Jump to content


Yo! You're not logged in. Why am I seeing this ad?

Photo

WPA Since the All-Star Break


  • This topic is locked This topic is locked
20 replies to this topic

#1 Eric Van


  • Kid-tested, mother-approved


  • 10,900 posts

Posted 17 August 2009 - 05:44 AM

Before you read the chart, try to guess who's at the bottom. Hint: I had a very sarcastic "Time to Bench X?" thread title in mind but thought better of it.

WPA Since the Break
Name WPA
Beckett 1.35
Papelbon 0.88
Martinez 0.71
Lowell 0.70
Youkilis 0.60
Ramirez 0.55
Okajima 0.53
Lester 0.42
Kotchman 0.18
Ortiz 0.15
Kottaras 0.15
Delcarmen 0.11
Cabrera 0.05
Buchholz 0.04
Masterson 0.00
Anderson -0.02
Gonzalez A. -0.04
Gonzalez E. -0.05
Baldelli -0.06
Saito -0.07
Woodward -0.08
LaRoche -0.10
Traber -0.14
Kotsay -0.16
Reddick -0.17
Bay -0.19
Tazawa -0.32
Ellsbury -0.38
Drew -0.53
Bard -0.54
Green -0.55
Lowrie -0.59
Penny -0.69
Varitek -0.73
Smoltz -1.24
Pedroia -1.29



#2 BosRedSox5


  • doesn't use Google


  • 745 posts

Posted 17 August 2009 - 06:59 AM

That's pretty surprising indeed. Even if you go by the whole year instead of just since the ASB Pedroia is in the negatives (-0.27). Judging from his more conventional numbers after the break he's hitting .272/.349/.465/.814 which seem to be pretty good. It just seems confusing that he's subtracting so much from win probability. Is he hitting into more double plays? Striking out in big spots?

Whatever it is, Petey is only 26, he's hit more homers since the break than he did before and over the rest of the season seems like a good candidate to straighten those numbers out.

Edited by BosRedSox5, 17 August 2009 - 07:02 AM.


#3 Pumpsie


  • The Kilimanjaro of bullshit


  • 10,446 posts

Posted 17 August 2009 - 07:05 AM

QUOTE (BosRedSox5 @ Aug 17 2009, 07:59 AM) <{POST_SNAPBACK}>
That's pretty surprising indeed. Even if you go by the whole year instead of just since the ASB Pedroia is in the negatives (-0.27). Judging from his more conventional numbers after the break he's hitting .272/.349/.465/.814 which seem to be pretty good. It just seems confusing that he's subtracting so much from win probability. Is he hitting into more double plays? Striking out in big spots?

Whatever it is, Petey is only 26, he's hit more homers since the break than he did before and over the rest of the season seems like a good candidate to straighten those numbers out.


I think part of his problem is that he's trying to hit more homers. They're going up but everything else is going down. It's a bad trade off. Someone should clue him in that he's not Mighty Mouse.

#4 BoSoxFink


  • Stripes


  • 3,530 posts

Posted 17 August 2009 - 07:23 AM

I don't have the full stats in front of me, but I know since the break I don't believe he has gotten a hit with men on base besides the big hit the other night in Texas. He has been leaving a ton of guys on and not only that but hitting into a ton of double plays as well. This has to be what is leading him to this effect. I also think his average has dropped from about .306 or so at the break to about .293 right now, in a rush so haven't checked those numbers but I will come back to them later when I have more time.

#5 NDame616


  • will bailey


  • 851 posts

Posted 17 August 2009 - 08:16 AM

QUOTE (BoSoxFink @ Aug 17 2009, 08:23 AM) <{POST_SNAPBACK}>
I don't have the full stats in front of me, but I know since the break I don't believe he has gotten a hit with men on base besides the big hit the other night in Texas. He has been leaving a ton of guys on and not only that but hitting into a ton of double plays as well. This has to be what is leading him to this effect. I also think his average has dropped from about .306 or so at the break to about .293 right now, in a rush so haven't checked those numbers but I will come back to them later when I have more time.


BA/OPB/SLG/OPS for Petey on July 12th and today:

July 12: 303/ .378/ .429/ .807
Today: .295/ .370/ .438/ .809


#6 Tudor Fever

  • 3,371 posts

Posted 17 August 2009 - 08:31 AM

QUOTE (BoSoxFink @ Aug 17 2009, 08:23 AM) <{POST_SNAPBACK}>
I don't have the full stats in front of me, but I know since the break I don't believe he has gotten a hit with men on base besides the big hit the other night in Texas. He has been leaving a ton of guys on and not only that but hitting into a ton of double plays as well. This has to be what is leading him to this effect. I also think his average has dropped from about .306 or so at the break to about .293 right now, in a rush so haven't checked those numbers but I will come back to them later when I have more time.

Why post just to announce that you're in such a rush that you don't have any useful information?

BB-ref has some data on this, although unfortunately, these categories aren't broken down between pre and post ASB. For all of 2009 here is his performance in the following situations:

- "Late and close" (7th inning or later, tied, ahead by 1, of with the tying run on deck): 203/257/313.
- "High leverage" (this category is obviously somewhat redundant with "late and close"): 213/329/328

Given that the BABIPs in the above situations are only .213 and .214, bad luck seems to be at least part of this. And the Buchholz play actually helped his BABIP.

He hit 13 GIDP pre-ASB, 4 thereafter, so his pace in this regard has actually slowed down a bit. (He had 393 PA pre-ASB, 130 thereafter.)

Edited by Tudor Fever, 17 August 2009 - 08:51 AM.


#7 Eric Van


  • Kid-tested, mother-approved


  • 10,900 posts

Posted 17 August 2009 - 08:37 AM

As I look at his Play Log, I see that the Buchholz play essentially broke FanGraph's allocation system, and that's part of the big negative number. Ordinarily, when a guy gets a hit and a runner is thrown out advancing, they break the play down and calculate the WP before the runner tries for the extra base, and credit the hitter with the base he got to and charge the runner with the out. Can't do that in this case. If Buchholz scores on that play, Pedroia's post-break WPA is "just" -0.78, still the worst of the offense but not hugely so (and not as bad as 'Tek per PA).

Four of Petey's eight most damaging PA have come since the break:

July 28 vs. Oakland, bottom of the 11th, tying run on 3B, 2 outs, flies out to LF to end the game (-.173).

July 30 vs. Oak, bottom of 7th, down 5-3, Ellsbury draws 4-pitch leadoff BB but Petey hits into 463 DP (-.138 and rendered moot three batters later by Papi's bomb).

Aug. 4 in TB, Sox load the bases with 1 out in the 10th and he hits into a 5-3 DP (a brutal -.353).

Aug. 7 in NY, Sox have 1st and 2nd and 2 outs in the top of the 10th, flies out to LF (-.116).

Except for the Buchholz play, his comparable positive plays since the break are all HRs:

8/4, top of the 6th off Garza ups Sox lead to 2-0 (.121).

8/6, top of 3rd off Joba gives Sox 1-0 lead (.111).

8/10, 2 run bomb off Jackson in bottom of 1st (.119).

8/16, HR off Mathis pulls Sox to within 1 (.110).

After correcting for the Buchholz play his WPA since 8/10 is +.284.







#8 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 03:22 PM

Well ... I'm honestly never been sure what WPA is supposed to be telling us, but it is interesting to see Pedroia so low. I think part of it is exactly what Pumpsie said about too many HR's. Lately I think he has been suffering from a severe case of "Monsterititis" which is leading to lots of fly balls and warning track power on the road. He's kind of morphing into Kevin Millar with a glove.

I think a lot of it may also reflect the absolute suck that has been the bottom of the lineup since the ASB not getting runners into scoring position in any meaninful way. Run production is very reliant on the three preceding slots in the batting order. Having added Anderson and SeaBass to the bottom of the lineup will surely not improve that.

#9 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 03:26 PM

QUOTE (Tudor Fever @ Aug 17 2009, 09:31 AM) <{POST_SNAPBACK}>
Given that the BABIPs in the above situations are only .213 and .214, bad luck seems to be at least part of this. And the Buchholz play actually helped his BABIP.


Sigh ... his tOPS+ of 41 and 64 for those situations say otherwise.

#10 Tudor Fever

  • 3,371 posts

Posted 17 August 2009 - 03:30 PM

QUOTE (Vermonter At Large @ Aug 17 2009, 04:26 PM) <{POST_SNAPBACK}>
Sigh ...
What is that supposed to mean?


#11 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 03:36 PM

QUOTE (Tudor Fever @ Aug 17 2009, 04:30 PM) <{POST_SNAPBACK}>
What is that supposed to mean?

Sorry, it got sent before I finished it (since edited). The sigh, however, was for BABIP=bad luck, though smile.gif I believe it goes that variations in pitcher BABIP are related to luck, but hitter BABIP is pretty much what it is. I personally think the luck part is crap altogether - BABIP just doesn't measure what we're trying to measure accurately enough to be able to see what we're supposed to see so since we can't see anything in the muck we label it as "bad luck". Or something like that.

The sample size in those BBR numbers is very small though (70-something ABs each - probably measuring the same set of ABs, more or less anyway), but for whatever reason, he hasn't been very clutchy this season. Honestly, with the suck at the bottom of the lineup and Ellsbury getting to second base at a very good rate, what Petey needs to be doing is hitting more singles when singles are called for rather than doing that warning track hero thing.

#12 Tudor Fever

  • 3,371 posts

Posted 17 August 2009 - 03:40 PM

QUOTE (Vermonter At Large @ Aug 17 2009, 04:36 PM) <{POST_SNAPBACK}>
Sorry, it got sent before I finished it (since edited). The sigh, however, was for BABIP=bad luck, though smile.gif
You don't think that BABIP depends a lot on luck? It seems fairly self-evident.

Edit: bad luck on balls in play will bring down his OPS. I think you may be confusing cause and effect.

Edited by Tudor Fever, 17 August 2009 - 03:42 PM.


#13 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 03:53 PM

QUOTE (Tudor Fever @ Aug 17 2009, 04:40 PM) <{POST_SNAPBACK}>
You don't think that BABIP depends a lot on luck? It seems fairly self-evident.

Edit: bad luck on balls in play will bring down his OPS. I think you may be confusing cause and effect.

Well ... I think that if you hit the ball hard a lot, both your BABIP and your OPS will look better. BABIP for hitters is a crude measurement of how hard a guy hits the ball. Luck certainly plays into that equation, since some hard-hit balls get caught. If he hits line drives or hard grounders a lot, his BABIP will be good. I just think a lot of his power lately has been elevated (have pitchers been pitching him down more?) which gets him doubles and HRs in Fenway, but not much anywhere else. It's been a while since we had righty hitters with Fenway power (Millar, Bellhorn, etc). Lowell is an interesting case, because he seems to have a Fenway swing and a road swing and (anecdotally at least) seems to hit more LDs on the road. Guys with true power (Manny, Bay, Rice, etc) don't really fall into this since they can clear the left field wall in most parks fairly easily.

I don't think this is a trend with Pedroia, so much as a rut, though.

#14 behindthepen


  • SoSH Member


  • 5,800 posts

Posted 17 August 2009 - 04:05 PM

For me, the stats EV posted just help to reinforce my opinion about WPA. I think it's pretty helpful in evaluating pitchers, although it pretty well tells us what we already know there ... Beckett has been great and Penny below average. But for hitters, while it gives us some color about the "clutchiness" of their historical performance, it has no predictability. In fact I think it obscures things for hitters, because it so overvalues late-inning contributions. Just because a guy missed an opportunity to win a game doesn't mean he lost it.

#15 Shelterdog


  • SoSH Member


  • 7,771 posts

Posted 17 August 2009 - 04:10 PM

QUOTE (Vermonter At Large @ Aug 17 2009, 04:53 PM) <{POST_SNAPBACK}>
Well ... I think that if you hit the ball hard a lot, both your BABIP and your OPS will look better. BABIP for hitters is a crude measurement of how hard a guy hits the ball. Luck certainly plays into that equation, since some hard-hit balls get caught. If he hits line drives or hard grounders a lot, his BABIP will be good. I just think a lot of his power lately has been elevated (have pitchers been pitching him down more?) which gets him doubles and HRs in Fenway, but not much anywhere else. It's been a while since we had righty hitters with Fenway power (Millar, Bellhorn, etc). Lowell is an interesting case, because he seems to have a Fenway swing and a road swing and (anecdotally at least) seems to hit more LDs on the road. Guys with true power (Manny, Bay, Rice, etc) don't really fall into this since they can clear the left field wall in most parks fairly easily.

I don't think this is a trend with Pedroia, so much as a rut, though.


Not over 70 at bats it won't be.

His BABip is close to 300 under all sorts of other situations this year- 2 out, RISP (364), Tie Games (287), 1 run games (312), men on base (324). The career numbers are similar as well - 300 or so BABip in situation after situation. It's incredible to me that he'd be a good hitter in every other pressure situation but fall to crap in "clutch situations."

And if he had hit in clutch situations at his career average, he gets about 5 more single in the 73 high leverage plate appearances all of the sudden he's a 300/400/490 guy in the leveraged situations and we think he's a stud.





#16 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 04:36 PM

QUOTE (Shelterdog @ Aug 17 2009, 05:10 PM) <{POST_SNAPBACK}>
Not over 70 at bats it won't be.

His BABip is close to 300 under all sorts of other situations this year- 2 out, RISP (364), Tie Games (287), 1 run games (312), men on base (324). The career numbers are similar as well - 300 or so BABip in situation after situation. It's incredible to me that he'd be a good hitter in every other pressure situation but fall to crap in "clutch situations."

And if he had hit in clutch situations at his career average, he gets about 5 more single in the 73 high leverage plate appearances all of the sudden he's a 300/400/490 guy in the leveraged situations and we think he's a stud.

Well ... you're correct in saying that the sample size is too small to assign any significance to it as a career trend or anything, but I think we can look at a 70-AB sample in this case to make the statement, "He hasn't hit well in clutchy situations this year." From that observation, you can look at other things to see if you can isolate reasons for it, it may be luck, the data may be spurious, there may be other information available, for instance an observation that he appears to be flying out more, or his Home-Road splits show something, etc.

To elaborate a little bit on BHTP's thoughts on WPA, I would say that it may have value as an indicator stat. You pull the data together and you get an unexpected result, in this case Pedroia being at the bottom of the list. You look at further data. Eric, for instance, looked at the game logs and added some qualitative data. Tudor went to look at the BBR splits to see if their numbers agreed. Et cetera ...

There are no perfect stats, no single number that tells you everything you need to know (well, there are things like VORP, WS, etc, but ...), so the best use of statistics is like the examples here, as indicators, as confirmation, or to raise an observation that we might have missed elsewhere (Would anybody, as EV asked us to try to do, have guessed that Pedroia was at the bottom of the list). The data speaks in riddles and metaphors, and it takes a lot to pull nuggets of "truth" of of them.

#17 Eric Van


  • Kid-tested, mother-approved


  • 10,900 posts

Posted 17 August 2009 - 05:34 PM

QUOTE (behindthepen @ Aug 17 2009, 03:05 PM) <{POST_SNAPBACK}>
For me, the stats EV posted just help to reinforce my opinion about WPA. I think it's pretty helpful in evaluating pitchers, although it pretty well tells us what we already know there ... Beckett has been great and Penny below average. But for hitters, while it gives us some color about the "clutchiness" of their historical performance, it has no predictability. In fact I think it obscures things for hitters, because it so overvalues late-inning contributions. Just because a guy missed an opportunity to win a game doesn't mean he lost it.

I like to begin threads like this with just the facts and leave my interpretation until later.

In this case, I think it's actually a positive sign that the guy hurting the offense most during the big slide is the reigning and nearly legitimate MVP (Mauer deserved it, but he was 2nd). You can expect that not to continue.

BTW, "Clutch" for hitters has been shown to be predictive; about 10% of it persists from season to season. I've just discovered that "Clutch" for teams is even more predictive, but the concept gets very tricky (in a nutshell, "Clutch" combines two different things: performing well when it really counts, and performing poorly when it absolutely doesn't.) I may start a thread on that.


#18 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 17 August 2009 - 08:05 PM

QUOTE (Eric Van @ Aug 17 2009, 06:34 PM) <{POST_SNAPBACK}>
I like to begin threads like this with just the facts and leave my interpretation until later.

In this case, I think it's actually a positive sign that the guy hurting the offense most during the big slide is the reigning and nearly legitimate MVP (Mauer deserved it, but he was 2nd). You can expect that not to continue.

BTW, "Clutch" for hitters has been shown to be predictive; about 10% of it persists from season to season. I've just discovered that "Clutch" for teams is even more predictive, but the concept gets very tricky (in a nutshell, "Clutch" combines two different things: performing well when it really counts, and performing poorly when it absolutely doesn't.) I may start a thread on that.

This would be a great thread. I've got some new metrics I've been working on that might have some application in studying clutchiness as well. The main problem with WPA is that puts extraordinary emphasis on leverage - probably way more emphasis than would facilitate a useful measurement. This is mostly a factor of the use of probalistic base-out states which give much greater weight to an event with, for instance, two outs. I think there has to be some levelling of the probability matrix to decrease the out-state warping of the run values in order to get a better measurement. For example, if you have a guy at second and third in the bottom of the ninth inning, WPA will assign a much higher leverage value if there are two outs than if there are none. Delivering the hit that plates both runners for a walk-off is equally clutch regardless of the out state. The only advantage with doing it with nobody out is that if the batter screws up, the next guy has a chance to come through.

I'm not even sure that I want to even consider the score or the inning in looking at clutch hitting. From the point of drama value, late and close situations seem to have more leverage, but are batters really affected by the situation? Does that hit with a runner on second and third really mean more in the 9th inning than in the first? Is it that much more difficult to deliver that hit in a close-late situation against the bullpen than it is with a two-run lead in the third inning against Cole Hamels?

Lots of food for thought here ...

#19 paulftodd


  • 133% banned


  • 1,470 posts

Posted 17 August 2009 - 08:33 PM

QUOTE (Tudor Fever @ Aug 18 2009, 04:40 AM) <{POST_SNAPBACK}>
You don't think that BABIP depends a lot on luck? It seems fairly self-evident.

Edit: bad luck on balls in play will bring down his OPS. I think you may be confusing cause and effect.


Last year Dustin was 389/450/547/997 in high leverage situations. His WPA was 3.21. His BABIP was 419 in high leverage situations. Was he lucky or MVP clutch.

The tendency to attribute luck to any BABIP that lies outside the MLB average, or even the players average, without any supporting data reduces BABIP to a luck-o-meter. We see bloops fall in for hits every day, or seeing eyed grounders, and LD's getting caught, so there is good luck and bad luck. But you can not see this from BABIP alone.

As alluded to already by VAL, BABIP is a function, given a large enough sample, of how well a player hits a ball. The harder a ball gets hit on the ground or a line the more likely it gets through the infield and in the case of a line drive drops without getting caught, with a FB of course, most hits are either bloops or hit far enough to hit the wall or go over it (in which case it is not a BIP and indirectly lowers a sluggers BABIP), or hard enough without being called a LD to drop in the gaps. If a FB stays in the park, the longer a ball stays in the air, the more likely it is to be caught (friendly fenway excepted with balls 20 ft off the wall).

From THT , Dustin has hit 40% FB this year, up from 36% last year. 86% are outs this year vs 83% last year. His HR/FB has dropped from 0.09 to 0.06. He went from + 3 runs vs avg last year on FB to -7 this year. He is hitting fewer GB and more of them are going for outs 72% vs 70%. Maybe fewer of his FB's are going off the wall (his doubles have went way down, 35 last year at home to 18 this year), maybe they are positioning him better on GB's, or he is not hitting his GB's as hard. How do you discern luck here?

His LD rate is about the same as last year. Perhaps fewer are falling in for hits, I don't have this data, but is this just regressing to mean (normal luck) from last years "lucky" season, or was last year normal luck and this year bad luck. If you want to play the luck card, you have to play both the good luck and bad luck cards. Too often luck is only rolled out when one of our guys is not performing as well as expected (not saying you do this).

There is some question as to whether clutch even exists, some say it is just luck, just like BABIP. Do some pitchers pitch better with ROB (as some claimed Dice-K does), or do some hitters hit better with ROB in close and late situations, or is it all luck as to when a hitter gets a hit or HR, maybe some years they come at clutchy times, other years not so clutchy times. I mean is it all luck Papi only has 1 HR and 10 RBI from innings 7-9 (compared to 8 HR, 27 RBI in innings 1-3, and 8 HR, 28 RBI in innings 4-6) or is there some other reason? For reference, in 2005 Papi had 17 HR and 45 RBI's from inning 7-9, but I am not comparing this papi with the 2009 version. If this is the case, WPA is just luck driven by the number of opportunities to hit in high leverage high WPA situations. Lucky hitters with plenty of opportunity have high WPA, unlucky hitters, or opportunity deprived hitters have low WPA. I believe clutch exists, but like RBI's and GIDP, WPA depends on opportunity, and there is some luck involved.

Edited by paulftodd, 17 August 2009 - 08:39 PM.


#20 JMDurron

  • 3,701 posts

Posted 18 August 2009 - 10:26 AM

QUOTE (paulftodd @ Aug 17 2009, 08:33 PM) <{POST_SNAPBACK}>
His LD rate is about the same as last year. Perhaps fewer are falling in for hits, I don't have this data, but is this just regressing to mean (normal luck) from last years "lucky" season, or was last year normal luck and this year bad luck.


According to baseball reference, Pedroia hit 705/699/920 on line drives in 2008, and has hit 753/744/1049 on line drives in 2009. Judging from those numbers, Pedroia is not having a problem with his LDs becoming hits.

#21 Vermonter At Large


  • SoxFan


  • 3,055 posts

Posted 18 August 2009 - 08:59 PM

If we remove the notion of leverage from the equation, Pedroia has actually hit very well with RISP - .343 BA/.862 OPS, so he's doing fine with generic runners in scoring position. I think this points to his poor leverage numbers from BBR as likely spurious.

He has, however, had fewer runners reaching scoring position so far this season. In 2008 he had RISP in 27.2% of his plate appearances, while this season he has only had 24.5% in scoring position. Of those 128 PAs with RISP this season, leadoff hitters have gotten into scoring position 70 times (it would be 78 had Ellsbury batted leadoff all season), so preceding hitters have only reached scoring position in front of Pedroia 58 times. So I think these numbers point to the bottom part of the order as being a factor in h is WPA results.




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users