Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Thursday, February 18, 2010

WEEI: Keith Law on D&H: The 2010 Sox, the farm system, and ‘useless’ statistics

Nick from comments…

“Keith Law is a total moron. To say RBI means nothing is absolutely ridiculous.”

To recap: Theo Epstein and Keith Law, two guys who have worked in front offices/player personnel, say that RBI is a worthless stat.

Harvey from the internet says they are “morons”.

Do you really think that RBIs are useless, rather than just overvalued?

Totally useless. In terms of measuring the value of a player’s performance, I find them absolutely useless because 1) it’s determined by how many opportunities you get — the guys who hit in front of you in the lineup, how often did they get on base; and 2) there’s no particular skill to driving runs in. There’s no such thing as a hitter who is significantly better in RBI opportunities. Guys might do that over a year or two over the course of their careers, but you are not seeing guys who are just substantially better than the norm with runners in scoring position. Obviously all hitters hit a little bit better with men on base and pitchers working out of the stretch, maybe he doesn’t generate the same velocity. But in general, a hitter’s a hitter, whether there’s nobody on base or there are guys on second and third.

Some guys take advantage of those opportunities that are there for them, and some don’t. Some guys have a knack for driving people in.

I disagree with that. I do not think that’s true, that there are guys out there who take extra advantage of those opportunities. Over the course of the season, yes, absolutely, there will be players who will be significantly better than the norm in knocking in runs or performing in clutch situations, which has been a sabermetric debate for 20 years, whether there is such a thing as a “clutch hitter.” There’s clutch hitting, obviously, but is there an individual animal who you could call a clutch hitter? I fall on the side that says no, there’s not really a guy who’s better in the clutch. There’s not really a guy out there who’s better in RBI situations. If you’re in an RBI situation, if you’re in a clutch situation, the guy you want at the plate is just your best hitter, period – the guy who’s going to produce the most offensively or give you the least chance of making an out, because obviously in a clutch situation, in an RBI situation, the last thing you want is an out. So get me the guy up there who’s the least likely to make an out or who’s most likely to get that extra-base hit, regardless of what the situation is, because I think if you really look deep down into it, over the course of multiple seasons, you won’t find that those guys who you’re talking about who step up in big situations really exist.

Repoz Posted: February 18, 2010 at 03:21 PM | 89 comment(s) Login to Bookmark
  Tags: business, red sox, sabermetrics

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Golfing Great Mitch Cumstein Posted: February 18, 2010 at 03:41 PM (#3462615)
I am surprised that the "RBI's rule" crowd was shouted down on WEEI.com by a 6-2 margin. All the mouthbreathers must be waiting 45 minutes to call in.
   2. Dale Sams Posted: February 18, 2010 at 03:49 PM (#3462626)
And the ones over the course of their career who *are* statistically better in the clutch?

Mike Lowell OPS bases empty:.805
OPS RISP: .848

or is 43 points insignificant? I'm not taking sides, I'm just asking. I just looked up Lowell because I thought his 120 RBI's in 2007 were remarkable, and am a little hesitant to say RBI's are completly useless.
   3. RoyalsRetro (AG#1F) Posted: February 18, 2010 at 03:54 PM (#3462631)
Pat Tabler is the only evidence that "clutch" exists in my mind.

Nobody on: .261/.324/.353
RISP: .317/.388/.432
Runner at 3rd, 2 out: .434/.427/.566
Bases loaded: .489/.505/.693 (109 PAs)
   4. RB in NYC (Now Semi-Retired from BBTF) Posted: February 18, 2010 at 03:55 PM (#3462632)
or is 43 points insignificant? I'm not taking sides, I'm just asking.
Is career the right way to measure it? Does Lowell do it year-in-and-year-out or are there a couple of really good seasons bringing up his number? It seems like you'd have to look at it the numbers a couple of different ways.
   5. SoSH U at work Posted: February 18, 2010 at 03:59 PM (#3462634)
or is 43 points insignificant? I'm not taking sides, I'm just asking. I just looked up Lowell because I thought his 120 RBI's in 2007 were remarkable, and am a little hesitant to say RBI's are completly useless.


I'll go with insignificant, in that most guys hit better with RISP than bases empty. Besides the windup vs. stretch argument Keith lists, there is a huge advantage to hitting with guys on third base and less than two outs. Flyballs that would be outs in bases-empty situtations become at-bat less sac flies, and grounders that might result in a 6-3 are more likely to get through a drawn-in infield for a base hit.
   6. Shooty Survived the Shutdown of '14! Posted: February 18, 2010 at 04:00 PM (#3462635)
I don't think RBI's are useless, but their importance is so overstated by the pro-RBI crowd and their usefulness is so subtle and hard to make use of, it's probably better to just ascribe zero value to them for the time being. They're useful as an historical record of what happened but as a tool for valuing players...it's probably not worth the effort to ferret it out at this point.
   7. depletion Posted: February 18, 2010 at 04:08 PM (#3462641)
Obviously all hitters hit a little bit better with men on base and pitchers working out of the stretch, maybe he doesn’t generate the same velocity.

There's #5 and there's also, "if the pitcher has just allowed one or more baserunners, then he is not as sharp as a pitcher who's been mowing down the lineup."
   8. Steve Treder Posted: February 18, 2010 at 04:08 PM (#3462642)
most guys hit better with RISP than bases empty. Besides the windup vs. stretch argument Keith lists, there is a huge advantage to hitting with guys on third base and less than two outs. Flyballs that would be outs in bases-empty situtations become at-bat less sac flies, and grounders that might result in a 6-3 are more likely to get through a drawn-in infield for a base hit.

Yes, and moreover there's a significant measurement bias at work, in that bases-empty situations are likely to be caused by a better pitcher on the mound and/or a pitcher having a better day and/or a better pitchers' park, and RISP situations just the opposite.

For all these reasons, the entire league hits better with RISP.

EDIT: Coke to #7.
   9. GGC don't think it can get longer than a novella Posted: February 18, 2010 at 04:12 PM (#3462644)
Don Hurst would like a word with these two.
   10. Dale Sams Posted: February 18, 2010 at 04:12 PM (#3462646)
For all these reasons, the entire league hits better with RISP.


Except for Derek Jeter.
   11. The Good Face Posted: February 18, 2010 at 04:18 PM (#3462652)
I don't think RBI's are useless, but their importance is so overstated by the pro-RBI crowd and their usefulness is so subtle and hard to make use of, it's probably better to just ascribe zero value to them for the time being. They're useful as an historical record of what happened but as a tool for valuing players...it's probably not worth the effort to ferret it out at this point.


Pretty much. I've long viewed RBI as a sort of indicator of good players, in the sense that if you look at the top 20 career RBI guys in ML history, you're going to be looking at some awfully good players, but of virtually no use for assigning value.
   12. Randy Jones Posted: February 18, 2010 at 04:21 PM (#3462656)
Pretty much. I've long viewed RBI as a sort of indicator of good players, in the sense that if you look at the top 20 career RBI guys in ML history, you're going to be looking at some awfully good players, but of virtually no use for assigning value.


You could say the same thing looking at the top 20 in career PA's or AB's though.
   13. WallyBackmanFan Posted: February 18, 2010 at 04:25 PM (#3462658)
Pretty much. I've long viewed RBI as a sort of indicator of good players, in the sense that if you look at the top 20 career RBI guys in ML history, you're going to be looking at some awfully good players, but of virtually no use for assigning value


Except for losses, if you look at the top 20 career list of any counting stat you'll get a list of mostly great players. You have to be considered good to get that much playing time.

Here's the top 20 career totals for outs made:

1. Pete Rose 10328 B
2. Hank Aaron+ 9136 R
3. Carl Yastrzemski+ 9126 L
4. Cal Ripken+ 8893 R
5. Eddie Murray+ 8569 B
6. Rickey Henderson+ 8510 R
7. Dave Winfield+ 8422 R
8. Robin Yount+ 8415 R
9. Brooks Robinson+ 8340 R
10. Craig Biggio 8272 R
11. Luis Aparicio+ 8110 R
12. Willie Mays+ 8056 R
13. Paul Molitor+ 8040 R
14. Rabbit Maranville+ 7906 R
15. Omar Vizquel (42) 7896 B
16. Rafael Palmeiro 7858 L
17. Lou Brock+ 7823 L
18. Ty Cobb+ 7748 L
19. Stan Musial+ 7704 L
20. George Brett+ 7673 L

EDIT: Coke to #12
   14. aleskel Posted: February 18, 2010 at 04:28 PM (#3462662)
I think there's a certain tautology about RBIs that, in some way, makes them an okay shorthand. If you drive in a lot of runs it probably means you hit in the middle of a lineup, and if you hit in the middle of a lineup chances are you were a good hitter.
   15. Jose Can Still Seabiscuit Posted: February 18, 2010 at 04:28 PM (#3462663)
Actually losses wouldn't be awful either. The top 20 features 13 Hall of Famers plus a few guys with good cases (Blyleven, John, Kaat).
   16. snapper (history's 42nd greatest monster) Posted: February 18, 2010 at 04:29 PM (#3462667)
Except for losses, if you look at the top 20 career list of any counting stat you'll get a list of mostly great players.

Including Losses.

The top 7 all time in losses are HoFers. 13 of the top 21 (tie for 20th).

The others include Blyleven, Tommy John, Jim Kaat, Frank Tanana and Bob Friend, plus a few real old timers whoe were at least above average and long lasting.

Edit: Coke to Jose. Serves me right for providing detail.
   17. Jose Can Still Seabiscuit Posted: February 18, 2010 at 04:34 PM (#3462675)
Keep the Coke snapper, I gave up soda for Lent. I hope you learned your lesson about providing detail though, don't let it happen again!
   18. GGC don't think it can get longer than a novella Posted: February 18, 2010 at 04:35 PM (#3462676)
So that's how that effects Frank Tanana. Leaders in walks and homers given up aren't too shabby. I didn't realize that Jamie Moyer could break Robin Roberts record if he pitches this year.
   19. DCA Posted: February 18, 2010 at 04:42 PM (#3462681)
I didn't realize that Jamie Moyer could break Robin Roberts record if he pitches this year.

Wow, that's right. That would be awesome. Not as awesome, though, as Moyer pitching 3-4 more years and getting 300 wins.
   20. Eric in Madison Posted: February 18, 2010 at 04:43 PM (#3462682)
So let me put this question to the group: what about "clutch pitching?" I actually posed this question to Bill James on his site, and he suggested that it was easier to imagine a "clutch pitcher" than a "clutch hitter" because pitching is proactive whereas hitting is essentially reactive.

It came up because I was looking at Nick Blackburn's stats, and trying to figure out why his ERA was better than his peripherals would suggest. Turns out that the reason is that his results with RISP are way better than his results overall. (774 OPS against overall; 665 OPS against with RISP).

My assumption was (and is) that this is merely a function of luck. James said it could be luck, but there could be other factors when it comes to pitching. What say you?
   21. heyyoo Posted: February 18, 2010 at 04:43 PM (#3462683)
One of these days I'll need to look closer at BP's RBI opportunities report. You can sort guys by OBI%.

Others Batted In Percentage -- the fraction of runners on base who were driven in during a batter's plate appearances. OBI is distinguished from RBI (runs batted in) in that OBI does not credit the batter for his own scoring on a home run. In otherwords OBI = RBI - HR


It would be kinda cool to run through that and see which, if any, guys show up in the top 10 or 20 with any consistency, and also how far above or below average a guy is in this stat in relation to how far above or below avg he is in say something like OPS.

OBI report
   22. Ron Johnson Posted: February 18, 2010 at 04:44 PM (#3462686)
For all these reasons, the entire league hits better with RISP.


That's mostly selection bias (non-hitters are far more likely to be hit for) plus the fact that almost all IBBs happen with RISP and that there's no reliable way to hit a SF without a runner in scoring position.

When you control for these factors the difference is not large.

I did a study of 187 players with 1,000+ PA with RISP. They hit .275/.346/.432/.340 overall and .278/.369/.430/.345 with RISP.
(The last number being OBP with IBB removed. Didn't think to adjust for SF. If I had ... well you're talking ~5 points of OBP versus maybe 7 points of SLG. It's a difference that would take a long time to manifest itself in runs scored)

To get to the topic at hand, there are players who drive in more runs than you'd expect (and #2, Mike Lowell is one of them -- not many others. Joe Carter is one as well, as is Michael Young) but their teams don't tend to score more runs than you'd expect given the counter stats. Pretty strong indication that "rbi alility" isn't of any particular value.

Arne Olsen did a study of these guys. Title of the study ("rbi vultures") sums up the results. Basically guys who are unusually good at driving in runs tend to be relatively poor at creating opportunities for teammates. So it pretty much comes out as a wash. I suspect that they'd have a little extra value batting 6th in a 6 hitters plus passengers type of offense (which isn't exactly uncommon). The fact that they're relatively weak at creating opportunities is at least partially mitigated by the guys who are at the plate. But it's no big deal.
   23. SouthSidePat Posted: February 18, 2010 at 04:44 PM (#3462687)
I agree that RBIs are completely useless for evaluating players and their performance. However, they do tell the story of what happened in any indivual game. For that reason, I still like them as a box score stat.
   24. BillP Posted: February 18, 2010 at 04:54 PM (#3462695)
My assumption was (and is) that this is merely a function of luck. James said it could be luck, but there could be other factors when it comes to pitching. What say you?

I think it's pitching out of the stretch. As a hitter, your job is (virtually) always the exact same--avoid making an out, while advancing as far as you can toward getting back to home plate. As a pitcher, you're actually doing two pretty different things in pitching from a windup vs. pitching from the stretch.

It's a lot easier for me to believe that some pitchers (like maybe Blackburn) are better at making that transition than others than it is for me to believe that some players are better at handling "pressure" than others, when you consider that ALL players have, at least, tens of thousands of pairs of eyes focused on everything they do in their professional capacity.
   25. SoSH U at work Posted: February 18, 2010 at 04:56 PM (#3462698)
So let me put this question to the group: what about "clutch pitching?" I actually posed this question to Bill James on his site, and he suggested that it was easier to imagine a "clutch pitcher" than a "clutch hitter" because pitching is proactive whereas hitting is essentially reactive.

It came up because I was looking at Nick Blackburn's stats, and trying to figure out why his ERA was better than his peripherals would suggest. Turns out that the reason is that his results with RISP are way better than his results overall. (774 OPS against overall; 665 OPS against with RISP).

My assumption was (and is) that this is merely a function of luck. James said it could be luck, but there could be other factors when it comes to pitching. What say you?


Glavine, for one, seems to have the track record to support the idea of the clutch pitcher.
   26. Famous Original Joe C Posted: February 18, 2010 at 04:58 PM (#3462700)
120 RBI's in 2007

Not so remarkable for a guy who hit over .320 with a crapload of doubles while hitting behind a bunch of high OBP hitters.
   27. Ron Johnson Posted: February 18, 2010 at 05:00 PM (#3462701)
#20, I'm not aware of any systematic look at clutch pitching, though Total Baseball did have a clutch ranking which was basically predicted runs allowed divided by actual runs allowed.

As James suggests, there are enough pitchers with odd situational stats to suggest that it "should" exist.

To pick a guy at not random: Nolan Ryan

Overall: .204/.307/.298
Empty: .191/.298/.279
Runners on: .221/.320/.325

Throw in free bases via wild pitches, poor DP support and not being very good at controlling the running game ... Ryan didn't do well in TB's clutch ranking IIRC.
   28. bookbook Posted: February 18, 2010 at 05:00 PM (#3462702)
I always found "clutch hitters" interesting:

Because amps don't actually go to eleven in any way that means anything, any clutch hitter can just as profitably be thought of as a non-clutch slacker.

Do you really want to celebrate the guys who don't give their all when the game's not at a critical juncture (or when the bases are empty)?
   29. Ron Johnson Posted: February 18, 2010 at 05:05 PM (#3462707)
#26, my rbi estimator came up with 100 for Lowell's 2007. Given that the standard error would be around 9 given his opportunities (299 is high but not extreme) ... well most errors are explained by baserunner distribution. I suspect he happened to have an unusual number of multiple baserunner ABS.

It's still on of the larger errors I've seen. And he has driven in more than you'd expect over the course of his career, not just 2007.
   30. PreservedFish Posted: February 18, 2010 at 05:13 PM (#3462724)
As James suggests, there are enough pitchers with odd situational stats to suggest that it "should" exist.

To pick a guy at not random: Nolan Ryan

Overall: .204/.307/.298
Empty: .191/.298/.279
Runners on: .221/.320/.325


The old argument is: I don't believe in clutch players, but I do believe in "unclutch" players, because it is much easier to imagine a person having their skills suffer due to anxiety (extremely common in regular life) than it is to imagine a person actually sharpening their abilities and improving in the same situation.
   31. Mayor Blomberg Posted: February 18, 2010 at 05:19 PM (#3462730)
Keith may have a second career as Robbie Cano's agent.
   32. heyyoo Posted: February 18, 2010 at 05:21 PM (#3462734)
The old argument is: I don't believe in clutch players, but I do believe in "unclutch" players, because it is much easier to imagine a person having their skills suffer due to anxiety (extremely common in regular life) than it is to imagine a person actually sharpening their abilities and improving in the same situation.


Perhaps, but even this is overblown. Players have to deal with pressure and anxiety thoughout their amateur career. A kid in a baseball tournament with college and pro scouts in the stands with radar guns knows exactly what is going on. All thats at stake THAT DAY is his entire future, not to mention the fate of their team's outcome on that particular day.

Then, the ones that pass through this crucible go on to the minor leagues where every day is tryout in an attempt to get to the next level. By the time these players reach the major leagues, they have been tested and pushed and faced tremendous pressures. For the vast majority of them, it was a fight for survival every step of the way, as only a very few were given signing bonuses that truly represent financial security.

Once they hit the show, there are new pressures, and sure, 50,000 screaming fans in a playoff scenario probably gets to some guys. But for the vast majority, they already have the ability to perform under pressure, or they would not have gotten there in the first place.

Like Tim Kurjian said one night, "These guys are not like you and me".

(Note: It was a different context, TK was talking about Edgar Gonzalez stepping back in the box after his bad beaning, but the point remains the same)
   33. Ron Johnson Posted: February 18, 2010 at 05:22 PM (#3462736)
#30, need not be any form of anxiety. Seems to me that it could be as simple as not pitching as well from the stretch.

In fact one plausible explanation for the run of success Alou/Kerrigan had with pitchers who had washed out in other organizations is that they didn't sweat the whole pitch different with runners on. I recall Alou explicitly saying that the pitcher's job was to get the batter out and if that sometimes made things tough on the catcher, too bad.
   34. Eddo Posted: February 18, 2010 at 05:29 PM (#3462747)
I wouldn't go so far as to call RBI* as "useless" statistic; it's useful in a different sense.

It's a descriptive statistic, in that it adds some data as to which players actually provided value in a given game, season, or career. Of course, RBI often come about due to a favorable situation over anything else, but the value was still provided by the player, given the opportunity.

Because of the situational aspect to RBI, we all know that it has little value in projecting players' values moving forward. It also provides little evidence to suggest any given hitter is better than any other given hitter. It's not a predictive statistic, not even a little bit.

I think what creates the conflict is that people like Law and Epstein are speaking a slightly different language than the person doing the questioning. When Law or Epstein uses the term "value", he means something along the lines of "true ability" or "what Player X should be paid this coming season".

The questioner, however, simply sees that there was "value" in driving in the runs a hitter did; you need to drive in runs in order to win games, after all. He's not necessarily trying to say that the RBI make Mike Lowell, hitting behind OBP machines, a better hitter than Hanley Ramirez, hitting in a less-favorable situation. He's just saying that Lowell's RBI did indeed provide value.

In some cases, this conflict could be alleviated a bit by the Laws and Epsteins avoiding calling RBI "totally useless", and instead explaining that RBI have no value in determining if one hitter is worth more than another moving forward. Of course, there are people on the "pro-RBI" side that won't even accept that fact, but it is totally useless to discuss baseball with them most of the time, anyway.

* Pet peeve: the plural form is "RBI", which stands for runs batted in. Saying "RBIs" is superfluous. It would be like saying someone has "6.5 WARs".
   35. jacksone (AKA It's OK...) Posted: February 18, 2010 at 05:31 PM (#3462751)
#20 - I hereby charge you with watching all of Daisuke Matsuzaka's starts in 2008 to try and determine his clutchiness.
   36. Cowboy Popup Posted: February 18, 2010 at 05:33 PM (#3462755)
Because amps don't actually go to eleven in any way that means anything, any clutch hitter can just as profitably be thought of as a non-clutch slacker.

I've seen this suggested before and I think its an easily refutable statement. We know that pitchers will change their approach with runners on and in big time situations. They are more willing to use more pitches in order to work around giving up a big hit. That alone suggests that hitting in clutch situations is a different animal than a normal at bat and it makes sense that some hitters would be better suited for it than others.

Beyond that, players definitely have periods of time when they are locked in and when they are not. Are players slacking when they hit .200 for two weeks and playing at full speed when they hit .500 for two weeks? Are you unwilling to admit that some times, players are in fact just seeing the ball better, anticipating the pitches better and are able to perform better for periods of time during the season due to conditions beyond their conscious control? Do you have any evidence to suggest anything to the contrary?

I'm also not sure why everyone in the ballpark and watching the game at home can feel the intensity of a clutch situation while ballplayers are supposedly impervious to it. I think its clear that players feel the energy of the moment and it can have a different effect on each player. I mean, just based on watching, it seems to matter a whole heck of a lot more to the players when they come through in big spots than when they come through in 13-5 games. Does that extra release of emotion when a closer gets a big K or a batter hits a walkoff only manifest itself after the play has happened?

Further, the players believe in it. The guys who go through this stuff and are watching others go through it a way we can't imagine have a firm belief that some guys respond better to the pressure than others. And the non-clutch camps response to that is wholly unconvincing. There's your "then those guys are slackers" argument and there's Keith Law's "based on a ridiculously small sample of a player's at bats, using broad generalities to determine what 'clutch' actually is without putting more than ten seconds of thought into how to narrow the sample, the results are inconclusive" argument. Neither has convinced me and I'm not surprised that it hasn't really convinced anyone else either.
   37. sunnyday2 Posted: February 18, 2010 at 05:38 PM (#3462761)
Of course there is clutch pitching. And its name is Jack M-;)rris.
   38. PreservedFish Posted: February 18, 2010 at 05:41 PM (#3462767)
#30, need not be any form of anxiety. Seems to me that it could be as simple as not pitching as well from the stretch.


The "old argument" to which I referred exists to explain choke artists. Armando Benitez, or whoever.
   39. Ron Johnson Posted: February 18, 2010 at 05:43 PM (#3462771)
#36, Back on usenet I used to have a standard request to posts such as yours. Provide me with your definition of clutch and I'll run the study.

I've looked at RISP (bottom line, it's not quite random. There are players who meet a definition of clutch by this method. It's just not very important)

I've looked at late/close (see above)
   40. PreservedFish Posted: February 18, 2010 at 05:46 PM (#3462777)
* Pet peeve: the plural form is "RBI", which stands for runs batted in. Saying "RBIs" is superfluous. It would be like saying someone has "6.5 WARs".


From the first Google hit:

Some people reason that since “RBI” stands for “runs batted in,” there is no need for an additional “S” to indicate a plural, and speak of “120 RBI.” However, though somewhat illogical, it is standard to treat the initialism as a word and say “RBIs.”

The same pattern applies to other such plural initialisms as “WMDs” (“weapons of mass destruction”), “POWs” (“prisoners of war”), and “MREs” (“meals ready to eat”)...


I think in this case the acronym has become so familiar that it has almost taken on its own life as a word, and thus is treated as one. It doesn't make sense but I wouldn't call it wrong because the use is so widespread and feels so natural.

Another difference between WAR and RBI is that RBI is very happily singular (Run Batted In) whereas the phrase "Win Above Replacement" is just about useless. This is also true of the other acronyms in the quote above, but not true of RPM, which the author gives as another acronym that does not take on an S to indicate plurality.
   41. Cowboy Popup Posted: February 18, 2010 at 05:49 PM (#3462779)
#36, Back on usenet I used to have a standard request to posts such as yours. Provide me with your definition of clutch and I'll run the study.

How about every time a batter comes to the plate and the crowd gets on its feet to watch the next play. I don't see this as the kind of thing you're going to be able to use retrosheet or simple situational stats to solve. That was my issue with Law's argument.
   42. Forsch 10 From Navarone (Dayn) Posted: February 18, 2010 at 05:59 PM (#3462796)
This is something I've wondered before--what's the "true" 100-RBI season? By that I mean, to traditionalists the 100-RBI season is, ipso facto, an indicator of quality. But we know that isn't true. Ruben Sierra had 101 RBI in 1993 and was pretty awful. Joe Carter in '90 had 115 RBI while toting around an OPS+ of 85. There are, of course, many other examples. So what's the lowest RBI total at which a player is necessarily productive in more meaningful ways? Stated another way, what's the lowest RBI total at which a player cannot be otherwise below average? Because of Carter, we know it's >115. 116? 120? 125? Maybe I'll poke around in the Play Index ...
   43. Eddo Posted: February 18, 2010 at 06:08 PM (#3462806)
Beyond that, players definitely have periods of time when they are locked in and when they are not. Are players slacking when they hit .200 for two weeks and playing at full speed when they hit .500 for two weeks? Are you unwilling to admit that some times, players are in fact just seeing the ball better, anticipating the pitches better and are able to perform better for periods of time during the season due to conditions beyond their conscious control? Do you have any evidence to suggest anything to the contrary?


What does this have to do with clutch ability? If clutch ability existed (I'm not saying it does or doesn't), then this would explain it not being constant. However, the fact that players don't play at one constant level has no bearing on whether or not clutch ability exists; one is due to natural fluctuations, the other would be due to an inherent difference in ability (or at least that's what the pro-clutch crowd is arguing).
   44. PreservedFish Posted: February 18, 2010 at 06:17 PM (#3462813)
However, the fact that players don't play at one constant level has no bearing on whether or not clutch ability exists;


I believe that this was a kind of building block argument. If we agree that players do not play at a constant level (which is to say, that they are not just Strat cards) then we can come closer to accepting that they will perform at different levels in different situations.

It doesn't prove that they can or will do so consistently, but if we want to prove it, it is a necessary first step to admit this.
   45. Cowboy Popup Posted: February 18, 2010 at 06:17 PM (#3462814)
What does this have to do with clutch ability?

#28 suggested that if players did perform better in the clutch, they must be slacking at all other times. I wanted to illustrate that players often perform better or worse over periods of time through no conscious decision of their own. Therefore, its unreasonable to suggest that if a player does perform better in the clutch, he is a slacker. The point is not meant to illustrate that clutch hitting exists, but rather to show that players perform better or worse without some intentional or conscious change in effort.

EDIT: Thanks Preserved, you nailed what I was trying to get at.
   46. Yeaarrgghhhh Posted: February 18, 2010 at 06:18 PM (#3462816)
Jim Palmer is often cited as an example of a clutch pitcher.

Overall: .230/.294/.340
Runners on base: .228/.292/.330
RISP: .213/.290/.307
Bases loaded: .196/.230/.234
   47. Sheer Tim Foli Posted: February 18, 2010 at 06:21 PM (#3462819)
How about every time a batter comes to the plate and the crowd gets on its feet to watch the next play.

Ha! Cowboy that was a dodge and you know it.

I found Tango's work on this in The Book so definitive I don't understand why it is still discussed. IIRC his conclusions were:
1) Some people have mathemtatically demonstrated improved results in certain pre-defined clutch situations. Others have declining "choke" or "unclutch" changes.
2) The increase or decrease provided in #1 is not significant. A .300 hitter who has a propensity to choke still bats better than a .279 clutch hitter in a clutch situation.

If you haven't read the chapter (or The Book) I recommend it. If you have read it and disagree with the findings please let us know.
   48. JPWF13 Posted: February 18, 2010 at 06:33 PM (#3462830)
#26, my rbi estimator came up with 100 for Lowell's 2007. Given that the standard error would be around 9 given his opportunities (299 is high but not extreme) ... well most errors are explained by baserunner distribution. I suspect he happened to have an unusual number of multiple baserunner ABS.


2007...
Here's the leaders, batters with the most runners on 3rd for them to drive in:
Justin Morneau    106
Mike Lowell    100
Matt Holliday    99
Vladimir Guerrero    99
Ryan Howard    98
Nick Markakis    94
Torii Hunter    93
Garrett Atkins    92
Todd Helton    92
Casey Blake    92 


Most runners on 2nd:
Alex Rodriguez    179
Ryan Howard    178
Magglio Ordonez    178
Andruw Jones    173
Aaron Rowand    173
Mike Lowell    172
Hideki Matsui    172
Jeff Francoeur    170
Orlando Cabrera    168
Garrett Atkins    166 


Most runners in scoring position overall:
Ryan Howard    276
Mike Lowell    272
Alex Rodriguez    267
Matt Holliday    263
Magglio Ordonez    260
Andruw Jones    260
Garrett Atkins    258
Jeff Francoeur    257
Orlando Cabrera    257
Nick Markakis    256 


Here's the top 20 with batters in scoring position, and their RBI totals:
NAME    rbi    
Ryan Howard    136    276
Mike Lowell    120    272
Alex Rodriguez    156    267
Matt Holliday    137    263
Magglio Ordonez    139    260
Andruw Jones    94    260
Garrett Atkins    111    258
Jeff Francoeur    105    257
Orlando Cabrera    86    257
Nick Markakis    112    256
Aaron Rowand    89    252
Vladimir Guerrero    125    251
Hideki Matsui    103    249
Justin Morneau    111    249
Bobby Abreu    101    248
Robinson Cano    97    247
Carlos Guillen    102    240
Brad Hawpe    116    240
Todd Helton    91    238
Travis Hafner    100    237 


If you rank the by % of base runners driven in, Lowell was 15th that year:
Magglio Ordonez    21.9
Matt Holliday    21.2
Emil Brown    21.1
Chase Utley    21
Ryan Braun    20.9
Vladimir Guerrero    20.5
Sammy Sosa    20.3
James Loney    20.2
Victor Martinez    19.8
Nick Markakis    19.6
Miguel Cabrera    19.6
Chone Figgins    19.6
Raul Ibanez    19.5
Alex Rodriguez    19.2
Mike Lowell    19 


The worst in 2007 at driving in runners?
Shannon Stewart    10.9
Ryan Garko    10.8
Alfredo Amezaga    10.7
David Eckstein    10.6
Juan Pierre    10.5
Chris Burke    10.2
Bobby Crosby    9.3
Brad Ausmus    8.7
Rickie Weeks    8
Nick Punto    7.7 
   49. PreservedFish Posted: February 18, 2010 at 06:37 PM (#3462836)
Wait, a coin flip is not random?
   50. Cowboy Popup Posted: February 18, 2010 at 06:44 PM (#3462842)
Ha! Cowboy that was a dodge and you know it.

Its not a dodge. I don't believe that you can define a clutch situation by simply looking at the men on base or the score or the inning. Is every 3-2 game of the same intensity? I certainly don't think so. Simply put, I do not think you can define clutch without judging the atmosphere and the intensity in the stadium and in responses by the players. If you want to believe that every tying runner on third with two outs after the 7th inning situation is the same, feel free to. I disagree.

I found Tango's work on this in The Book so definitive I don't understand why it is still discussed. IIRC his conclusions were:
1) Some people have mathemtatically demonstrated improved results in certain pre-defined clutch situations. Others have declining "choke" or "unclutch" changes.
2) The increase or decrease provided in #1 is not significant. A .300 hitter who has a propensity to choke still bats better than a .279 clutch hitter in a clutch situation.


I think you've already shown that it requires a different argument to convince you that clutch doesn't exist than to convince me of the same.

If you haven't read the chapter (or The Book) I recommend it. If you have read it and disagree with the findings please let us know.

Well, I'm not going to do that. Nor do I think that my opinion should be discredited simply because I haven't read Tango's book. I've read plenty of his stuff that I've disagreed with in the past and I reserve the right to have my opinions about baseball with out reading math heavy books that I would have to shell out cash for. If you want to try and persuade me right here, feel free to.

Edit: I hope that doesn't come off as defensive or nasty. Its not intended to be.
   51. OsunaSakata Posted: February 18, 2010 at 06:46 PM (#3462846)
Some acronyms that were intended to be plural have become singularized. "John McCain was a POW." "Derek Jeter had one RBI yesterday."

It might have been Bill James' managers book where he simulated lineup construction. I think he put a Mark Belanger-type cleanup behind Ricky Henderson and Tim Raines and still came with 100 RBIs. It shouldn't be too diffult too experiment with an OOTP team.
   52. Walt Davis Posted: February 18, 2010 at 06:59 PM (#3462859)
2) there’s no particular skill to driving runs in. There’s no such thing as a hitter who is significantly better in RBI opportunities.

Poorly phrased. Of course Albert Pujols is significantly better in RBI opps than Willy Taveras. No doubt Keith means that, after controlling for BA/OBP/SLG (or whatever), there is no detectable "run-producing" effect. Probably fine when talking with this crowd but not a good way to go about things on the radio (unless you just want to generate calls ... which is the purpose of talk radio.)
   53. Sheer Tim Foli Posted: February 18, 2010 at 07:02 PM (#3462862)
Its not a dodge. I don't believe that you can define a clutch situation by simply looking at the men on base or the score or the inning.


I do totally agree - but if we really want to know the truth throwing out conditional at bats altogether is, I think, overkill. Maybe all of them won't be true clutch situations but surely most of them are - at least in the games I go to having, as an example, the tying run on base in the bottom of the ninth gets they butts out of seats.

Well, I'm not going to do that. Nor do I think that my opinion should be discredited simply because I haven't read Tango's book.


I didn't meant to imply your opinion is not worth mine because I read the book. Many people I know have never read the book. Some of them good friends of mine. I only meant to suggest that for those who find the topic interesting it is the most thorough analysis I have seen.

If you want to try and persuade me right here, feel free to.


Actually if you believe "in clutch" we agree. I was in the camp that disagreed until I read The Book and I decided it was the best argument I have seen that supports it so I changed my mind. As I said previously, however, his math argues that clutch or choke numbers may not be as significant as fans think. I do not know how closely that aligns with your approach.

Also I haevn't read the book in years so I may be misrepresenting the chapter. I am sure someone smarter than me will comment.
   54. BDC Posted: February 18, 2010 at 07:02 PM (#3462863)
I do not think you can define clutch without judging the atmosphere and the intensity in the stadium and in responses by the players

This is the single greatest problem in convincing the unconvinced that "clutch" is not a predictable or repeatable ability. If you compare batting with RISP, there's batting with RISP close and late that's even clutchier. If there's batting RISP close/late, there's Sept./Oct./postseason RISP-extra-innnings batting that's exponentially clutchier. If there's that, then there's the Francisco Cabrera pennant-or-elimination situation that's a hundredfold clutchier. If there's Cabrera, then there's Matt Wieters on chemo attached to an oxygen tank batting against Curt Schilling in his bloody socks with the freedom of humanity from its new alien overlords on the line. You can always find a clutchier subset of situations, but the sample size gets smaller and smaller, and the results less and less reliable.
   55. PreservedFish Posted: February 18, 2010 at 07:05 PM (#3462868)
then there's Matt Wieters on chemo attached to an oxygen tank batting against Curt Schilling in his bloody socks with the freedom of humanity from its new alien overlords on the line.


Who's playing for the aliens?
   56. RJ in TO Posted: February 18, 2010 at 07:07 PM (#3462869)
Who's playing for the aliens?


Otis Nixon.
   57. Pasta-diving Jeter (jmac66) Posted: February 18, 2010 at 07:11 PM (#3462874)
c'mon you guys, get real--it's TRIVIAL to define a clutch situation:

a clutch situation is any situation where Derek Jeter gets a hit

QED
   58. ?Donde esta Dagoberto Campaneris? Posted: February 18, 2010 at 07:15 PM (#3462880)
2) there’s no particular skill to driving runs in. There’s no such thing as a hitter who is significantly better in RBI opportunities.

Poorly phrased. Of course Albert Pujols is significantly better in RBI opps than Willy Taveras. No doubt Keith means that, after controlling for BA/OBP/SLG (or whatever), there is no detectable "run-producing" effect


I'm not clear that your caveat even saves that statement (though I may be misreading it.) Runs score on a few type of events; hits, outs, or walks. There are a few other events, but they rarely occur. The vast majority of runs score on hits. As such, a "high hit" guy (230 hits + 20 walks/year) will typically drive in more runs than a guy who gets 180 hits + 70 walks/year. Even though their non-outs are equal, the former player is generating much more advancement and will, therefore, drive in more runs assuming equal opportunities.

Obviously, you have to factor in the number of extra base hits- but my point is that a guy who gets lots of hits will move more runners and drive in more runs than a guy who gets a good number of hits and lots of walks. Is that wrong?
   59. heyyoo Posted: February 18, 2010 at 07:16 PM (#3462882)
Jim Palmer is often cited as an example of a clutch pitcher.



Yeah, but his clutchness is matched by Doug Davis

Palmer
Overall: .230/.294/.340 .634 OPS
Runners on base: .228/.292/.330 .622 OPS (better by 12 points of OPS)
RISP: .213/.290/.307 .597 OPS (better by 38 points of OPS)
Bases loaded: .196/.230/.234 .564 OPS (Better by 70 points of OPS)

Davis
Overall: .271/.349/.418 .767 OPS
Runners on Base: .261/.339/.391 .730 OPS (Better by 30 points of OPS)
RISP: .255/.340/.380 .720 OPS (Better by 47 points of OPS)
Bases Loaded: .265/.318/.391 .709 OPS (Better by 58 points of OPS)

:-)
   60. Cowboy Popup Posted: February 18, 2010 at 07:17 PM (#3462886)
Maybe all of them won't be true clutch situations but surely most of them are - at least in the games I go to having, as an example, the tying run on base in the bottom of the ninth gets they butts out of seats.

I won't argue with you there, but i still think that the more general situational stats include too many non clutch situations and that specific ones like the one you mention end up excluding a lot of them. I don't think that there is any subset of situations that's ever going to give us a real good read on what situations are truly clutch.

I only meant to suggest that for those who find the topic interesting it is the most thorough analysis I have seen.

My bad, I obviously misinterpreted a lot of what you were saying in your post.

Actually if you believe "in clutch" we agree. I was in the camp that disagreed until I read The Book and I decided it was the best argument I have seen that supports it so I changed my mind. As I said previously, however, his math argues that clutch or choke numbers may not be as significant as fans think. I do not know how closely that aligns with your approach.

I honestly have no idea how I think clutch translates into production. I believe that some players consistently have better at bats and approaches in clutch situations, I can't really speak to it more than that. I always squirm when people throw out rate stats for this too because to me, its just as much about approach as it is about production.

It wouldn't shock me if the statistical difference in their production is limited. I believe Ruben Sierra at the end of his career was a legitimate clutch hitter. It didn't mean I'd rather see him up there instead of A-rod or Nick Johnson (I still can't believe Torre did that). He was still a terrible hitter who did not deserve a spot on the roster, he just always had a seemingly good approach in big spots and got a lot of big hits for such a crappy hitter. Torre seemingly agreed with me but I obviously feel he took it too far.

You can always find a clutchier subset of situations, but the sample size gets smaller and smaller, and the results less and less reliable.

Almost makes you think that people should stop trying to convince other people that clutch hitting isn't a repeatable talent and vice versa. I think this whole argument has to do with personal belief rather than any objective data.
   61. Pasta-diving Jeter (jmac66) Posted: February 18, 2010 at 07:22 PM (#3462891)
I think this whole argument has to do with personal belief rather than any objective data.


you thought I was kidding in #58, dinnya?
   62. JPWF13 Posted: February 18, 2010 at 07:29 PM (#3462903)
It might have been Bill James' managers book where he simulated lineup construction. I think he put a Mark Belanger-type cleanup behind Ricky Henderson and Tim Raines and still came with 100 RBIs.


The 1985 Cardinals did that for real, Tommy Herr got 100 rbis hitting .302/.379/.416 with 8 homers he had 110 ribbies (Ok Belanger never had a .416 OBP, but still..)

Bill Brubaker in 1936 batted .289 with 6 homers, slugged .384 and still had 102 ribbies

That team had 3 batters with OBPs over .400, Suhr .410, Vaughn .453 and Waner .446
I have a sneaking suspicion that at least one of them were batting in front of Brubaker
   63. Joe Bivens, Minor Genius Posted: February 18, 2010 at 07:30 PM (#3462904)
Getting back to the article and its comments, I was glad to see someone make the point that if you put Bay on KC he wouldn't have nearly as any RBI as he had w/the Sox. A no brainer.

I'm gonna check BBRef and see what guys like Teahan and Butler and Buck had for OPS w/RISP. My gut tells me their numbers were comparable to hitters' on the Red Sox or other good hitting teams.
   64. BDC Posted: February 18, 2010 at 07:33 PM (#3462909)
believe that some players consistently have better at bats and approaches in clutch situations, I can't really speak to it more than that. I always squirm when people throw out rate stats for this too because to me, its just as much about approach as it is about production

That's OK, Cowboy, but it seems to lift the discussion from facts into hunches based on impressions. If I note that Ruben Sierra in 2004 hit .264/.300/.491 in a tiny sample of "late and close PAs," marginally better than his .244/.296/.456 overall, you will squirm and we'll have to stop talking about it. If you opine that the older Ruben Sierra got, the finer a human being and the more professionally prepared for his PAs he became, though, that's cool :)
   65. Josh1 Posted: February 18, 2010 at 07:34 PM (#3462910)
I'm not sure if this is exactly what Dan meant by a coin flip not being random, but the idea seems reasonable. The outcome is determined by how hard and what angle the coin is flicked. With practice repeating the exact same motion, I'm sure someone could bias the landing, and apparently per the internet (which is never wrong) there are con artists who try to do biased-flip betting as a scam.

Cowboy, I don't want to pick a fight, but it seems you are deliberately defining your argument in a way to make it impossible to disprove. It seems you are starting with your conclusion (from faith?) and then blocking any attempts to test it. You are of course free to have any opinion you want, but the more informed and researched your opinion is, the more convincing it will be to others if your goal is to convince them to agree with you.
   66. Pasta-diving Jeter (jmac66) Posted: February 18, 2010 at 07:36 PM (#3462912)
Cowboy, I don't want to pick a fight, but it seems you are deliberately defining your argument in a way to make it impossible to disprove

of course he is--one can define "clutch" in a completely circular fashion, should one choose to, as use this as "proof" of the existence of clutchiers and chokers

(not unlike St. Anselm, when it comes down to it)
   67. bibigon Posted: February 18, 2010 at 07:37 PM (#3462915)
Obviously, you have to factor in the number of extra base hits- but my point is that a guy who gets lots of hits will move more runners and drive in more runs than a guy who gets a good number of hits and lots of walks. Is that wrong?


Probably, yes. The guy with the lower batting average gets has to get more extra base hits to make up for the lower batting average (in order for his SLG to be identical, as was the initial condition).
   68. Joe Bivens, Minor Genius Posted: February 18, 2010 at 07:43 PM (#3462921)
BBref doesn't show OPS w/RISP (that I could find), but Butler had 93 RBI on that shitful team last year, with an OPS+ of 124, so I think it's safe to say his #'s were comparable to many of the usual suspects.
   69. Cowboy Popup Posted: February 18, 2010 at 07:44 PM (#3462922)
Cowboy, I don't want to pick a fight, but it seems you are deliberately defining your argument in a way to make it impossible to disprove.

Not really, its entirely possible to go through the game tapes and figure out when the crowd stands. No one wants to do that of course, but its certainly possible.

It seems you are starting with your conclusion (from faith?) and then blocking any attempts to test it.

Not really. If you want to convince me that RISP in a certain situation is always clutch, explain to me why I should accept it. What I'm not willing to accept right now is a conclusion that clutch hitting doesn't exist based on numbers that are based on generalities (performance "close and late" or "RISP after the 7th" or something like that).

You are of course free to have any opinion you want, but the more informed and researched your opinion is, the more convincing it will be to others if your goal is to convince them to agree with you.

Well I don't think that's fair, I read nearly every clutch hitting thread on this site and I read the arguments and the math behind all of it. And none of it convinces me. There are easily identifiable holes in many of the arguments that suggest clutch hitting doesn't exist. I've mentioned some of my issues with those arguments above and no one has really addressed my points other than to ask me specifically to define what a clutch situation is. I raised issues that don't require a set definition of clutch and why clutch ABs might be different than other ABs and no one has addressed them.

What's your definition of clutch? I'd love to see what a more acceptable and less "circular" definition of clutch is.
   70. Cowboy Popup Posted: February 18, 2010 at 07:50 PM (#3462930)
That's OK, Cowboy, but it seems to lift the discussion from facts into hunches based on impressions.

I don't think it really changes anything. I used Sierra as an example. But the notion that you can't objectively show a player has a good approach in a big time situation doesn't seem right to me. Things such as swing at pitches outside the zone, or swinging at first pitch breaking balls seem like easily identifiable ways to determine who has a good approach in those situations. Maybe no one wants to do the leg work, I certainly don't, but it doesn't seem beyond the realm of possibility to study something like that.

If I note that Ruben Sierra in 2004 hit .264/.300/.491 in a tiny sample of "late and close PAs," marginally better than his .244/.296/.456 overall, you will squirm and we'll have to stop talking about it.

We don't have to stop talking about it. But I don't think rate stats are the end all. Frankly, they aren't the end all for any discussion of true offensive talent. We have things like platoon splits and LD% and performance on particular pitches to better evaluate hitters and I don't see why those stats shouldn't factor into evaluating a hitters clutch performance as well.

EDIT: I should add, why are rate stats seemingly acceptable in samples of 30-50 PA spread throughout a season in clutch situations while the first 30-50 PAs of a season, or any 30-50 PAs of the season are understood to be subject to a substantial amount of luck?
   71. BDC Posted: February 18, 2010 at 08:59 PM (#3463005)
why are rate stats seemingly acceptable in samples of 30-50 PA spread throughout a season in clutch situations

I think my point (and many other people's point) was that they're not. One less hit "close and late" in '04 brings Sierra's BA in those situations down to his season average; two less and he's well below it.

The problem with these small sample sizes is that the research (by Tango and others) doesn't pick up any significant effect. If you had a .260 hitter like the elderly Sierra and he went out every year and consistently hit .633 in certain defined clutch situations, then even at 30-50 PAs a year that would become significant – at least in a seat-of-the-pants sense, something a manager should sit up and notice. But when a .260 hitter bats .300 in 60 "clutch" PAs one year and .220 the next year, that really doesn't mean anything.

However, if it's not about results, then your observations are as valid as anyone's :)
   72. 'zop sympathizes with the wrong ####### people Posted: February 18, 2010 at 09:13 PM (#3463022)
I don't think it really changes anything. I used Sierra as an example. But the notion that you can't objectively show a player has a good approach in a big time situation doesn't seem right to me. Things such as swing at pitches outside the zone, or swinging at first pitch breaking balls seem like easily identifiable ways to determine who has a good approach in those situations. Maybe no one wants to do the leg work, I certainly don't, but it doesn't seem beyond the realm of possibility to study something like that.

But this is just so stupid. A "good approach" is only defined as such because it leads to good results. Good results are measured in statistics. If a hitter is using an approach in a clutch situation that makes little Cowboy Popup pop up, that's fine and dandy, but if it has no impact upon tangible results, who cares?
   73. Cowboy Popup Posted: February 18, 2010 at 09:27 PM (#3463040)
I think my point (and many other people's point) was that they're not.

Well that's also my point. That you simply can't use these results to show one thing or another because the sample isn't big enough. And that's why one way or another, this sort of thing can't be proved. All of the data is based on extremely small samples. And yet when I say it, I'm given a tremendous amount of crap for it and when the "clutch doesn't exist" crowd says it, it somehow validates their point.

If you had a .260 hitter like the elderly Sierra and he went out every year and consistently hit .633 in certain defined clutch situations, then even at 30-50 PAs a year that would become significant – at least in a seat-of-the-pants sense, something a manager should sit up and notice.

Well, why does Sierra have to hit .633 in certain defined clutch situations to be clutch? Even if he were a true talent .633 clutch hitter, you wouldn't expect him to perform like that in such small samples. Because the numbers are so completely unreliable, I tried to move the discussion away from that into reasoning about changes in pitchers performance and stadium atmosphere and things of that nature, but since that was dismissed out of hand because there are no numbers, I don't know how else you all expect to prove anything. The proof, one way or another, about clutch hitting is not going to come from numbers based on samples of 30-50 PAs.

If a hitter is using an approach in a clutch situation that makes little Cowboy Popup pop up, that's fine and dandy, but if it has no impact upon tangible results, who cares?

Well, if you're trying to isolate clutch from the luck inherent in the small samples you might try something like. Until you find a way to show that these clutch samples are somehow meaningful, the numbers and the studies based on them are equally meaningless. Because the numbers are essentially meaningless, I fail to see how everyone can be so certain one way or another, relying just on the numbers. I mean, I know numbers are dogma among the intellectual elite around here, but at least you normally have some sort of reason for believing that. All I see here is counter-culture.
   74. Dylan B Posted: February 18, 2010 at 09:50 PM (#3463066)
I shall not today attempt further to define the kinds of material I understand to be embraced within that shorthand description; and perhaps I could never succeed in intelligibly doing so. But I know it when I see it, and the motion picture involved in this case is not that.- Justice Potter Stewart

Seems to fit perfectly with the talk regarding clutchiness.
   75. Josh1 Posted: February 18, 2010 at 09:58 PM (#3463078)
Well I don't think that's fair, I read nearly every clutch hitting thread on this site


Fair enough, there was no intent to offend: the way you phrased your previous post did not make this clear.

Not really. If you want to convince me that RISP in a certain situation is always clutch, explain to me why I should accept it... What's your definition of clutch?


(1) People have investigated many definitions of clutch, and as far as I know, the results have been similar. Even if close and late isn't clutch 100% of the time, it is most of the time. If 75% of close and late PA are clutch and 25% are regular pressure, if there really is a reasonable-sized "clutch" effect, it would still show up in the summation (sum of two random variables plus a trend still contains the trend). To answer the question of my definition of clutch, I would think using high LI PA would be pretty reliable; I don't see why standing fans would be better.
(2) I think you are backwards on the burden of proof. The null hypothesis needs to be "there is no clutch hitting." This statement can be rejected with a counterexample. The null "there is clutch hitting" cannot be tested completely. There have been plenty of studies trying to reject the proper null hypothesis, and most have not been able to do so. The Book IIRC (have not read in a while, so feel free to correct me) showed some small evidence of clutch hitting existing, but the effects were small enough to ignore for practical purposes, and it is almost impossible to know in advance who the real clutch guys are due to sample size. Maybe some day someone will find a study that shows more significant clutch effects, and I'll change my mind. Until then, I'm using the best evidence we have to form my opinion, and it seems you have a different opinion based on something this is totally unclear to me.
   76. Ron Johnson Posted: February 18, 2010 at 10:07 PM (#3463095)
#75 I agree. Here's my problem. People who talk about clutch tend to speak as they they think it's important. And yet they can't come up with anything approaching a working definition.

As I've noted above I'm not out to prove clutch doesn't exist. I honestly don't care. Paul Molitor and Tony Fernandez meet a definition of clutch (using above 95% confidence as the test. While that's not precisely statistical significance it'll have to do)

I will note that the predicted number of players to meet this standard (given the number of players in the study) is right about 2.
   77. Cowboy Popup Posted: February 18, 2010 at 10:22 PM (#3463116)
People have investigated many definitions of clutch, and as far as I know, the results have been similar.

Based on the same numbers, I'm not surprised that they had similar results.

Even if close and late isn't clutch 100% of the time, it is most of the time.

Based on what? I don't agree with the assumption.

If 75% of close and late PA are clutch and 25% are regular pressure, if there really is a reasonable-sized "clutch" effect, it would still show up in the summation (sum of two random variables plus a trend still contains the trend).

Again, these are just random numbers and I have no reason to accept them as truth. What if 25% of them are clutch and 75% of them are of regular pressure? Would the trend still hold?

To answer the question of my definition of clutch, I would think using high LI PA would be pretty reliable; I don't see why standing fans would be better.

Since clutch is something of a moving definition, standing fans would indicate something of an agreement among the people watching the game that the moment was fairly important. Using high LI PAs seems to just be more of the same to me, picking situations where the game is close and theoretically in balance. I don't think this automatically or even likely makes a situation clutch. And even if it did, the number of high LI PAs would consistently be small enough to be subject to wild variations that don't make the numbers they generate meaningful for any sort of conclusive analysis.

Until then, I'm using the best evidence we have to form my opinion, and it seems you have a different opinion based on something this is totally unclear to me.

My point is that the best evidence is almost completely worthless, its based off of small samples and general situations that don't precisely encompass the situations we are attempting to evaluate. Because I believe the numbers are useless, I'm throwing them out the window. There are plenty of other things that can allow you to deduce when clutch situations do exist (fan and player reactions), that certain batters would likely be better at it (based on the change in pitching strategy) and that there is reason to believe there are players who are better at it than others (consistent player and manager testimonials).

I don't expect people here to buy into those things, but I'm also surprised at how much faith is being put into some very questionable numbers over the common wisdom that's existed for more than a century. I think its likely that primates, when they find common wisdom that can't be proved, automatically assume it is wrong. I'm perfectly willing to disregard the common wisdom when it is demonstrable that it is wrong, but that is not the case here. The notion that clutch hitting has to be proved is a fallacious one in my opinion. I don't have to prove anything and neither do you, we've based our opinion on different things, neither of our opinions can be proven by some infallible objective data. Whether anyone wants to accept the foundation of my belief as valid or logical is somewhat inconsequential to me, I've read what I've written over a couple of times and it makes sense unless you are unwilling to accept that there must be numbers to support any baseball related conclusion.

I shared what I thought because I've seen some uninspiring analysis in regards to clutch hitting floating around these parts and I wanted to discuss why I didn't think it held up. I didn't expect to convince anyone, I only hoped to make the argument that these things aren't as clear or certain as they were being presented in post #28 and in what Law said in his interview.
   78. Tracy Posted: February 18, 2010 at 10:23 PM (#3463120)

The 1985 Cardinals did that for real, Tommy Herr got 100 rbis hitting .302/.379/.416 with 8 homers he had 110 ribbies (Ok Belanger never had a .416 OBP, but still..)

Bill Brubaker in 1936 batted .289 with 6 homers, slugged .384 and still had 102 ribbies

That team had 3 batters with OBPs over .400, Suhr .410, Vaughn .453 and Waner .446
I have a sneaking suspicion that at least one of them were batting in front of Brubaker


All three of them did, according to Retrosheet. Brubaker hit mostly 6th, while Waner, Vaughan, and Suhr hit 3-4-5.

Retrosheet is my Jesus, btw.
   79. Dylan B Posted: February 18, 2010 at 10:32 PM (#3463136)
Since clutch is something of a moving definition, standing fans would indicate something of an agreement among the people watching the game that the moment was fairly important.


Not to pile on(I actually do think there is such a thing a clutch players), but wouldn't fans standing have a bit of preformance bias? Are fans more likely to stand when Pujols is up at bat vs Aaron Miles?

Like I said I do think that it exists, that is it important and that it is a repeatable skill, just that I have no clue how to actually measure it on a consistent basis outside of observing it in action, and understand that isn't really useful in projecting across a large pool of individuals.
   80. Cowboy Popup Posted: February 18, 2010 at 10:41 PM (#3463152)
Not to pile on(I actually do think there is such a thing a clutch players), but wouldn't fans standing have a bit of preformance bias? Are fans more likely to stand when Pujols is up at bat vs Aaron Miles?

Sure, but is it any less reliable than any of the splits that have been suggested? Given that no one is suggesting that any of these splits will perfectly define clutch, I'm not sure why its an issue to suggest something that might also miss some clutch situations or include some that aren't. My only point in making suggesting this method was to add some kind of subjective quality to the evaluation since what we are measuring is subjective.

I believe you can define how clutch is however you want to, the sample of PAs will still be useless.

Like I said I do think that it exists, that is it important and that it is a repeatable skill, just that I have no clue how to actually measure it on a consistent basis outside of observing it in action, and understand that isn't really useful in projecting across a large pool of individuals.

I have the same feeling about it, I didn't mean to suggest anything else. That's why I made my faith statement earlier, I don't think there is any way to prove or disprove it and you simply have to accept what you see in the ballgames.
   81. CrosbyBird Posted: February 18, 2010 at 10:57 PM (#3463162)
Well that's also my point. That you simply can't use these results to show one thing or another because the sample isn't big enough. And that's why one way or another, this sort of thing can't be proved. All of the data is based on extremely small samples. And yet when I say it, I'm given a tremendous amount of crap for it and when the "clutch doesn't exist" crowd says it, it somehow validates their point.

There's no particular evidence that certain pitchers are "wind-lovers," who perform better than expected given the particular wind conditions of the day. A pitcher might allow relatively fewer extra HR when the wind is blowing out or in, as compared to his peers. I can certainly rationalize why a given pitcher's style might be impacted more or less by wind conditions.

Should I behave as if this phenomenon exists? Should I choose between Pitcher A and Pitcher B based on their propensity to allow more/fewer HR with the wind blowing out?

Something that doesn't show up significantly in the data isn't proof that it doesn't exist at all, but it's generally good policy not to make decisions based on effects that don't show up in the noise, particularly when we have many other reasons to make a decision one way or the other, all supported by more accurate data points.

Because the numbers are essentially meaningless, I fail to see how everyone can be so certain one way or another, relying just on the numbers.

There is not "certainly not." There is only "certainly not enough evidence to reach a positive conclusion."

I am not certain that clutch ability does not exist at the ML level. I am fairly certain that there's isn't reasonable evidence to conclude that clutch ability is useful. It can't be predicted and acted upon in a significant way (toward furthering the goal of winning games).

When people suggest that you'd rather have David Ortiz than Alex Rodriguez in a clutch situation, they are ignoring significant data that definitively shows A-Rod is the better hitter in favor of data which is not so supportable. That's a bad gamble.
   82. Josh1 Posted: February 19, 2010 at 12:08 AM (#3463175)
Again, these are just random numbers and I have no reason to accept them as truth. What if 25% of them are clutch and 75% of them are of regular pressure? Would the trend still hold?

Yes, it would hold but would appear smaller and be more noisy. I didn't mean to put out any exact numbers by the way. It seems to me that if some player could "rise to the occasion" for some number of at bats per year, you'd want him to do it in late and close situations or even better when LI is high. If a player is "stepping it up" at other times, I don't see the point in being clutch, as it doesn't then have a big impact in winning games. If clutch is performance in vaguely defined and does not necessarily mean critical times for winning the game, we're just devolving into having a silly rhetorical discussion.

It would be good if sample sizes were larger, but the samples are large enough to rule out big effects at least. We might still be missing small but real effects. The evidence isn't perfect, but decent evidence is a lot better than no evidence. My default position with respect to any issue isn't to believe the conventional wisdom: conventional wisdom is often very wrong and sometimes ridiculously stupid. I try to look at whatever evidence exists and make up my own mind. Conventional wisdom isn't even evidence.

Digressing a bit, I just watched "Sugar" the other day (which it seems Dominican players say is a very realistic movie). The pressure even in Rookie League/A- every day is immense. A few bad games in a foreign environment can end what a man has worked on his entire life, and he has nothing to fall back on. It is pretty hard to imagine many chokers make it to the majors.
   83. Forsch 10 From Navarone (Dayn) Posted: February 19, 2010 at 12:11 AM (#3463176)
Digressing a bit, I just watched "Sugar" the other day (which it seems Dominican players say is a very realistic movie).

I think most baseball movies (non-docs, that is) are lousy, but "Sugar" was a most pleasing exception.
   84. Don Malcolm Posted: February 19, 2010 at 01:24 AM (#3463204)
Player X: don't forget to check out Maurice Van Robays...

I think that this thread is either an example of questionable selectivity by Repoz, or else an illuminating experiment in determining exactly how few of the people who post at BTF actually RTFA. The discussion of RBIs and "clutch" is a saw so old that its teeth are no longer even visible on the blade, and while some of you are trying valiantly, there is precious little here that is far removed from the odor of dead fish and bullet holes in an empty barrel.

The portion of Law's interview that is actually worth comment is found lower down, when he is asked to explain/defend his actions with respect to the NL Cy Young voting. There is much more of interest to contemporary issues and viewpoints to be found in what Law says there:

I haven’t criticized anybody for putting Carpenter on their ballot over Vazquez. I thought those four guys were all fairly close. But Carpenter did get a little bit of a boost from, whether you call it luck, or from the defense, probably a little bit of both. Vazquez, on the other hand, was probably hurt a little bit by some of those same factors.

Not a lot of precision in "advanced metrics" to be found here. Law sounds more like a "trad" writer doing a little mixology than a guy wedded to WAR or VORP or FIP or QMAX or WHATNOT. Maybe that's good, but maybe it would be nice if Law weren't so slippery.

There was one other factor in Law's comp of Carpenter and Vazquez that's worthy of note:

And I gave Vazquez a boost also for pitching in a tougher division. I go out and see all these clubs every year, at least the contending clubs. The NL East is a much better division offensively than the N.L. Central was, and I think that skews the stats. What Vazquez did, coming out roughly equal or slightly ahead of Carpenter in a tougher division, was worth putting him higher on the ballot.

Now that is something that you guys might want to discuss, rather than flog the glue-factory nag of RBI/clutch. Does the unbalanced schedule call for yet another type of adjustment, or is it counterbalanced by other contextual stats--for instance, the stat that shows that within the NLC last year, where the Cards played to a draw with the Cubs and Reds, Chris Carpenter was 11-0 with a 1.67 ERA? He and Wainwright dominated the NLC, going a combined 21-2, which means that Cards were just 25-32 in the NLC games those two didn't pitch.

Does winning a division/pennant matter in terms of award voting? If it's a mixology type of thing, when does it matter and when not? Vazquez pitched well for an also-ran team, but he didn't dominate his own division: he merely held his own (4-5, 4.14). He was 11-4 against the East and West, had only 40% of starts against NLE opponents.

With all this adjustment to factor in, Law should be able to do better than simply say that the "stats are skewed." Let's face it--FIP isn't the stat addressing this issue. None of the new "advanced metrics" are doing that, either. Yet this seems to be his ace in the hole for his Vazquez vote, but I think the actual value of the card is below face-card level, say around a five or six in the deck.

Here's the data for the Top 5 NL Cy pitchers broken out by division. ERA in the first three, pct of starts vs. division in the next three (sorry, still can't get it together to display tables). How to adjust for "the skew" is something I will leave to others. These five guys really dominated in the NLC, with a combined W-L record of 37-7, and were solid against the NLW (20-10). They were just .500 against the East (14-14), however. Carpenter's 6.16 ERA there stands out to the naked eye, but it was compiled in only three starts. He was the #1 pitcher vs. NLC and NLW opponents.

Pitcher, E, C, W, E, C, W
Lincecum, 1.67, 3.06, 2.65, .22, .25, .44
Carpenter, 6.16, 1.67, 2.12, .10, .52, .28
Wainwright, 4.26, 2.56, 2.67, .18, .50, .24
Vazquez, 4.14, 2.4, 2.54, .41, .34, .22
Haren, 1.67, 2.4, 3.07, .18, .24, .48
AVG TOP 5, 4.23, 2.34, 2.69, .22, .37, .33
LG AVG, 4.69, 4.35, 4.49, .28, .34, .28
   85. PreservedFish Posted: February 19, 2010 at 01:29 AM (#3463205)
The reason that Sugar is not lousy is that it's not really a sports movie. Baseball is just the setting, the detail. The narrative does not hinge on in-game events.

I liked it too.
   86. Josh1 Posted: February 19, 2010 at 02:55 AM (#3463246)
Don,

I can see thinking about performance against competitive division rivals also competing for the title (not the dogs of the division) similarly to how people think about clutch performance after the fact (or WPA). There is definite value in hindsight, but who knows if it is predictive or if it was just luck. I could see some people totally ignoring your data and others finding it very important. It certainly does help the Cardinals marginally more toward playoff probability to put a great start against the Cubs and a bad one against the Padres rather than vice versa.

A number of people have argued strength of opposition should be accounted for when evaluating Halladay's stats the past few seasons against his competitors for awards.
   87. Something Other Posted: February 19, 2010 at 03:05 AM (#3463251)
Once they hit the show, there are new pressures, and sure, 50,000 screaming fans in a playoff scenario probably gets to some guys. But for the vast majority, they already have the ability to perform under pressure, or they would not have gotten there in the first place.

Like Tim Kurjian said one night, "These guys are not like you and me".
They do indeed have more money.

You're neglecting the possibility that within the ability to reach the majors and not become paralyzed playing in front of 50,000 screaming fans there are real gradations in players' ability to perform under pressure.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Backlasher
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogOT: The Soccer Thread March, 2014
(942 - 11:55am, Apr 20)
Last: I am going to be Frank

NewsblogTim Federowicz has rough night behind the plate - True Blue LA
(6 - 11:52am, Apr 20)
Last: Tom Nawrocki

NewsblogJohn Torres: Baseball must bag sickening farewell tours
(13 - 11:51am, Apr 20)
Last: Tom Nawrocki

NewsblogOMNICHATTER for April 20, 2014
(1 - 11:48am, Apr 20)
Last: Rickey! In a van on 95 south...

NewsblogRB: Carlos Beltran: more of a center fielder than Mickey Mantle, Ty Cobb or Duke Snider. So what?
(43 - 11:46am, Apr 20)
Last: BDC

NewsblogA’s Jed Lowrie “flabbergasted” by Astros’ response to bunt
(18 - 11:43am, Apr 20)
Last: What did Billy Ripken have against ElRoy Face?

NewsblogOT: The NHL is finally back thread, part 2
(155 - 11:31am, Apr 20)
Last: zack

NewsblogPirates Acquire Ike Davis From Mets
(42 - 11:19am, Apr 20)
Last: formerly dp

NewsblogRaissman: After cheating scandal, former Mets GM Steve Phillips back calling the shots
(2 - 11:11am, Apr 20)
Last: spike

NewsblogDesign Room: Top 10 Logos in MLB History.
(79 - 11:05am, Apr 20)
Last: Nats-Homer-in-DC

NewsblogDoug Glanville: I Was Racially Profiled in My Own Driveway
(369 - 10:54am, Apr 20)
Last: CrosbyBird

NewsblogChase Utley is the hottest hitter in baseball and has a shot at .400
(64 - 10:51am, Apr 20)
Last: GregD

NewsblogOTP April 2014: BurstNET Sued for Not Making Equipment Lease Payments
(1743 - 10:49am, Apr 20)
Last: Morty Causa

NewsblogDaniel Bryan's 'YES!' chant has spread to the Pirates' dugout
(101 - 10:35am, Apr 20)
Last: Infinite Joost (Voxter)

NewsblogBryce Harper benched for 'lack of hustle' despite quad injury
(66 - 9:45am, Apr 20)
Last: The Id of SugarBear Blanks

Demarini, Easton and TPX Baseball Bats

 

 

 

 

Page rendered in 1.0043 seconds
52 querie(s) executed