SoxProspects News
|
|
|
|
Legal
Forum Ground Rules
The views expressed by the members of this Forum do not necessarily reflect the views of SoxProspects, LLC.
© 2003-2024 SoxProspects, LLC
|
|
|
|
|
Forum Home | Search | My Profile | Messages | Members | Help |
Welcome Guest. Please Login or Register.
2016 Red Sox Rotation Discussion
|
Post by jmei on Jul 20, 2016 15:07:47 GMT -5
It's amazingly speculative, almost irresponsibly so, because your hypothesis now comes replete with character judgments about players' personal lives.
|
|
|
Post by James Dunne on Jul 20, 2016 15:13:41 GMT -5
en.wikipedia.org/wiki/Fooled_by_RandomnessEDIT: Also, if lavarnwayguy had claimed the Rodriguez had pitched poorly because Latinos love to party in LA, he'd have been crucified. It shouldn't be any different just because there are cherry-picked numbers alongside improvable presumptions.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Jul 22, 2016 8:43:02 GMT -5
It's amazingly speculative, almost irresponsibly so, because your hypothesis now comes replete with character judgments about players' personal lives. en.wikipedia.org/wiki/Fooled_by_RandomnessEDIT: Also, if lavarnwayguy had claimed the Rodriguez had pitched poorly because Latinos love to party in LA, he'd have been crucified. It shouldn't be any different just because there are cherry-picked numbers alongside improvable presumptions. I should make it clear that it's not a character judgment whatsoever. I would say that going out with your teammates in the evening is normative (especially on an off day). Nor am I talking about doing anything that would register as misbehavior on anyone's radar. Getting one hour less sleep than normal, for a guy who ordinarily keeps a regular schedule, would be plenty sufficient as a cause. I think it's hopelessly naive to think that it's a coincidence that the only two road day games -- which is to say, the only two games where getting less sleep than on average would be normative -- are the incredibly large outliers in his performance. Using t-tests on xFIP-, there's no significant difference between home games at day, home games at night, or road games at night. You have two outliers, the day games on the road, with p=.001 for being random versus the other 17 starts. That's two-tailed. The odds of having two starts that much worse than the other 17 at random are 1 in 1762. Now, what are the odds that those two outliers were the two day games on the road? 180 to 1 (p = 2 / 19^2). So we're talking about a statistical anomaly that has a 318,278 to 1 chance of happening at random. Which is to say, if you programmed a random number generator to spit out xFIP- scores for E-Rod games based on the observed distribution, with an assumption for normality, you would get two games this bad or worse in his only two road day games once every 318,278 of your simple brainless simulation. That is the fact that we are starting with here. And, oh yeah ... once again, the variance in pitching performance across outings not only isn't random, it should not be treated as such (unlike, say, BABIP). The quality of the pitches is a direct product of general neurological function. Yes, what happens when the pitch gets to the plate has a large treat-it-as-random factor, but big splits in pitching performance -- by home / road, day / night, days rest, and so on -- should always be treated first as real when they are significant enough statistically (after you correct for the number of splits you're looking at). Edit: That the cities were L.A. and Miami (as opposed to, say, Minnesota and Pittsburgh) just makes his behavior more normative. That's why I asked jimoh if he would have gone out with his teammates in L.A. under those circumstances. I'm more aware of how sleep affects my day-to-day function than 999 of 100 people, and I probably would have done the same thing. It's worth noting that the Anaheim start was far worse; in Miami, he settled down after two rough innings but couldn't get anyone out on the 6th, blowing his line back up. That's only an outlier by ERA-, because the bad was clustered where you might well expect it to be with a guy just a bit shy of sleep (after learning a lesson in L.A.) -- finding his mechanics, and stamina. Yeah, I'm really constructing a narrative now ... but if you use ERA- instead of xFIP-, which is, after all, what actually determines his value, the above odds go up to 521,804,716 to 1.
|
|
|
Post by jimedsright on Jul 22, 2016 9:15:32 GMT -5
It's amazingly speculative, almost irresponsibly so, because your hypothesis now comes replete with character judgments about players' personal lives. en.wikipedia.org/wiki/Fooled_by_RandomnessEDIT: Also, if lavarnwayguy had claimed the Rodriguez had pitched poorly because Latinos love to party in LA, he'd have been crucified. It shouldn't be any different just because there are cherry-picked numbers alongside improvable presumptions. I should make it clear that it's not a character judgment whatsoever. I would say that going out with your teammates in the evening is normative (especially on an off day). Nor am I talking about doing anything that would register as misbehavior on anyone's radar. Getting one hour less sleep than normal, for a guy who ordinarily keeps a regular schedule, would be plenty sufficient as a cause. I think it's hopelessly naive to think that it's a coincidence that the only two road day games -- which is to say, the only two games where getting less sleep than on average would be normative -- are the incredibly large outliers in his performance. Using t-tests on xFIP-, there's no significant difference between home games at day, home games at night, or road games at night. You have two outliers, the day games on the road, with p=.001 for being random versus the other 17 starts. That's two-tailed. The odds of having two starts that much worse than the other 17 at random are 1 in 1762. Now, what are the odds that those two outliers were the two day games on the road? 180 to 1 (p = 2 / 19^2). So we're talking about a statistical anomaly that has a 318,278 to 1 chance of happening at random. Which is to say, if you programmed a random number generator to spit out xFIP- scores for E-Rod games based on the observed distribution, with an assumption for normality, you would get two games this bad or worse in his only two road day games once every 318,278 of your simple brainless simulation. That is the fact that we are starting with here. And, oh yeah ... once again, the variance in pitching performance across outings not only isn't random, it should not be treated as such (unlike, say, BABIP). The quality of the pitches is a direct product of general neurological function. Yes, what happens when the pitch gets to the plate has a large treat-it-as-random factor, but big splits in pitching performance -- by home / road, day / night, days rest, and so on -- should always be treated first as real when they are significant enough statistically (after you correct for the number of splits you're looking at). However, are there not still other factors associated with the time of the games, daylight conditions, etc that could also play a part? You've done some good isolation...I'm just not totally convinced it's comprehensive. That said, I'm not alone in agreeing that sleep could be a huge factor here. It's not coincidental that so many pro sports franchises have sleep specialists now. I spent 24 years in a career that too-often left me sleep deprived; and in childish ways even encouraged it. It is only now, that I'm able to get good sleep consistently, that I can really feel the difference and realize I probably took at least 3-4 years off my lifespan.
|
|
|
Post by jmei on Jul 22, 2016 10:01:36 GMT -5
This is the exact situation where hypo testing p-values are worthless and misleading, and you know that. It's the kind of data dredging that gets you laughed out of conferences. When you have literally hundreds of variables available to compare two individual starts, you're going to find a lot of false correlations.
|
|
jimoh
Veteran
Posts: 3,966
|
Post by jimoh on Jul 22, 2016 11:51:34 GMT -5
However, are there not still other factors associated with the time of the games, daylight conditions, etc that could also play a part? You've done some good isolation...I'm just not totally convinced it's comprehensive. That said, I'm not alone in agreeing that sleep could be a huge factor here. It's not coincidental that so many pro sports franchises have sleep specialists now. I spent 24 years in a career that too-often left me sleep deprived; and in childish ways even encouraged it. It is only now, that I'm able to get good sleep consistently, that I can really feel the difference and realize I probably took at least 3-4 years off my lifespan. The idea that sleep could be a factor here is perfectly fine, but the claim that lack of sleep due to partying is undeniably the correct explanation for this person you have never spoken to and never examined is still shamefully, embarrassingly ridiculous. The chain of assumptions is preposterous.
|
|
|
Post by Don Caballero on Jul 22, 2016 13:57:35 GMT -5
Somehow, this link led me to DeepDream and now I'm going to spend the rest of my afternoon making psychedelic images out of pretty much everything on my computer. Thank you!
|
|
ianrs
Veteran
Posts: 2,414
|
Post by ianrs on Jul 22, 2016 14:13:01 GMT -5
You have two outliers, the day games on the road, with p=.001 for being random versus the other 17 starts. Eric, I don't even think what you are suggesting is that absurd at face value (that Rodriguez had two outlier starts). Its the presentation of your argument that seems...quite off. You could have simply used other statistical indicators like Cook's distance (or other indicators of statistical leverage/influence), comparing the starts to either the rest of his season or the rest of his career, then argued that those two starts met outlier criterion by definition. It would be a more descriptive statistical approach, rather than a confirmatory hypothesis testing approach (you even could have used a sensitivity analysis to see if the day starts disproportionately skewed p-values and/or the prediction of magnitude and direction of performance). Which would seem to make more sense when you're working directly with numbers rather than as the Red Sox team psychologist. Since outliers are by definition those data points with strong statistical influence, particularly on p-values, you could have argued that these starts should not be given as much weight going forward when projecting the remainder of Rodriguez' season. Just a thought...
|
|
|
Post by joshv02 on Jul 22, 2016 14:17:03 GMT -5
en.wikipedia.org/wiki/Fooled_by_RandomnessEDIT: Also, if lavarnwayguy had claimed the Rodriguez had pitched poorly because Latinos love to party in LA, he'd have been crucified. It shouldn't be any different just because there are cherry-picked numbers alongside improvable presumptions. I've read enough van to not want to go down that rabbit hole, but I should point out that Talib is a wonderful story teller, and very interesting, and even though some of his work is incomprehensible I actually really liked Fooled by Randomness, but in the end he is a complete jerk on Twitter and in all the interviews I see him. Beyond jerky, really. He loves to pick a weird fight. I've never met him, of course. (His views on Labenese cultural are really beyond interesting, too.)
|
|
|
Post by mandelbro on Jul 22, 2016 15:39:35 GMT -5
This is the exact situation where hypo testing p-values are worthless and misleading, and you know that. It's the kind of data dredging that gets you laughed out of conferences. When you have literally hundreds of variables available to compare two individual starts, you're going to find a lot of false correlations. And even if you accept the premise that struggling in the two road games is sufficient, there are a million reasons that one could struggle on the road. It could mess with his routine or his preparation in other ways. Do we know how to calculate Latino Nightlife Runs Above Average too?
|
|
|
Post by trotfan on Jul 22, 2016 17:30:47 GMT -5
Chris Sale is available ... Well that is if you want to gut the entire top half of our prospects
|
|
|
Post by blizzards39 on Aug 15, 2016 16:58:08 GMT -5
Maybe a little early to start thinking about, but if Sox are to go into the playoffs hoe does this rotation shake out??? This is all assuming the Wright is healthy and in form and obviously the stretch run may self determine how it shakes out but? Who is 1? Who is left out? I'd tend to lean 1-porcello (pitching like an ace. Would you rather him or Lester? Definatly like porcello a contract better. Still has to prove tho) 2-price (I still think he is doing much better than his outlying numbers. I would be opposed or scared with David in game 1) 3-Wright( I like Wright after price. Assuming health) 4-Pomeranz/Rodriquez (depending who is hotter. Maybe the edge to Erod. Maybe)
|
|
|
Post by ryan24 on Aug 15, 2016 18:20:15 GMT -5
Maybe a little early to start thinking about, but if Sox are to go into the playoffs hoe does this rotation shake out??? This is all assuming the Wright is healthy and in form and obviously the stretch run may self determine how it shakes out but? Who is 1? Who is left out? I'd tend to lean 1-porcello (pitching like an ace. Would you rather him or Lester? Definatly like porcello a contract better. Still has to prove tho) 2-price (I still think he is doing much better than his outlying numbers. I would be opposed or scared with David in game 1) 3-Wright( I like Wright after price. Assuming health) 4-Pomeranz/Rodriquez (depending who is hotter. Maybe the edge to Erod. Maybe) Definitely go with first 3. erod as 4 and drew as long relief. You going with 11 or 12 pitchers? Do you include clay?
|
|
|
Post by amfox1 on Aug 15, 2016 23:10:03 GMT -5
Last year, the schedule was 10/6 (WC game), 10/8, 10/9, 10/11, 10/12, 10/14 (ALDS), which means 4 starters.
The order would likely depend on whether the Red Sox can actually set a rotation, but I would be aiming for the following rotation:
Wild card: 1. Porcello (right now, he's the obvious choice for a road elimination game)
ALDS: 1. Price (road game) 2. Wright (road game, he's been significantly better on the road this year) 3. Pomerantz (home game) 4. Porcello (home game) 5. Price (road game)
Bullpen (8): Closer - Kimbrel 8th - Ziegler 7th - Barnes Mix-and-match - Abad, Ross, Rodriguez, Tazawa Long man - Buchholz
On the outside - Kelly, Hembree, Uehara (health)
|
|
|
Post by ryan24 on Aug 16, 2016 4:26:38 GMT -5
Last year, the schedule was 10/6 (WC game), 10/8, 10/9, 10/11, 10/12, 10/14 (ALDS), which means 4 starters. The order would likely depend on whether the Red Sox can actually set a rotation, but I would be aiming for the following rotation: Wild card: 1. Porcello (right now, he's the obvious choice for a road elimination game) ALDS: 1. Price (road game) 2. Wright (road game, he's been significantly better on the road this year) 3. Pomerantz (home game) 4. Porcello (home game) 5. Price (road game) Bullpen (8): Closer - Kimbrel 8th - Ziegler 7th - Barnes Mix-and-match - Abad, Ross, Rodriguez, Tazawa Long man - Buchholz On the outside - Kelly, Hembree, Uehara (health) Like the rotation setup. Drew, erod, and clay are a roll of the dice. One of them would pitch game 4. Whoever has pitched the best. But only keep 2 of the 3. I would look at hembree. maybe Kelly depending on what they have done in September.
|
|
|
Post by soxfan1615 on Aug 16, 2016 4:52:24 GMT -5
There's absolutely no way E-Rod slots ahead of Pomeranz in the playoff rotation. Pomeranz is in, E Rod is out
|
|
|
Post by ryan24 on Aug 16, 2016 8:24:01 GMT -5
There's absolutely no way E-Rod slots ahead of Pomeranz in the playoff rotation. Pomeranz is in, E Rod is out Yup I think you are right based on today. Close enough to say who Farrell thinks is pitching the best. The problem I see is how does erod fit coming out of the pen. The other 2 drew and clay have pitched out of the pen.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Aug 16, 2016 9:52:54 GMT -5
Last year, the schedule was 10/6 (WC game), 10/8, 10/9, 10/11, 10/12, 10/14 (ALDS), which means 4 starters. The order would likely depend on whether the Red Sox can actually set a rotation, but I would be aiming for the following rotation: Wild card: 1. Porcello (right now, he's the obvious choice for a road elimination game) ALDS: 1. Price (road game) 2. Wright (road game, he's been significantly better on the road this year) 3. Pomerantz (home game) 4. Porcello (home game) 5. Price (road game) Bullpen (8): Closer - Kimbrel 8th - Ziegler 7th - Barnes Mix-and-match - Abad, Ross, Rodriguez, Tazawa Long man - Buchholz On the outside - Kelly, Hembree, Uehara (health) If you had 8 quality relievers and a manager who could slice and dice them against opposing hitters, an 8-man pen might make sense. We don't, and we most certainly don't. I see Uehara as the 7th guy if he's healthy, and Hembree, Abad, and Buchholz competing for it if he's not. There's no way that the second best healthy guy of that quartet is more valuable to you in the post-season than Moncada as a pinch-runner (and one who may be far less of an offensive downgrade than the usual PR if he has to bat). E-Rod in the pen can be a weapon. And you especially don't need an 8-man pen when he can come into a tie game in extra innings and pitch for a while. Having a 5th starter who is of comparable quality or better than the guys in the bottom half of your pen is like having two or three extra guys out there.
|
|
|
Post by sox fan in nc on Aug 16, 2016 9:55:11 GMT -5
Pomz looked pretty good against the best team in the AL yesterday (at least record wise). Texas lost Choo for the year so that brings them down a peg. Not sure E-Rod gets the automatic bid over Pomz.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Aug 16, 2016 10:01:59 GMT -5
There's absolutely no way E-Rod slots ahead of Pomeranz in the playoff rotation. Pomeranz is in, E Rod is out Yup I think you are right based on today. Close enough to say who Farrell thinks is pitching the best. The problem I see is how does erod fit coming out of the pen. The other 2 drew and clay have pitched out of the pen. Pomz looked pretty good against the best team in the AL yesterday (at least record wise). Texas lost Choo for the year so that brings them down a peg. Not sure E-Rod gets the automatic bid over Pomz. If it were based strictly on performance, the most likely dilemma at season's end would be whether E-Rod or Price should be the 4th starter. Price's struggles have been so persistent that you have to wonder whether he changed his off-season regimen because of his free agency. I have a lot of confidence that he'll be himself next year. I think it would now be naive to think that the turnaround comes any sooner.
|
|
|
Post by soxfan1615 on Aug 16, 2016 17:04:20 GMT -5
Yup I think you are right based on today. Close enough to say who Farrell thinks is pitching the best. The problem I see is how does erod fit coming out of the pen. The other 2 drew and clay have pitched out of the pen. Pomz looked pretty good against the best team in the AL yesterday (at least record wise). Texas lost Choo for the year so that brings them down a peg. Not sure E-Rod gets the automatic bid over Pomz. If it were based strictly on performance, the most likely dilemma at season's end would be whether E-Rod or Price should be the 4th starter. Price's struggles have been so persistent that you have to wonder whether he changed his off-season regimen because of his free agency. I have a lot of confidence that he'll be himself next year. I think it would now be naive to think that the turnaround comes any sooner. But it shouldn't be based strictly on this season's performance. It should be based on expected performance going forward. Price still has excellent peripherals and projections. If we sort by projected ERA (Depth Charts) we get Price 3.40 Pomeranz 3.67 Porcello 4.00 Wright 4.16 (I think this is wrong, knuckleballers are hard to project) Rodriguez 4.28
|
|
|
Post by blizzards39 on Aug 16, 2016 17:11:15 GMT -5
David price ranks 7th amounts AL starting pitchers in fWAR. Not a cy young type year but still solid.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,923
|
Post by ericmvan on Aug 16, 2016 17:33:45 GMT -5
If it were based strictly on performance, the most likely dilemma at season's end would be whether E-Rod or Price should be the 4th starter. Price's struggles have been so persistent that you have to wonder whether he changed his off-season regimen because of his free agency. I have a lot of confidence that he'll be himself next year. I think it would now be naive to think that the turnaround comes any sooner. But it shouldn't be based strictly on this season's performance. It should be based on expected performance going forward. Price still has excellent peripherals and projections. If we sort by projected ERA (Depth Charts) we get Price 3.40 Pomeranz 3.67 Porcello 4.00 Wright 4.16 (I think this is wrong, knuckleballers are hard to project) Rodriguez 4.28 My point is that such ROS projections have been consistently wrong all year. And they are projections based on a standard weighting of this year's results to previous ones. Clearly, in reality, each season in a career has its own weight. Once Hanley got hurt last year, he fell short of his ROS projection the rest of the way. Any genuine breakout hitter will exceed his ROS. Mid-August seems to me to be about the right point in the season where you can say that if there has been no observable regression towards the ROS projection so far, we're seeing something anomalous and our ROS projections should be accordingly tempered. Porcello (includes his first 2/3 of last year) and E-Rod (includes his injured first half, and his pitch-tipping games and other possible outliers from last year) are also unreliable. As is Pomeranz, since the system doesn't know he added a cutter this year, which means that he's being lumped with all of the guys whose apparent breakouts were random. Hey, that's everybody!
|
|
|
Post by jmei on Aug 17, 2016 10:34:09 GMT -5
But it shouldn't be based strictly on this season's performance. It should be based on expected performance going forward. Price still has excellent peripherals and projections. If we sort by projected ERA (Depth Charts) we get Price 3.40 Pomeranz 3.67 Porcello 4.00 Wright 4.16 (I think this is wrong, knuckleballers are hard to project) Rodriguez 4.28 My point is that such ROS projections have been consistently wrong all year. And they are projections based on a standard weighting of this year's results to previous ones. Clearly, in reality, each season in a career has its own weight. Once Hanley got hurt last year, he fell short of his ROS projection the rest of the way. Any genuine breakout hitter will exceed his ROS. Mid-August seems to me to be about the right point in the season where you can say that if there has been no observable regression towards the ROS projection so far, we're seeing something anomalous and our ROS projections should be accordingly tempered. Porcello (includes his first 2/3 of last year) and E-Rod (includes his injured first half, and his pitch-tipping games and other possible outliers from last year) are also unreliable. As is Pomeranz, since the system doesn't know he added a cutter this year, which means that he's being lumped with all of the guys whose apparent breakouts were random. Hey, that's everybody! Actual non-cherry-picked empirical analysis of this question suggests that you're dead wrong. Here's MGL: More at the link (and here's the equivalent analysis for hitters). Yes, there are sometimes real breakouts (and the opposite of breakouts-- death spirals?) that the projections can't take into account. But the vast majority of the time? Trust the projections.
|
|
|
Post by bosox81 on Aug 17, 2016 11:51:19 GMT -5
I'm late to the post-season rotation discussion. Mainly, because I don't like to count my chickens before they're hatched, so to speak. But, I think ERod in, Pomeranz out. ERod's superior performance post all-star has been mentioned, but also some of you are forgetting the extreme jump in innings for Pomeranz if they make a deep run in the playoffs. That has to be factored too.
|
|
|