happens with intent: many players make small steps in the intended direction sometimes, others seem to be able to leap into it. And overall team results shift because of it.

When it's working, they're very tough. Not optimal, but they weren't in 2002, and they went home w/more hardware than True Value had.

Jeff, you and I are coming from different philosophies, so to speak. I'll leave it at that. I enjoy your articles and respect your expertise in business managment, or whatever you call your profession. BTW, and I work for a major league team, the batter in a hit and run is instructed to swing at any pitch except one in the dirt. Most of the time, he does not have enough time to discern a pitch in the dirt, so he generally swings at everything. There is no way he has the time (or can devote his attention) to discern how good of a jump the runner at first gets on a hit and run. This is a minor point though and nothing really to do with your article. Obviously a person or organization does not have to use an optimalstrategy to be successful. An optimal strategy, does, however, lead to optimal success, by definition. Of course you must define "success" as it may have different meanings for different entities (make money for the owener, win games and championships for the fans, etc.). Once we define success, however, optimal is optimal. A good strategy yields good results and a bad strategy yields bad results, again, by definition. What we have to be careful with, in many endeavors, as I already articulated is using the wrong thing to identify and evaluate our strategies and therefore our chances of success in the future. Using a team's w/l record is an extremely poor way to evaluate "strategy" (personnel, etc.). The results of one's blackjack way is another. The results of one's stock transactions is another. Again, anything with a significantly large uncertainty by chance as compared to the difference in expectancy among various strategies is a poor measure to use to evaluate one's efficacy...

blackjack "way" should be blackjack "play." I'm watching RJ get tatooed as I am writing these posts...

Plus I maintain that some guys both try harder with RISP and do not try as hard with the bases empty. And by some guys I mean Garret Anderson.

This was very prescient.

BTW, and I work for a major league team, the batter in a hit and run is instructed to swing at any pitch except one in the dirt.

Sorry, mgl, the fact that you *work* for a MLB team does not mean that you are smarter than anyone else who has an opinion about baseball and how the game is played. I'm not trying to be disrespectful. You're entitled to your opinions. But don't confuse *your* opinions with facts.

There is no way he has the time (or can devote his attention) to discern how good of a jump the runner at first gets on a hit and run.

If I was you, I would just say you are flat out wrong here. But I'm not you so I will say that I disagree that "there is no way (a batter) has the time to discern how good of a jump the runner at first gets on a hit and run" with *no way* being the operative words.

Exactly that scenario happened with the 2002 Angels. They had a medicore team and won the WS, presumably with the help of a lot of luck. They chose to do nothing with their roster (they thought they were already a great team - after all, they just won the WS) and the next year they finished almost exactly where they should have (3rd place with a sub-.500 record).

That is total BS. Scot Spiezio's season was a total fluke (at least his right handed hitting was), yes but NOTHING else was out of the question.

They had one of the best defenses of the current decade with Darin Erstad in centerfield and had enough offense to finish 4th in runs scored. They lead the league in both "runners reaching scoring position" and "hitting with runners in scoring position" and overall batting average. They also had IMO the best bullpen in recent memory. Scot Shields didn't even get in a game until the World Series that year (in the playoffs). The team lead the league in pythag w% and by the end of the world series they had the best winning percentage in all of baseball, all while playing in an extremely tough AL West with 3 teams winning over 90 games. If my memory serves me right this was the toughest division in baseball history. This current Angel team would get its hat handed to it by the 2002 team. How... the hell... did you get the idea that this was an average 'sub .500 team'???

As for 2003, the team obviously could've used some retooling, but they should have been fine without them. The 2003 team had huge injury problems, the most important, if not the most overlooked, was Darin Erstad's. His missing most of the season in center coincided with a huge drop in the teams DER and thus the teams era. Fullmer was the teams best hitter at bat for at bat in 2002 but he contributed nothing after his injury. Glaus had his first of two years of shoulder injuries that sidelined him almost the whole season. A minor note was that the teams bench was downgraded for the 2003 team, which didn't help when the injuries hit. You don't know how many times I cringed when Eric Owens found his way into the starting lineup for days on end in 2003. By the end of the year that team shared no resemblance to the one that won the World Series in 2002.

With the exception of Glaus, that data seems to be pointing in the same direction....

Not to me. That's one of the problems - you know what you want the data to say, so it says that. I don't care one way or the other, so I see it differently.

Finley doesn't "point that way".

It looks to me that each player has a typical range (I STRONGLY DOUBT that a player has a specific level of AB/GDP).

So:
Guerrero's "true level" is ~31-32 (3 of 4 seasons - with and without the Angels)

Guillen may be plateauing at ~40 (he did hte same on a different team)

Rivera's mark went down from last season (but appears to have a "normal" rate around 21-23). I'd say last year for Rivera was just a fluke.

Cabrera appears to have a rate around 34-38, but popped up to 54 with the Angels. But I can't tell if that is a fluke high (like Rivera's last year) or if it is real. It's definitely different from what appears to be his "other team norm", but I can't tell by the last season what it means.

Finley went 50, 87, 45, 68 - that's not a pattern at all. It certainly isn't evidence of anything teh Angels did.

Eckstein's true talent is ~48-51 - with a fluke high season in 2002. No change after leaving the Angels.

Glaus declined as he got injured (probably related to a weaker bat and more GBs). This season, more FBs, fewer GBs. But no real pattern with teh Angels - not even a "level" like the other players have.

That chart says nothing to me about what the Angels do wrt hitting into DPs.

I work for a major league team,

I'll read that as "I have access to information you do not", rather than an appeal to authority based on the fact itself

the batter in a hit and run is instructed to swing at any pitch except one in the dirt.

However, they don't always.

Most of the time, he does not have enough time to discern a pitch in the dirt,

This isn't close to true. Even with the H&R on, something like 95% of the time players don't swing at pitches in the dirt, so they have no trouble discerning that.

However, in a H&R call, they focus more on making sure they "protect", so this is true:

so he generally swings at everything.

But:
There is no way he has the time (or can devote his attention) to discern how good of a jump the runner at first gets on a hit and run.

And he couldn't tell in that time anyway.

Sure, if the runner got 5 steps before the pitcher started his motion, but typically, the runner has only taken three or four strides, and the hitter cannot tell how great of a jump it is at all.

The runner is moving - that's enough *especially* if the H&R is on, because the hitter knows he's been called upon to swing.

You may be thinking of a straight steal where it looks like the hitter checks his swing on a great jump, but those are different situations.

If I was you, I would just say you are flat out wrong here. But I'm not you so I will say that I disagree that "there is no way (a batter) has the time to discern how good of a jump the runner at first gets on a hit and run" with *no way* being the operative words.


Sure, Rich, and mgl does have tone issues, but you are picking a HUGE nit here.

We are always speaking in probabilities - that should *always* be understood (it isn't - not mgl either).

It's a brutally small chance (extremely unlikely) that the batter can tell. Better?

"Using a team's w/l record is an extremely poor way to evaluate "strategy" (personnel, etc.)."


So how do we evaluate "strategy" (personnel, etc.)? The Beane Count? Pythag standings? Some BP metric? Sim league stats (brought to you by Danny)?
I'll take a team's W/L record in a 162 game season over some hypothetical metric that you've created.

So how do we evaluate "strategy" (personnel, etc.)?

Well, I haven't read all the debates about 'luck' that have been taking place on this site recently, but strategy is best evaluated by probability of success, rather than actual success. If I take my monthly IRA money and put it in lottery tickets, and win the big prize, most people would still probably say it was a stupid strategy. I just got lucky.

However, the bald statement of the original quote does sabermetrics about as much good service as an interview with Mike Gimbel in the mid 1990s did.

We do know there's a differnce between a hit & run and a run & hit, don't we? Not tyring to sound snippy but I just want to make sure those in the discussion are aware of the distinction. And it does go some way to dispelling the notion that a batter cannot tell if a runner has a good jump or not.

Separately, I'm not quite comfortable with the short shrift given to the idea that batters do change their approch with runners in scoring position (esp on thrid) and more worringly that some won't accept that some batters are better at it than others. Working on a method to study the idea statstically is an inetersting exercize and worthy of debate, but claiming the idea isn't true because a satisfactory method may not exist is upsetting to me.

Sabermetrics is a wonderful concept and I definitely fall into the "supporter" camp, but sometimes it seems common sense leaves the room when the mathematical navel gazing takes over.

I should not have invoked the "I work for a team" rhetoric. I stand by my opinion about the hit and run. It is not fact. It is based on my observation and experience in speaking with baseball people. I could be wrong. BTW, I don't think there is any distinction between the hit and run and the run and hit anymore. There are only hit and runs and straight steals. Rich, chill out. I haven't gotten into any arguments with anyone for a long time. I have mellowed out. Thanks Chris.

BTW, I don't think there is any distinction between the hit and run and the run and hit anymore. There are only hit and runs and straight steals.

That I find interesting. What do you base this on, mgl? If its based on knowing signs or inside information, just give me a no comment and I'll understand. If its based on something that is not a trade secret, I'd be interested.

Because after you said this, I began thinking, and I'm inclined to agree. The "run and hit" to the best of my recollection is a strategy that is premised on having a superior base stealer at first. Moreover, it was used a lot when the run producers were guys without tons of power.

Based on the offensive explosition, I could see that a manager specifically calling a "run and hit" would be limited.

Nevertheless, we do see managers employ the "hit and run" And I would imagine the better base stealers may still have open green lights to take a base when they feel they can. So I would imagine there is still:

(1) Hit and Run - Batter must make contact on any ball they can, excluding dirt stuff, etc. previously discussed.

(2) Straight Steal, Take - Manager calls for the steal and the batter has a take sign. Probably not used often, but still in play.

(3) Discretionary Steal - The runner moves on his own accord, the batter treats the at bat as he normally would. (without entering the debate as to whether he will protect, albeit I agree with Angus).

I could see where the normal

(4) Run and Hit - Manager gives the steal sign, the batter hits only if its a good pitch.

May have fallen out of favor.

mgl -- I'm still waiting for you to explain how "the Angels did not have a very good team in 2002 (sabermetric talent-wise)." Is it just that hitting singles isn't a "sabermetric talent"?)

And as someone who has had the displeasure of facing 90 mph pitching (against Larry Casian, no less!), I can testify that it *is* possible for even a relatively talentless right-handed batter to notice if a base-stealer got a killer jump from first, and to alter his decision accordingly. This usually happens against LHPs with slow times to the plate. Also, catchers can notice this, too; which I believe is why sometimes they don't even leave their crouch to attempt to throw a guy out.

That's one of the problems - you know what you want the data to say, so it says that. I don't care one way or the other, so I see it differently.

I used "seems to be pointing to," instead of "it says that," for a reason.

That's good Matt, but do you see why I say "it doesn't "seem" to be pointing to"?

Yeah, it's on the borderline of usefulness. Still, if Glaus and Cabrera cancel each other out, that leaves you with Guerrero, Guillen and Eckstein all having their best GIDP rates with the Angels & their worst without, and Finley & Rivera having their second-best rates in their lone year with the Angels. Probably doesn't rise to the level of statistical significance, but it might just point slightly in one direction.

But I really have far less interest in arguing about it than I have in posting the table in the first place.

But I really have far less interest in arguing about it than I have in posting the table in the first place

Heh. fair enough.

There is no way he has the time (or can devote his attention) to discern how good of a jump the runner at first gets on a hit and run.

I've played against guys who threw between 80-90, and in my case there's no way I can alter my decision to swing based on the batter's jump. I pretty much tune out the batter and focus completely on the pitch.

Of course, I'm also devoid of talent. Since you work for the Cards, MGL, if you have access could you ask David Eckstein if this is possible?

He's pretty good on the H&R, knows something about how the Angels do things. I'd accept his opinion and let it stand.

I don't know if I'm the little boy noticing the naked emperor or just a fool missing something obvious, but it seems to me the discussion of why the Angel offensive philosophy is working misses the point that the Angel offensive philosophy is NOT working. With a position player payroll rank of 3rd or 4th in the AL in each of the past 6 years, their rank in terms of runs scored has been:

2000 7th (out of 14)
2001 12th
2002 4th
2003 11th
2004 7th
2005 7th

Overall, a team whose average offensive production ranks 5 places below its offensive payroll is doing a lousy job offensively. This does not seem to me to be even remotely a counterexample to orthodox sabermetric views.

Yes, the Angels have been successful in 3 of the past 6 years, but that is because of the conversation not held, about pitching. They made the playoffs only in the years that they ranked absolute best or second best in fewest runs allowed in the AL.

Given the Angel payroll, the offense has been a flop in 5 of the 6 years that the current management team has been operating.

What am I missing?