PDA

View Full Version : Griffey, Phillips, Dunn



RedsManRick
08-16-2007, 06:46 PM
Griffey: .276/.384/.516, .900 OPS
Phillips: .272/.319/.462, .781 OPS
Dunn: .257/.364/.540, 904 OPS

One of these things is not like the others... one of these thing does not belong....

And if you're worried about Dunn batting cleanup against lefties:

Dunn vL: .238/.355/.474, .829 OPS
Phillips vL: .282/.327/.451, .778 OPS

Really, there's just no reason he should be batting cleanup and Dunn should be batting 5th, even to split up the lefties.

High average, low OBP, decent SLG, good speed. Perfect guy for the 5th or 6th spot. Dunn gets on base more, Dunn slugs more. Those more frequent hits of Phillips will help more when he's got an OBP machine ahead of him. And Dunn's homers will help more when he doesn't have a frequent out maker and double play machine ahead of him. Let Phillips drive in Dunn, not the other way around.

top6
08-16-2007, 07:56 PM
I'm sorry, this can't be right, because I've heard nothing but praise for Phillips on the radio whereas Dunn is simply not a run producer.

dougdirt
08-16-2007, 07:58 PM
RMR.... I think you are preaching to the chior, problem is the Reds just don't seem to care about things like stats that make sense.

icehole3
08-16-2007, 08:21 PM
I would say something about Dunn and Griffey"s attitu... nevermind.

VR
08-16-2007, 08:54 PM
I'm totally fine with Phillips in the 4 spot. He's always hit better with men on...he appears to thrive on the chance.


Dunn should hit second in any lineup...end of discussion.

KronoRed
08-16-2007, 09:11 PM
I wish Dunn would say he's a switch hitter so the splitting leftie argument would be moot ;)

RedsBaron
08-16-2007, 09:21 PM
Adam Dunn has never played for a manager who knew where to bat him in the lineup. I agree with VR-bat him second.

edabbs44
08-16-2007, 09:43 PM
Griffey: .276/.384/.516, .900 OPS
Phillips: .272/.319/.462, .781 OPS
Dunn: .257/.364/.540, 904 OPS

One of these things is not like the others... one of these thing does not belong....

And if you're worried about Dunn batting cleanup against lefties:

Dunn vL: .238/.355/.474, .829 OPS
Phillips vL: .282/.327/.451, .778 OPS

Really, there's just no reason he should be batting cleanup and Dunn should be batting 5th, even to split up the lefties.

High average, low OBP, decent SLG, good speed. Perfect guy for the 5th or 6th spot. Dunn gets on base more, Dunn slugs more. Those more frequent hits of Phillips will help more when he's got an OBP machine ahead of him. And Dunn's homers will help more when he doesn't have a frequent out maker and double play machine ahead of him. Let Phillips drive in Dunn, not the other way around.

Dunn has been borderline pathetic vs LHPs this season. It's great that he hit well versus them last season, but the .685 OPS vs LHP in '07 is more telling. Thanks, but this season I'd rather see Phillips breaking up Griffey and Dunn. The stats don't lie. If the opponent brings in a LHP to face Griffey, it's a big decision if they want to leave him in vs Phillips since he has flat out mashed them in '07.

TOBTTReds
08-16-2007, 09:49 PM
Dunn should hit second in any lineup...end of discussion.

Wow, that opens a whole new one. I think he should hit 3rd in most lineups.

RedsManRick
08-16-2007, 10:05 PM
I'd bat him 2nd as well. I don't want him coming up with the bases empty and 2 outs in the first inning 50% of the time. Strikeouts avoid doubles plays and it maximizes his AB and his OBP. He's also a good base runner with decent speed. If he bats cleanup, he's either batting with men on or to leadoff the inning. Anything higher than 5th.

And I'll take the career split numbers over the single year sample. I don't think Dunn's ability to hit lefties has somehow gotten significantly worse this year.

GAC
08-16-2007, 10:08 PM
I think most would tend to agree with you RMR; but here is an interesting scenario....

Like most, I am a firm believer in having your high OB% guys (table setters) at the top of the batting order. At one time I was a strong advocate for Dunn at the #2 spot. But that situation may have changed with players like Hamilton (.377 OB%) and Keppinger (.413 OB%). I liked seeing them at the top of the batting order. Plus, I want to see what these kids can do.

But here are Dunn's stats batting in the #4 spot....


AB R H 2B 3B HR RBI BB HBP SO SB CS AVG OBP SLG OPS
81 17 26 3 0 11 24 10 2 18 0 1 .321 .400 765 1.165

My biggest concern is this team's inability to hit LH'd pitching. And especially guys like Dunn and Jr with RISP...

Jr .... .340 OB% .342 SLG% .184 B/A
Dunn .... .318 OB% .375 SLG% .224 B/A

edabbs44
08-16-2007, 10:09 PM
I'd bat him 2nd as well. I don't want him coming up with the bases empty and 2 outs in the first inning 50% of the time. Strikeouts avoid doubles plays and it maximizes his AB and his OBP. He's also a good base runner with decent speed. If he bats cleanup, he's either batting with men on or to leadoff the inning. Anything higher than 5th.

And I'll take the career split numbers over the single year sample. I don't think Dunn's ability to hit lefties has somehow gotten significantly worse this year.

I'll take performance over ability in determining this year's lineup. If Dunn starts to hit LHPs then he can be moved back to a better spot in the lineup vs LHPs. But for now, he shouldn't be hitting as much vs LHPs as Phillips.

mth123
08-16-2007, 11:01 PM
I'll take performance over ability in determining this year's lineup. If Dunn starts to hit LHPs then he can be moved back to a better spot in the lineup vs LHPs. But for now, he shouldn't be hitting as much vs LHPs as Phillips.

Of course you realize that Phillips gets the benefit of hitting between Griffey and Dunn against lefties so he is naturally the guy that other teams are going to pitch to. He should be outdoing both of them by a wide margin as far as batting average goes.

Dunn on the other hand has had Gonzalez, Conine, or Ross behind him. If he lets those pitches go by to make the pitcher pitch to him and takes his walks (which he should do) he isn't "clutch" and needs to "expand his zone." When he hacks a little at that crap he gets himself out. Its really a situation where the rules have been designed against him and the results (like a .120 or so OPS advantage don't seem to matter). According to some, the only thing that would categorize Dunn as successful is for him to swing at everything and put in the river every time.

I like Phillips a lot, but the fact that he has had a pretty good year while sitting in the catbird seat doesn't change the fact that he is clearly a lesser offensive player than Griffey or Dunn (or even Hamilton). Make no mistake as to who makes this line-up go. Griffey and Dunn are the guys teams pitch around. Phillips enjoys that benefit quite a bit.

edabbs44
08-16-2007, 11:17 PM
Of course you realize that Phillips gets the benefit of hitting between Griffey and Dunn against lefties so he is naturally the guy that other teams are going to pitch to. He should be outdoing both of them by a wide margin as far as batting average goes.

Dunn on the other hand has had Gonzalez, Conine, or Ross behind him. If he lets those pitches go by to make the pitcher pitch to him and takes his walks (which he should do) he isn't "clutch" and needs to "expand his zone." When he hacks a little at that crap he gets himself out. Its really a situation where the rules have been designed against him and the results (like a .120 or so OPS advantage don't seem to matter). According to some, the only thing that would categorize Dunn as successful is for him to swing at everything and put in the river every time.

I like Phillips a lot, but the fact that he has had a pretty good year while sitting in the catbird seat doesn't change the fact that he is clearly a lesser offensive player than Griffey or Dunn (or even Hamilton). Make no mistake as to who makes this line-up go. Griffey and Dunn are the guys teams pitch around. Phillips enjoys that benefit quite a bit.

The excuses for Dunn are getting pretty old. The fact of the matter is that Phillips is hammering LHPs this season. Why change it? Because Dunn doesn't get pitches grooved to him? Because Adam will have to deal with being pitched around?

Phillips has a .958 OPS vs LHPs this season. There is no reason why he can't hit 4th vs LHPs. The only reason why anyone seems to have a problem with Phillips hitting 4th vs LHPs is that his name isn't Adam Dunn. If Conine had a .958 OPS vs LHPs, he should be hitting 4th. Same with Castro or anyone else who is teeing off on lefties.

mth123
08-16-2007, 11:48 PM
The excuses for Dunn are getting pretty old. The fact of the matter is that Phillips is hammering LHPs this season. Why change it? Because Dunn doesn't get pitches grooved to him? Because Adam will have to deal with being pitched around?

Phillips has a .958 OPS vs LHPs this season. There is no reason why he can't hit 4th vs LHPs. The only reason why anyone seems to have a problem with Phillips hitting 4th vs LHPs is that his name isn't Adam Dunn. If Conine had a .958 OPS vs LHPs, he should be hitting 4th. Same with Castro or anyone else who is teeing off on lefties.

Look, I give Phillips credit for taking advantage of the situation. He does have good numbers against lefties this year. But LH or RH if I have Griffey, Phillips and Dunn followed by a selection of out machines, I pitch around Griffey, go after Phillips and then Pitch around Dunn. I do it every time all the time no matter how much Phillips may burn me.

I do think you have one valid argument and that is by hitting Phillips in the middle it makes all three a threat. I don't believe that Phillips would be nearly the threat he is if he were hitting ahead of Gonzalez etc. Griffey and Dunn are always threats and you may be right that the configuration maximizes the whole even if Dunn individually may suffer a bit.

BTW, I really give a lot of credit to Griffey for making a lot of this work. In 2006, Griffey was a major hacker who swung at lots of junk. With Phillips behind him, there was a huge potential for him to be completely taken out of the offense by chasing pitches out of the zone. To his credit he layed off, got his OBP back up where it should be and made pitchers pitch to him. It got him on base ahead of Phillips quite a bit (and really made Phillips a threat with a runner on base and Dunn on deck) and his other numbers all improved accordingly. I believe that if Griffey is hacking like he was in 2006, all three of these guys would have suffered.

RedsManRick
08-16-2007, 11:57 PM
I'll take performance over ability in determining this year's lineup. If Dunn starts to hit LHPs then he can be moved back to a better spot in the lineup vs LHPs. But for now, he shouldn't be hitting as much vs LHPs as Phillips.

In general, proven ability, as established by past performance, is a better predictor of future performance than is a smaller sample of said performance. It's not like I'm using minor league numbers or anything. I'll take Dunn's 1263 PA sample over Phillips' 442.

Not to mention that Phillips has accrued his recent numbers batting largely between Dunn and Junior, where as Dunn has had who, EE? AGone? behind him.

Even so, if I concede the point due to the very large difference this year, what business does BP have batting 4th against righties? That's my primary point.

So fine, BP kills lefties. Bat him 2nd behind Hopper or Freel where we don't really have another great option. He'll get more PA and can put his speed to use when our other guys aren't as productive. Bottom line, there's no reason Dunn should be batting lower than 4th, where it minimizes the value of his OBP and gives more PA to less productive hitters ahead of him.

Batting Phillips 4th against righties is just plain criminally stupid.

flyer85
08-17-2007, 12:08 AM
Dunn has been borderline pathetic vs LHPs this season. It's great that he hit well versus them last season, but the .685 OPS vs LHP in '07 is more telling. Thanks, but this season I'd rather see Phillips breaking up Griffey and Dunn. The stats don't lie. If the opponent brings in a LHP to face Griffey, it's a big decision if they want to leave him in vs Phillips since he has flat out mashed them in '07.last year Jr was bad against LHP(671) and Dunn was good(896). I would say the odds are likely they will both regress to their mean ... and that likely means Dunn would be a better bet for the 4 hole than Phillips.

Against a RH starter I would not break up Jr/Dunn

SteelSD
08-17-2007, 02:33 AM
Batting Phillips 4th against righties is just plain criminally stupid.

Bingo. And frankly, creating a lineup based on small sample size matchups and performance is also criminally stupid.

Welcome to the world of Cincinnati Reds baseball operations.

edabbs44
08-17-2007, 06:47 AM
last year Jr was bad against LHP(671) and Dunn was good(896). I would say the odds are likely they will both regress to their mean ... and that likely means Dunn would be a better bet for the 4 hole than Phillips.

Against a RH starter I would not break up Jr/Dunn

Phillips doesn't have much of a mean. This could be his mean, since he only has 2 years you should really be paying attention to.

jojo
08-17-2007, 07:03 AM
IMHO, Phillips might possibly become one of the more overrated Reds of the 2000's. I'm not suggesting that he hasn't been a positive for the Reds. I'm suggesting that I think a lot of Reds fans believe his abilities/impact have been and will be greater than what they actually are....

jojo
08-17-2007, 07:05 AM
Bingo. And frankly, creating a lineup based on small sample size matchups and performance is also criminally stupid.

Welcome to the world of Cincinnati Reds baseball operations.

That is a huge pet peeve of mine.

edabbs44
08-17-2007, 08:50 AM
Bingo. And frankly, creating a lineup based on small sample size matchups and performance is also criminally stupid.

Welcome to the world of Cincinnati Reds baseball operations.

Small sample size being this entire season so far? 4.5 months? There is such a thing called "riding the hot hand." Why would anyone want Adam Dunn hitting vs a lefty this year? It's wonderful that he's had some success versus them in certain years in the past, but this year he just isn't producing in those situations. So giving the opposition an opportunity to bring in a lefty to materially lessen the production of Griffey and Dunn in one fell swoop in a close situation is borderline criminally stupid, since then that LHP would be replaced with a righty to negate Phillips.

puca
08-17-2007, 09:33 AM
Small sample size being this entire season so far? 4.5 months? There is such a thing called "riding the hot hand." Why would anyone want Adam Dunn hitting vs a lefty this year? It's wonderful that he's had some success versus them in certain years in the past, but this year he just isn't producing in those situations. So giving the opposition an opportunity to bring in a lefty to materially lessen the production of Griffey and Dunn in one fell swoop in a close situation is borderline criminally stupid, since then that LHP would be replaced with a righty to negate Phillips.

Backward thinking it is to create a lineup based soley on the last few innings of a game.

puca
08-17-2007, 09:40 AM
The real problem is the makeup of this team. Their best hitters - in the majors and minors - are all left handed. I like Phillips, but at the rate he produces outs he should be batting no higher than 6th at least against righthanders.

edabbs44
08-17-2007, 09:52 AM
Backward thinking it is to create a lineup based soley on the last few innings of a game.

Bottom line is we are talking about vs LHPs. So if you want to keep the discussion related to starting pitchers, Dunn should not be hitting in front of Phillips in those matchups this season. If he wants to see more ABs vs starting LHPs, he should perform as such.

It is also backwards to think that last year's stats vs LHPs have any meaning when it comes to this year's lineup. So Adam's solid year vs LHPs in '06 is basically meaningless for right now. We can talk more about it if he chooses to show up at some point vs lefties.

flyer85
08-17-2007, 09:54 AM
Phillips doesn't have much of a mean. He has a minor league track record, which is not very impressive. If the Reds had a good RH hitter he would be nowhere near cleanup. 6th or 7th is his likely best spot.

puca
08-17-2007, 10:03 AM
Bottom line is we are talking about vs LHPs. So if you want to keep the discussion related to starting pitchers, Dunn should not be hitting in front of Phillips in those matchups this season. If he wants to see more ABs vs starting LHPs, he should perform as such.

It is also backwards to think that last year's stats vs LHPs have any meaning when it comes to this year's lineup. So Adam's solid year vs LHPs in '06 is basically meaningless for right now. We can talk more about it if he chooses to show up at some point vs lefties.


It seems to me that the discussion was about where Dunn should bat relative to Phillips in general. If you want to change it to where Dunn should bat relative to Phillips against lefthanded starting pitchers, then I might even agree with you. Against righthanded starters Dunn should not be batting behind Phillips yet he consistently does. Just in case the game is close late.

edabbs44
08-17-2007, 10:08 AM
He has a minor league track record, which is not very impressive. If the Reds had a good RH hitter he would be nowhere near cleanup. 6th or 7th is his likely best spot.

For this season, Phillips is perfectly fine hitting 4th vs LHPs. It doesn't matter if his career OPS is .400 in those situations. For now, he mashed lefties and should be seeing them as much as possible. If the situation were to change, then lineup changes should be discussed.

That's the beauty of a lineup. As we have seen, they can change day to day. If Phillips starts to "regress to the mean" vs LHPs, he can be dropped in the order at the drop of a hat. If Dunn starts to hit better vs lefties, he can be moved up pretty quickly.

But why would anyone want Dunn seeing more ABs vs LHPs right now? Because his name is Adam Dunn? As I said before, if Dunn wants to move up in the order vs LHPs, then maybe he should start to produce in those situations. For now, I would much rather see Phillips than Dunn vs a lefty.

RedsManRick
08-17-2007, 10:13 AM
Small sample size being this entire season so far? 4.5 months? There is such a thing called "riding the hot hand."

Yes. Less than 200 is a small sample size. Sure, it tells a better story than 20 or 100, but it's still significantly suspect to random variation.

The 1,090 PA Dunn had against lefties coming in to 2007 hold more weight that do the 173 PA he's had this year.

Furthermore, there's the sniff test. Yes, you can do tests of significance to see if difference is likely due to random variation or is "real". However, before you accept a refutation of the null hypothesis, you should always ask "does this make sense". Does it make sense that an uninjured 28 year old suddenly became 200 OPS points worse against lefties? Frankly, no, it doesn't.

What we often call the hot hand is not a change in the likelihood of an event occurring, but merely a random streak which is prone to end at any point in time. They think, "There's on way I could win 10 times in a row. That's impossible!" Well, it is possible. Just like it's possible for Dunn's true ability to hit lefties (ie. the likely level at which he'll hit lefties in the future) hasn't changed at all despite 173 PA which suggest otherwise.

People cite the "hot hand" in games like roulette all the time. Humans are just prone to assigning meaning to random things -- it's how our brains are wired. Yes, it's possible that Dunn's ability to hit lefties has significantly degraded, but I want more than 173 PA before I accept that.

With BP, given his age and relative lack of a greater sample, I'm willing to accept that he's improving and can hit lefties better than rigties, but the difference isn't as big as the current year to date splits suggest.

And if anybody wants to split up lefties against a righty to avoid the late game LOOGY, they're telling you that 1 at bat in the 8th inning is more important than the 3 at bats they'll get against the righty before that.

Again, if you want Phillips to get more ABs versus lefties, I'm cool with that. But don't take them away from Adam Dunn. Bat BP 2nd. If we have 4 guys on the roster with a track record of hitting lefties better than Dunn, then I'd have no problem moving him down a spot. Absent that, splitting the lefties is done by the manager to make himself feel productive and avoid criticism rather than actually improving our chance to win.


"But why would anyone want Dunn seeing more ABs vs LHPs right now?"
Because "right now" isn't as good of a predictor of what's likely to happen the next time(s) he steps up to the plate as is a track record stretch further back. And because even if BP deserves more ABs, he's just one guy and there's room for him above Dunn without pushing Dunn further down the order.

But really, the point of my post was more in reference to the 70% of at bats in which Phillips continues to bat 4th despite a .688 OPS this year and .673 career OPS. And if we consider Phillips' career splits last year:

03 vR 291 PA, .218/.247/.324, .571 OPS
03 vL 102 PA, .179/.228/.274, .502 OPS (-.069)

It's not the amount of OPS I'm looking at here, but the difference in the split.

06 vR: 436 PA, .268/.315/.429, .744 OPS
06 vL: 151 PA, .299/.351/.423, .774 OPS (+.030)

07 vR: 355 PA, .242/.295/.393, .688 OPS
07 vR: 169 PA, .327/.361/.591, .952 OPS (+.264)

So, did Phillips suddenly learn how to hit lefties and forget how to hit righties, to the tune of going from meh to all-star, or is it more likely that split stats, while significant over large samples, are not as significant over the course of any given year? And if they aren't significant, then that means that they are due more to luck than to a true, repeatable and thus predictive, change in ability level. I imagine that there's a little bit of both going on here. Phillips is probably improving, but the amount is overstated by the year to date split. I'd bet $100 on a regression to the mean next year, even if that mean is a bit higher, call it .750 OPS vR and .850 OPS vL, than it was previously.

edabbs44
08-17-2007, 10:24 AM
Because "right now" isn't as good of a predictor of what's likely to happen the next time(s) he steps up to the plate as is a track record stretch further back. And because even if BP deserves more ABs, he's just one guy and there's room for him above Dunn without pushing Dunn further down the order.

Phillips SLG vs LHPs this season is .591. With the superstars Cincy has had leading off this season, hitting him 2nd would be a travesty.

RedsManRick
08-17-2007, 10:49 AM
Phillips SLG vs LHPs this season is .591. With the superstars Cincy has had leading off this season, hitting him 2nd would be a travesty.

And the following guys have OBP higher than .340 vL this season:

Phillips: .361
Hopper: .359
Junior: .341

Who exactly do you want him to drive in batting 4th? This is the same faulty logic for batting Dunn 5th. If you have great OBP at the top of the lineup already, then keep your great SLG guys behind them. But if your great OBP are your great SLG guys, you don't improve your chances giving a guy with worse OBP and worse SLG more at bats ahead of him.

edabbs44
08-17-2007, 12:17 PM
And the following guys have OBP higher than .340 vL this season:

Phillips: .361
Hopper: .359
Junior: .341

Who exactly do you want him to drive in batting 4th? This is the same faulty logic for batting Dunn 5th. If you have great OBP at the top of the lineup already, then keep your great SLG guys behind them. But if your great OBP are your great SLG guys, you don't improve your chances giving a guy with worse OBP and worse SLG guy more at bats ahead of him.

In that case, I'd rather have the guy who performs better in those situations getting more ABs than those who don't perform as well.

SteelSD
08-17-2007, 12:48 PM
Small sample size being this entire season so far? 4.5 months? There is such a thing called "riding the hot hand." Why would anyone want Adam Dunn hitting vs a lefty this year? It's wonderful that he's had some success versus them in certain years in the past, but this year he just isn't producing in those situations. So giving the opposition an opportunity to bring in a lefty to materially lessen the production of Griffey and Dunn in one fell swoop in a close situation is borderline criminally stupid, since then that LHP would be replaced with a righty to negate Phillips.

Ah. The "hot hand". So far you've determined that Adam Dunn can't hit lefties due to 151 AB this season regardless of his history. In the fourth slot, Phillips is going to need to produce with ducks on the pond...

Phillips w/Runners On 2007: 212 AB- .255 BA/.306 OBP/.415 SLG (.721 OPS), 20 GIDP

Is that the "hot hand"? Now, if we're going to forward the theory that Adam Dunn can't hit lefties because of 150-odd AB this season, then we'd likewise need to conclude that Brandon Phillips shouldn't be anywhere near the fourth slot in the lineup.

If you don't want to see Dunn up versus a LHP, then why would you want to see Brandon Phillips up with ducks on the pond?

That's the problem with lineup construction based on small sample size assumptions. On one hand, such an analysis would provide rationale to break up Griffey and Dunn. But on the other hand, the data suggests that the guy breaking them up (Phillips) doesn't perform well in situations he's likely to see in that lineup spot. Of course, given the larger sample size history of both players, I believe neither. Dunn can hit LHP. Phillips can hit with runners on. But I see no indication that breaking up a powerful LH/LH run in the lineup by handing PA to a league league average offensive 2B (or Conine) is a smart move.

The result of the Reds poor logic is that even with Dunn's 1.165 OPS in 81 cleanup AB this year, they've produced the 5th worst 4-slot OPS in the National League this season. They've been so intent on avoiding LH matchups that they've handed a bunch of PA to hitters who happen to be right-handed but have no business hitting cleanup regardless of the matchups. At best, that's voodoo lineup construction geared toward sacrificing early-game performance in order to avoid late-game matchups that might not come into play and that aren't necessarily bad based on the larger-size historical data.

edabbs44
08-17-2007, 01:15 PM
Those numbers are overall...what are Phillips' numbers vs LHPs with runners on? That would be more relevant for this discussion.

SteelSD
08-17-2007, 01:22 PM
Those numbers are overall...what are Phillips' numbers vs LHPs with runners on? That would be more relevant.

Even if we assume that Phillips takes the same % of his AB versus LHP w/Runners On as he does overall, the result would be 69 AB to this point in the season.

That's actually drilling down to the point of being completely irrelevant. There's simply no predictive value in a sample size that small.

IslandRed
08-17-2007, 01:35 PM
The 1,090 PA Dunn had against lefties coming in to 2007 hold more weight that do the 173 PA he's had this year.

Furthermore, there's the sniff test. Yes, you can do tests of significance to see if difference is likely due to random variation or is "real". However, before you accept a refutation of the null hypothesis, you should always ask "does this make sense". Does it make sense that an uninjured 28 year old suddenly became 200 OPS points worse against lefties? Frankly, no, it doesn't.

What we often call the hot hand is not a change in the likelihood of an event occurring, but merely a random streak which is prone to end at any point in time. They think, "There's on way I could win 10 times in a row. That's impossible!" Well, it is possible. Just like it's possible for Dunn's true ability to hit lefties (ie. the likely level at which he'll hit lefties in the future) hasn't changed at all despite 173 PA which suggest otherwise.

People cite the "hot hand" in games like roulette all the time. Humans are just prone to assigning meaning to random things -- it's how our brains are wired. Yes, it's possible that Dunn's ability to hit lefties has significantly degraded, but I want more than 173 PA before I accept that.


While I'm generally of the analysis bent, I think you have to invoke the "eyes test" as well. The whole notion of regressing to the mean implies that the baseline hasn't changed, that the variations are random in nature. That isn't always the case, and while I haven't attempted to analyze Dunn's at-bats against lefties this year, I'd want to know whether he's still having good at-bats or whether he's genuinely flailing around up there. If there's an actual problem, the regression to the mean doesn't happen until it's fixed. Ask all the Braves fans who have waited all season for Andruw Jones to snap out of it.

That's not saying it's a good idea to make major decisions on a handful of plate appearances. Dunn shouldn't be discounted against lefties based on 173 PAs. At the same time, just waving off bad results as "small sample size" isn't good enough, either, not for the actual decision-makers. As always, it comes down to properly analyzing ALL the sources of information.

edabbs44
08-17-2007, 01:40 PM
Even if we assume that Phillips takes the same % of his AB versus LHP w/Runners On as he does overall, the result would be 69 AB to this point in the season.

That's actually drilling down to the point of being completely irrelevant. There's simply no predictive value in a sample size that small.

So we should ignore ABs vs LHPs when the discussion is mostly focused on that scenario? Makes little to no sense.

RedsManRick
08-17-2007, 01:58 PM
While I'm generally of the analysis bent, I think you have to invoke the "eyes test" as well. The whole notion of regressing to the mean implies that the baseline hasn't changed, that the variations are random in nature. That isn't always the case, and while I haven't attempted to analyze Dunn's at-bats against lefties this year, I'd want to know whether he's still having good at-bats or whether he's genuinely flailing around up there. If there's an actual problem, the regression to the mean doesn't happen until it's fixed. Ask all the Braves fans who have waited all season for Andruw Jones to snap out of it.

That's not saying it's a good idea to make major decisions on a handful of plate appearances. Dunn shouldn't be discounted against lefties based on 173 PAs. At the same time, just waving off bad results as "small sample size" isn't good enough, either, not for the actual decision-makers. As always, it comes down to properly analyzing ALL the sources of information.

This is a great point IR. I would definitely want to be clear that I agree with you 100%. No decision should be made without consideration of all available information, including stats (quantitative), scouting, and other relevant information like the player feels like crap today, is nursing a sore back, etc (qualitative).

However, that's not a defense I've heard invoked in the "bat Dunn 5th versus lefties" argument. If we have both qualitative and quantitative evidence to suggest a real change in ability, then by all means we should act on it appropriately.

Absent qualitative supporting arguments which suggest Dunn is now a worse hitter versus lefties than we has was previously, the quantitative evidence is insufficient for the argument being made.

RedsManRick
08-17-2007, 01:59 PM
So we should ignore ABs vs LHPs when the discussion is mostly focused on that scenario? Makes little to no sense.

We should ignore or at the very least severely discount any statistic based argument when the sample size is insufficient.

edabbs44
08-17-2007, 02:12 PM
We should ignore or at the very least severely discount any statistic based argument when the sample size is insufficient.

Here's the time line:

1) Discussions about whether Phillips or Dunn should be hitting 4th vs LHPs

2) Stats vs LHPs are used as backup for why Phillips should be hitting 4th in those situations, as he has mauled LHPs this year

3) Stats of how Phillips does with runners on enters the fray, showing that Phillips hasn't been doing that well with runners on this year

4) Since discussion is limited to situations involving LHPs, stats are requested in "runners on" situations vs LHPs, since those are the situations being discussed

5) Sample size is described as being insignificant

Now, shouldn't the argument used in #3 be determined as being insignificant, since the discussion is limited to only LHP situations?

nate
08-17-2007, 03:13 PM
Here's the time line:

1) Discussions about whether Phillips or Dunn should be hitting 4th vs LHPs

2) Stats vs LHPs are used as backup for why Phillips should be hitting 4th in those situations, as he has mauled LHPs this year

3) Stats of how Phillips does with runners on enters the fray, showing that Phillips hasn't been doing that well with runners on this year

4) Since discussion is limited to situations involving LHPs, stats are requested in "runners on" situations vs LHPs, since those are the situations being discussed

5) Sample size is described as being insignificant

Now, shouldn't the argument used in #3 be determined as being insignificant, since the discussion is limited to only LHP situations?

Your timeline doesn't make the sample size of the statistic you requested any larger. But because I care, here (http://www.baseball-reference.com/pi/shareit/4QuZ) ya go:




Brandon Phillips
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
45 63 59 18 3 0 3 18 3 2 7 0 0 1 1 5 .305 .333 .508 .842

jojo
08-17-2007, 03:14 PM
Here's the time line:

1) Discussions about whether Phillips or Dunn should be hitting 4th vs LHPs

2) Stats vs LHPs are used as backup for why Phillips should be hitting 4th in those situations, as he has mauled LHPs this year

3) Stats of how Phillips does with runners on enters the fray, showing that Phillips hasn't been doing that well with runners on this year

4) Since discussion is limited to situations involving LHPs, stats are requested in "runners on" situations vs LHPs, since those are the situations being discussed

5) Sample size is described as being insignificant

Now, shouldn't the argument used in #3 be determined as being insignificant, since the discussion is limited to only LHP situations?

The argument really has been that splits are being misapplied to support the conclusion that Phillips should break up Jr and Dunn.

Splits are only interesting to the extent that they inform a player's true skill level.

Splits actually are meaningless a majority of the time because even though they may speak directly to a specific point of debate, they very often suffer from a lack of information. Context is everything and all too often, too much can be made of a split. The reason people harp on sample size is because players can often experience streaks of wild variation in their results such that their numbers in small doses frankly do not reflect their true repeatable skill set.

Generally, many fans error by approaching player evaluation from a results-based standpoint. Splits used in this way all too often are nothing more than fool's gold made more attractive because the split itself adds an illusion of precision even though the data it holds can be dramatically influenced by randomness.

So really splits need to be eyed with a lot of caveats and caution rather than profited with blanket trust.

edabbs44
08-17-2007, 03:25 PM
This thread is a tremendous example of how RZ posters love to twist the facts and manipulate real life in an effort to pump up Adam Dunn.

edabbs44
08-17-2007, 03:26 PM
Your timeline doesn't make the sample size of the statistic you requested any larger. But because I care, here (http://www.baseball-reference.com/pi/shareit/4QuZ) ya go:




Brandon Phillips
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
45 63 59 18 3 0 3 18 3 2 7 0 0 1 1 5 .305 .333 .508 .842


I only said it b/c I was told that Phillips shouldn't hit 4th vs LHPs, since he struggles with runners on. It's clearer now that he doesn't struggle with runners on vs LHPs, so that argument is worthless.

RedsManRick
08-17-2007, 03:33 PM
Here's the time line:

1) Discussions about whether Phillips or Dunn should be hitting 4th vs LHPs

2) Stats vs LHPs are used as backup for why Phillips should be hitting 4th in those situations, as he has mauled LHPs this year

3) Stats of how Phillips does with runners on enters the fray, showing that Phillips hasn't been doing that well with runners on this year

4) Since discussion is limited to situations involving LHPs, stats are requested in "runners on" situations vs LHPs, since those are the situations being discussed

5) Sample size is described as being insignificant

Now, shouldn't the argument used in #3 be determined as being insignificant, since the discussion is limited to only LHP situations?

Absolutely, Steel's post was giving you the benefit of the doubt. Perhaps you missed it, because it was implied. His point was:

1.) The sample size problem renders the reliance on in-season data moot.
2.) If we ignore point 1, you would still be wrong. (which perhaps you've shown to be based on incorrect data).

Let's keep things civil here. We're not "manipulating real life". We're using data to reach conclusions which requires a set of assumptions which cannot be overlooked. Calling in to question certain conclusions because they are based on faulty assumptions is not twisting facts. They are faulty not because we say so, but because the rules of statistical analysis, as taut in any stats class in the country, say so.

Nobody has a hidden agenda to pump up Adam Dunn. There's nothing to be gained. We all want the Reds to win the most games as possible and to do so requires the best use of the available talent. What is clear is that there is a miscommunication between the points being made -- I would posit it is largely due to a difference in background with statistical analysis. In any case, maybe that's our fault for not communicating our point of view clearly or confusing the conversation with irrelevant side points.

To simplify, here are my points. Let me know where you disagree.

1.) To date, in their careers, Adam Dunn has hit lefties better than Brandon Phillips.
2.) To date, this season, Brandon Phillips has hit lefties much better than Adam Dunn.
3.) The amount of PA each has had versus lefties this season is not enough to conclude that, moving forward, Brandon Phillips is likely to be more productive than Adam Dunn against lefties.
4.) Point 1 must serve as our working "null" hypothesis as it is the only one based in a large enough sample of events.
5.) If we were to move on the assumption that Phillips really is the superior hitter versus lefties, the conclusion I reach is that he should bat higher in the lineup than 4th, because other available, significantly inferior options.

CTA513
08-17-2007, 03:41 PM
This thread is a tremendous example of how RZ posters love to twist the facts and manipulate real life in an effort to pump up Adam Dunn.

Bat Phillips 4th against LHP and don't bat him 4th against RHP.
The Reds see more RHP than LHP and Phillips will still be batting 4th anyway.

nate
08-17-2007, 03:42 PM
I only said it b/c I was told that Phillips shouldn't hit 4th vs LHPs, since he struggles with runners on. It's clearer now that he doesn't struggle with runners on vs LHPs, so that argument is worthless.

You should compare this season to his career numbers under those circumstances.

edabbs44
08-17-2007, 03:43 PM
Absolutely, Steel's post was giving you the benefit of the doubt. Perhaps you missed it, because it was implied. His point was:

1.) The sample size problem renders this conversation moot.
2.) If we ignore point 1, you're still wrong.

That's nice.

This is starting to get ridiculous. Phillips hammers lefties this season. Slice it up into whatever sample size you like.

He outperforms Adam Dunn vs LHPs in 2007 in any way you would like to manipulate the numbers.

You guys can use last year all you want...that doesn't make him perform any better vs LHPs this year. If that's the case, maybe Stanton should be closing games out and Coffey should be setting him up.

edabbs44
08-17-2007, 03:43 PM
Bat Phillips 4th against LHP and don't bat him 4th against RHP.
The Reds see more RHP than LHP and Phillips will still be batting 4th anyway.

I agree.

nate
08-17-2007, 03:47 PM
This thread is a tremendous example of how RZ posters love to twist the facts and manipulate real life in an effort to pump up Adam Dunn.

Adam Dunn's numbers pump him up:




Adam Dunn
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
118 493 417 108 20 2 32 79 69 5 134 3 0 4 3 10 .259 .365 .547 .912

BRM
08-17-2007, 03:50 PM
The tag says 2007 but those must be his career numbers vs LHP with runners on. Pretty impressive though.

edabbs44
08-17-2007, 03:52 PM
Adam Dunn's numbers pump him up:




Adam Dunn
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
118 493 417 108 20 2 32 79 69 5 134 3 0 4 3 10 .259 .365 .547 .912


You might want to rerun that search.

Patrick Bateman
08-17-2007, 03:52 PM
This thread is a tremendous example of how RZ posters love to twist the facts and manipulate real life in an effort to pump up Adam Dunn.

I'd suggest it's the other way around.

edabbs44
08-17-2007, 03:52 PM
The tag says 2007 but those must be his career numbers vs LHP with runners on. Pretty impressive though.

Those are this year's total numbers.

edabbs44
08-17-2007, 03:53 PM
I'd suggest it's the other way around.

To each their own.

BRM
08-17-2007, 03:56 PM
Those are this year's total numbers.

Ah, okay.

nate
08-17-2007, 03:58 PM
Ah yes!

It did seem like a lot of plate appearances in that situation:




Adam Dunn
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
57 84 71 11 0 1 1 12 12 0 30 1 0 0 1 6 .155 .286 .225 .511

Man, just smack another single a week and you'll be smokin' in these situations, Adam!

RedsManRick
08-17-2007, 03:59 PM
That's nice.

This is starting to get ridiculous. Phillips hammers lefties this season. Slice it up into whatever sample size you like.

He outperforms Adam Dunn vs LHPs in 2007 in any way you would like to manipulate the numbers.

You guys can use last year all you want...that doesn't make him perform any better vs LHPs this year. If that's the case, maybe Stanton should be closing games out and Coffey should be setting him up.

Nobody disagrees with what they've done to date this year. Phillips has clearly outhit done versus lefties this year.

What we disagree with is what they are likely to do in the future.

These are two very different arguments.

dabvu2498
08-17-2007, 04:11 PM
Some good lineup reading:
http://www.hardballtimes.com/main/article/constructing-lineups/
http://www.retrosheet.org/Research/RuaneT/lineup_art.htm

As much as I love lineup talk, I hate to say that I've kinda fallen in with this quote from Bill James:


Several people, maybe a dozen, have done simulation studies of lineups, and have all (as far as I know) reported that it really doesn't make any difference, that one lineup is as good as another.

Of course, his next line was "I still don't buy it" and I'm not 100% sure I do either. The Book says


Your three best hitters should bat somewhere in the #1, #2 and #4 slots. Your fourth- and fifth-best hitters should occupy the #3 and #5 slots. The #1 and #2 slots will have players with more walks than those in the #4 and #5 slots. From slot #6 through #9, put the players in descending order of quality.


Sounds about right to me.

But what are we talking about in the long run here? 1 win a season? 3 wins a season? 5? 7???? if your lineup is optimally constructed. And would that be assuming that the other team's lineup is not optimally contructed as well???

Anyway, as Steel said in another thread...


First, the failure of this team to win anything has everything to do with pitching and nothing to do with the Reds offensive performance.
http://www.redszone.com/forums/showpost.php?p=1444575&postcount=92

Complaining about the Reds' lineups is like complaining about the Confederates wearing only grey during the Civil War.

VR
08-17-2007, 04:19 PM
Ah yes!

It did seem like a lot of plate appearances in that situation:




Adam Dunn
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
57 84 71 11 0 1 1 12 12 0 30 1 0 0 1 6 .155 .286 .225 .511

Man, just smack another single a week and you'll be smokin' in these situations, Adam!


To me that speaks a lot about being pitched around. Just adds to the 'bat Dunn second' argument.

The guy needs protection to force pitchers to pitch to him.

edabbs44
08-17-2007, 04:20 PM
Nobody disagrees with what they've done to date this year. Phillips has clearly outhit done versus lefties this year.

What we disagree with is what they are likely to do in the future.

These are two very different arguments.

My argument is let Phillips hit 4th vs LHPs until there is a reason to switch him and Dunn in the order. Dunn may very well start to hit better vs LHPs. But until that time, Phillips should be hitting ahead of him in those situations.

edabbs44
08-17-2007, 04:22 PM
To me that speaks a lot about being pitched around. Just adds to the 'bat Dunn second' argument.

The guy needs protection to force pitchers to pitch to him.

No offense to you, but it is never Dunn's fault.

I guess only LHPs pitch around Dunn when there are runners on? He doesn't seem to have much of a problem in those situations vs RHPs.

jojo
08-17-2007, 04:55 PM
Your timeline doesn't make the sample size of the statistic you requested any larger. But because I care, here (http://www.baseball-reference.com/pi/shareit/4QuZ) ya go:




Brandon Phillips
2007
vs. LHP
with runners on

G PA AB H 2B 3B HR RBI BB IBB SO HBP SH SF ROE GDP BA OBP SLG OPS

+----+-----+-----+----+---+---+---+----+----+---+----+---+---+---+---+---+-----+-----+-----+-----+
45 63 59 18 3 0 3 18 3 2 7 0 0 1 1 5 .305 .333 .508 .842


Here's the danger of viewing such small samples as being meaningful...

Take out the two at bats Phillips had against Randy Keisler (who incidentally is pitching now for the ever dangerous Memphis Redbirds) and Phillips' line against lefties with runners on looks like this:

.281/.339/.421 OPS: .760

For comparison, here is Phillips' career line against lefties:

.281/.327/.450 OPS: .777

edabbs44
08-17-2007, 05:18 PM
Here's the danger of viewing such small samples as being meaningful...

Take out the two at bats Phillips had against Randy Keisler (who incidentally is pitching now for the ever dangerous Memphis Redbirds) and Phillips' line against lefties with runners on looks like this:

.281/.339/.421 OPS: .760

For comparison, here is Phillips' career line against lefties:

.281/.327/.450 OPS: .777

How did his OBP go up?

You can do that with anyone. But take those two ABs away from his overall numbers vs LHPs and they don't go down as significantly as they do with runners on. He is still doing fairly well.

However, I was under the impression that stats like "with runners on" are meaningless, especially when Dunn is in the conversation. I must have missed the memo that said these are relevant again.

VR
08-17-2007, 05:28 PM
No offense to you, but it is never Dunn's fault.

I guess only LHPs pitch around Dunn when there are runners on? He doesn't seem to have much of a problem in those situations vs RHPs.

this year, or in his career?

SteelSD
08-17-2007, 08:00 PM
How did his OBP go up?

You can do that with anyone. But take those two ABs away from his overall numbers vs LHPs and they don't go down as significantly as they do with runners on. He is still doing fairly well.

However, I was under the impression that stats like "with runners on" are meaningless, especially when Dunn is in the conversation. I must have missed the memo that said these are relevant again.

You've again missed the point. No one is saying that Phillips' 2007 numbers with ducks on the pond has any predictive value. Well, no one but you as it appears that you assign great predictive value to sample sizes of 150 to 212 AB. From there you've had to drill down to an even smaller sample size (59 AB) for Phillips vs. LHP w/Runners On while also assigning predictive value to that sample size. What jojo is showing you is how easy it is for those small sample sizes to be impacted by random single AB events. That randomness is why small sample size data isn't predictive in the first place.

And stop with the whole "Adam Dunn bias" thing. This isn't about Dunn. It's about the data. Doesn't matter who the players are.

edabbs44
08-17-2007, 08:47 PM
You've again missed the point. No one is saying that Phillips' 2007 numbers with ducks on the pond has any predictive value. Well, no one but you as it appears that you assign great predictive value to sample sizes of 150 to 212 AB. From there you've had to drill down to an even smaller sample size (59 AB) for Phillips vs. LHP w/Runners On while also assigning predictive value to that sample size. What jojo is showing you is how easy it is for those small sample sizes to be impacted by random single AB events. That randomness is why small sample size data isn't predictive in the first place.

And stop with the whole "Adam Dunn bias" thing. This isn't about Dunn. It's about the data. Doesn't matter who the players are.

You are missing my point by a rather wide margin. I am not predicting anything. I am not saying that Phillips will do anything. What I am saying is that he is doing materially better in 2007 vs LHPs than Adam Dunn is.

So, why not hit him 4th vs LHPs for the time being? Lineups are great because they only need to be in effect for one game at a time. If a team had to put their lineup for the rest of the year in during spring training, I would probably agree with you. But that's not how it works. You don't have to predict anything. If Dunn sarts producing vs LHPs, then move him up. If Phillips starts to crap out in the same situation, move him down.

You say it's about the data? The data shows that Phillips is dwarfing Dunn's production for the first 4.5 months of the season vs LHPs. That's not a sample size anymore, it's reality. Face facts.

And FYI, here's where the whole sample size thing started:


Ah. The "hot hand". So far you've determined that Adam Dunn can't hit lefties due to 151 AB this season regardless of his history. In the fourth slot, Phillips is going to need to produce with ducks on the pond...

Phillips w/Runners On 2007: 212 AB- .255 BA/.306 OBP/.415 SLG (.721 OPS), 20 GIDP

Is that the "hot hand"? Now, if we're going to forward the theory that Adam Dunn can't hit lefties because of 150-odd AB this season, then we'd likewise need to conclude that Brandon Phillips shouldn't be anywhere near the fourth slot in the lineup.

If you don't want to see Dunn up versus a LHP, then why would you want to see Brandon Phillips up with ducks on the pond?

That's the problem with lineup construction based on small sample size assumptions. On one hand, such an analysis would provide rationale to break up Griffey and Dunn. But on the other hand, the data suggests that the guy breaking them up (Phillips) doesn't perform well in situations he's likely to see in that lineup spot. Of course, given the larger sample size history of both players, I believe neither. Dunn can hit LHP. Phillips can hit with runners on. But I see no indication that breaking up a powerful LH/LH run in the lineup by handing PA to a league league average offensive 2B (or Conine) is a smart move.

The result of the Reds poor logic is that even with Dunn's 1.165 OPS in 81 cleanup AB this year, they've produced the 5th worst 4-slot OPS in the National League this season. They've been so intent on avoiding LH matchups that they've handed a bunch of PA to hitters who happen to be right-handed but have no business hitting cleanup regardless of the matchups. At best, that's voodoo lineup construction geared toward sacrificing early-game performance in order to avoid late-game matchups that might not come into play and that aren't necessarily bad based on the larger-size historical data.

The above argument holds no merit because you took a discussion about where Phillips and Dunn should hit vs LHPs and started talking theoretically why Phillips shouldn't be hitting 4th because of stats taken from situaitons outside of the LHP one we were discussing. When I said you should look at that data only vs LHPs if you were going to make that argument, that's when everyone started going off because of the small sample size.

Read the posts before freaking out about small sample sizes. I in no way brought that up. I only brought it up because you distorted the stats in your favor for unknown reasons.

jojo
08-17-2007, 09:43 PM
What I am saying is that he is doing materially better in 2007 vs LHPs than Adam Dunn is.

Phillips HAS done better than Dunn versus LHPs thus far in '07. There is no "is doing". You're referencing history absent a tangible underlying explanation for why this might be expected to continue. It's not a trivial point especially for this discussion.


So, why not hit him 4th vs LHPs for the time being?

Why not? Well here's one reason-each slot moved down in the order is associated with roughly 20 less PAs over the course of a season versus the spot preceding it. Batting Dunn 5th takes PAs away from him in favor of a hitter that has inferior repeatable skills.

Apparently the rationale for this is that Phillips has been uncharacteristically hot against LHPs this year while Dunn has had lower than normal numbers against them. You avoid actually stating it but the implication is that in their next at bat against lefties, Phillips should therefore be expected to be more effective than Dunn. It's possible of course, Aurillia road a lucky streak against lefties for better than a year as a Red. It's just not the best bet when considering their true repeatable skill levels. Really this discussion has less to do with Dunn/Phillips than it does with rationale.


Lineups are great because they only need to be in effect for one game at a time.

So now the appropriate sample size is down to 4-5 PAs?


You say it's about the data? The data shows that Phillips is dwarfing Dunn's production for the first 4.5 months of the season vs LHPs. That's not a sample size anymore, it's reality. Face facts.

Once again, the key word should be HAS rather than is.... Their splits for this year are not nearly as informative for their next PA as their career splits.


If Dunn sarts producing vs LHPs, then move him up. If Phillips starts to crap out in the same situation, move him down.

Here is their relative performances over their last 25 PA's against LHPs which is roughly their last 5 games (please feel free to double check my spreadsheet):

Phillips: .174/.269/.217 OPS: .487
Dunn: .278/.480/.556 OPS: .948

How many games does it take to determine if Dunn is producing vs LHPs or if Phillips is starting to crap out in the same situation?


You're a smart fellow. I have no idea why you consistently chose to be what I perceive to be purposefully obtuse on the issue of sample size and splits.

SteelSD
08-18-2007, 12:56 AM
You are missing my point by a rather wide margin. I am not predicting anything.

Yes you are.


So, why not hit him 4th vs LHPs for the time being?

And there it is.

Your rationale is:

Player A hasn't produced as well as Player B versus LHP this season, therefore Player B should be hitting higher in the lineup versus LHP than Player A until Player B disproves that he can hit LHP.

Problem is that you're not using anything resembling large enough sample sizes to jump to the initial conclusion, nor do you have any way of knowing when Player B will stop producing in those situations. You are speaking to probability when you state that Player B should be taking away PA from Player A. As jojo aptly noted, you've produced an opinion that Player B's performance should be expected to continue. No two ways about it.

jojo has also done an excellent job noting the severe issues with your low sample size "hot hand" approach. The most recent 25 PA he noted equal around 14% to 15% of each player's total PA versus LHP this season. If your position is that Phillips was "hot" over a small sample of unrelated PA prior to today, is that enough to figure out that he's no longer "hot"? Or does he need to put up twice as many PA? What's the sample size you're looking for? Well, that's rhetorical because the sample sizes you appear to be comfortable with as predictors simply aren't large enough to be predictive.


And FYI, here's where the whole sample size thing started:

...

The above argument holds no merit because you took a discussion about where Phillips and Dunn should hit vs LHPs and started talking theoretically why Phillips shouldn't be hitting 4th because of stats taken from situaitons outside of the LHP one we were discussing. When I said you should look at that data only vs LHPs if you were going to make that argument, that's when everyone started going off because of the small sample size.

Read the posts before freaking out about small sample sizes. I in no way brought that up. I only brought it up because you distorted the stats in your favor for unknown reasons.

Actually, you were in real trouble regarding sample sizes well before I posted the response you cited. Rick was well ahead of me in that respect at the top of page 2. All I did was push you further toward the corner you ended up painting yourself into.

Don't think I wasn't expecting your demand that Phillips' performance be drilled down to how he hit vs. LHP w/Runners On this season. If you accepted Dunn's sample size vs. LHP as "truth", then you had no choice but to accept Phillips' Runner On sample size as "truth" as well, unless you attempted to drill down to an even less predictive sample size (59 AB). In fact, that's what I was hoping for because it left you in another completely untenable position.

Sorry, but you were set up without any stat twisting or "distorting". Such claims are completely moot- especially considering that I clearly noted that I didn't believe in the predictive nature of any of the small sample size data (2007 Dunn vs. LHP, Phillips w/Runner On) I cited. After I brought up Phillips' overall performance with ducks on the pond, you (predictably) immediately attempted to drill down to a completely non-predictive 60-odd unrelated PA sample in those situations to justify Phillips' presence in the 4-slot in the lineup over a better hitter- "matchups" or otherwise.

What you've forwarded is exactly the kind of reasoning I was speaking to in my initial post in this thread- justification the press can get sold on for why a manager produced a poor lineup or made a poor "matchup" decision, but nothing resembling actual logic based on probability. But as long as the manager has a "reason" (as illogical as it may be) he's offered quarter.

And most disappointing, you've decided to again play both the "stat twisting" card as well as the "Dunn bias" card in this thread. Yet the only person actually drawing conclusions from non-predictive sample sizes is you and it appears that only you care about the names of the players. Hmn...

osuceltic
08-18-2007, 01:09 AM
If you'd stop talking about sample sizes for a minute, you could consider some things your eyes can tell you. Dunn is bailing against left-handers this season. Hasn't always been the case, and there are times when he hangs in there, but this season-long struggle boils down not to some statistical anomaly, but to the very real fact that Dunn has a mechanical flaw right now that is affecting his performance against left-handers. It's not as obvious as Encarnacion or Ross pulling off of every pitch, but it's pretty obvious if you watch Dunn consistently.

When I see Dunn put a couple of weeks together where he hangs in there against lefties and puts some good swings on the ball, maybe then I consider making a change. But he hasn't had a stretch like that this season.

SteelSD
08-18-2007, 01:31 AM
If you'd stop talking about sample sizes for a minute, you could consider some things your eyes can tell you. Dunn is bailing against left-handers this season. Hasn't always been the case, and there are times when he hangs in there, but this season-long struggle boils down not to some statistical anomaly, but to the very real fact that Dunn has a mechanical flaw right now that is affecting his performance against left-handers. It's not as obvious as Encarnacion or Ross pulling off of every pitch, but it's pretty obvious if you watch Dunn consistently.

When I see Dunn put a couple of weeks together where he hangs in there against lefties and puts some good swings on the ball, maybe then I consider making a change. But he hasn't had a stretch like that this season.

You mean that he needs a stretch of performance versus lefties like the one jojo noted in his last post?

osuceltic
08-18-2007, 11:17 AM
You mean that he needs a stretch of performance versus lefties like the one jojo noted in his last post?

25 PAs? Not quite. And let's see if he's staying in better or just getting mistake pitches he can pull. I'm not sure yet.

But if he can correct the glaring flaw in his swing that has given him so many problems this season, then it's absolutely time to reevaluate how to use him.

edabbs44
08-18-2007, 03:50 PM
Yes you are.



And there it is.

Your rationale is:

Player A hasn't produced as well as Player B versus LHP this season, therefore Player B should be hitting higher in the lineup versus LHP than Player A until Player B disproves that he can hit LHP.

Problem is that you're not using anything resembling large enough sample sizes to jump to the initial conclusion, nor do you have any way of knowing when Player B will stop producing in those situations. You are speaking to probability when you state that Player B should be taking away PA from Player A. As jojo aptly noted, you've produced an opinion that Player B's performance should be expected to continue. No two ways about it.

jojo has also done an excellent job noting the severe issues with your low sample size "hot hand" approach. The most recent 25 PA he noted equal around 14% to 15% of each player's total PA versus LHP this season. If your position is that Phillips was "hot" over a small sample of unrelated PA prior to today, is that enough to figure out that he's no longer "hot"? Or does he need to put up twice as many PA? What's the sample size you're looking for? Well, that's rhetorical because the sample sizes you appear to be comfortable with as predictors simply aren't large enough to be predictive.



Actually, you were in real trouble regarding sample sizes well before I posted the response you cited. Rick was well ahead of me in that respect at the top of page 2. All I did was push you further toward the corner you ended up painting yourself into.

Don't think I wasn't expecting your demand that Phillips' performance be drilled down to how he hit vs. LHP w/Runners On this season. If you accepted Dunn's sample size vs. LHP as "truth", then you had no choice but to accept Phillips' Runner On sample size as "truth" as well, unless you attempted to drill down to an even less predictive sample size (59 AB). In fact, that's what I was hoping for because it left you in another completely untenable position.

Sorry, but you were set up without any stat twisting or "distorting". Such claims are completely moot- especially considering that I clearly noted that I didn't believe in the predictive nature of any of the small sample size data (2007 Dunn vs. LHP, Phillips w/Runner On) I cited. After I brought up Phillips' overall performance with ducks on the pond, you (predictably) immediately attempted to drill down to a completely non-predictive 60-odd unrelated PA sample in those situations to justify Phillips' presence in the 4-slot in the lineup over a better hitter- "matchups" or otherwise.

What you've forwarded is exactly the kind of reasoning I was speaking to in my initial post in this thread- justification the press can get sold on for why a manager produced a poor lineup or made a poor "matchup" decision, but nothing resembling actual logic based on probability. But as long as the manager has a "reason" (as illogical as it may be) he's offered quarter.

And most disappointing, you've decided to again play both the "stat twisting" card as well as the "Dunn bias" card in this thread. Yet the only person actually drawing conclusions from non-predictive sample sizes is you and it appears that only you care about the names of the players. Hmn...

Whatever you say.

The sample size I used was 100% of this season vs LHPs. Not a sample size, but rather an entire population. If you want to use career numbers in a vacuum then Phillips shouldn't have even stepped on the field last season. Hell, Krivsky should have not even bothered to acquire him since the largest sample size you could have gotten would have showed that he flat-out sucked. A case could even be made that Bellhorn should be playing over Phillips, since he has the higher career OPS than BP.

But thanks anyway for trying to "set me up." Very Henry Gondorff like.

jojo
08-18-2007, 04:23 PM
The sample size I used was 100% of this season vs LHPs. Not a sample size, but rather an entire population.

If the conversation is about estimating a player's true repeatable performance level as accurately as possible, an entire season's PAs against LHPs is not a population. It's simply a larger sample than a more defined split.

150 PA's is simply too small to meaningfully estimate a player's true skill level in the face of the typical variation inherent in outcomes. If an estimate can't serve as an accurate picture of a player's true skill level then it stands to reason then that the estimate has no predictive power.

RedsBaron
08-18-2007, 08:25 PM
No offense to you, but it is never Dunn's fault.



To some people it seems as if it is ALWAYS Dunn's fault. ;)