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.
9/21-9/24 Red Sox vs. Rays Series Thread
|
Post by sarasoxer on Sept 22, 2015 10:14:20 GMT -5
I now stand "acquainted". Thanks for the info Chris.
|
|
|
Post by Oregon Norm on Sept 22, 2015 10:33:56 GMT -5
and with that, Bayes rolls in his grave And without researching where he is buried, north or south of the equator we don't know if that's clock or counter clock wise. We know exactly where the Reverend is buried, at Bunhill Fields, in London:
|
|
|
Post by thursty on Sept 22, 2015 11:01:46 GMT -5
And without researching where he is buried, north or south of the equator we don't know if that's clock or counter clock wise. We know exactly where the Reverend is buried, at Bunhill Fields, in London: Now *that's* a posterior distribution
|
|
|
Post by ramireja on Sept 22, 2015 11:27:38 GMT -5
Is predicting 3 home runs in 140 PA going out on a limb? Is 3 hr in 140 PA evidence of real power? Not saying I don't think he has raw power, or that he will eventually develop into a power hitter of sorts......I'm just not exactly sure we should all give you a pat on the back quite yet. And BTW, it is 2.5 times as many HRs as expected (1.2) at his previous rate (including the 8/16 HR). The HR/PA split that I cite is statistically significant. Thats all fine and dandy, I'm basically saying these numbers are so ridiculously low that its pretty silly to bring stats into this (and even sillier to be claiming statistical significance). I think if you polled the site 140 PA ago and asked for estimates on Bogaerts HR over this period of time, the vast majority would have predicted 3 or more.
|
|
|
Post by Oregon Norm on Sept 22, 2015 11:56:31 GMT -5
My own take is that, in having to adjust to the superior breaking stuff of ML pitchers, he developed a different approach and a swing to go with it. I'm sure that was done with the understanding, encouragement, and help of the staff, with Chili Davis playing an especially large role, I'll bet.
That swing has allowed him to largely negate the effectiveness of those outside pitches. He's hurt a lot of teams this year by going the other way on a very regular basis. They've started to probe him with other pitches in different locations, and he's now re-incorporating his power swing into the mix. What that means, long term, is that he has to have a good enough knowledge of how he's being pitched, along with an understanding of who's doing the pitching and what they bring, to know when to use one or the other. That's not easy.
But if there's one thing he's shown it's how intelligent he is, how he adapts and how willing he is to go at it systematically with that long term vision in mind. His approach to fielding has been exactly the same, systematic and thorough.
The analysis that showed him vulnerable because of bat speed is static, and this guy is dynamic by any measure you'd like to use. He builds on his skillset in a very advanced way. In that sense, he's really a perfect target for using Bayesian analysis, given the way he incorporates new information into the mix.
He's going to be a great ballplayer, hell he's already there.
|
|
|
Post by jchang on Sept 22, 2015 12:01:38 GMT -5
I may not be able to watch tonights game. If someone could be so kind as to report on how Owens fares against L vs R, I would greatly appreciate it. so far, his (MLB) line is:
WHIP AVG OBP SLG OPS ERA
- 1.35 .250/.326/.411/.737 - 4.33
L 1.96 .303/.410/.515/.925 - 8.22
R 1.22 .237/.304/.385/.689 - 3.50 I am not sure how ERA is calculated in splits? what if a R got on, then scored with hit by a L?
|
|
ianrs
Veteran
Posts: 2,418
|
Post by ianrs on Sept 22, 2015 12:08:52 GMT -5
And BTW, it is 2.5 times as many HRs as expected (1.2) at his previous rate (including the 8/16 HR). The HR/PA split that I cite is statistically significant. Thats all fine and dandy, I'm basically saying these numbers are so ridiculously low that its pretty silly to bring stats into this (and even sillier to be claiming statistical significance). I think if you polled the site 140 PA ago and asked for estimates on Bogaerts HR over this period of time, the vast majority would have predicted 3 or more. I absolutely agree with ramireja here. Not choosing an alpha or power level, then conducting a test of statistical significance post-hoc is almost the same as saying, "Look, Xander is hitting more homeruns than he did before!" It doesn't really tell us anything about next season or about his next X number of PA's.
|
|
|
Post by Chris Hatfield on Sept 22, 2015 12:58:52 GMT -5
I may not be able to watch tonights game. If someone could be so kind as to report on how Owens fares against L vs R, I would greatly appreciate it. so far, his (MLB) line is: WHIP AVG OBP SLG OPS ERA
- 1.35 .250/.326/.411/.737 - 4.33
L 1.96 .303/.410/.515/.925 - 8.22
R 1.22 .237/.304/.385/.689 - 3.50 I am not sure how ERA is calculated in splits? what if a R got on, then scored with hit by a L? I believe it's the results of a plate appearance by a batter of that handedness. So when a RHB is up, anything that happens during that PA goes in the RHB bucket. For what it's worth, the LHB sample size is 39 PA at this point. Say you make the wind blow in harder on one of the two LHB home runs he's given up, turning it into an out, the line drops from .303/.410/.505 to .272/.384/.393. Still not great, but markedly different looking. There's still a lot of noise in that sample, at this point, before we should worry too much about it.
|
|
|
Post by jmei on Sept 22, 2015 14:12:40 GMT -5
I may not be able to watch tonights game. If someone could be so kind as to report on how Owens fares against L vs R, I would greatly appreciate it. so far, his (MLB) line is: WHIP AVG OBP SLG OPS ERA
- 1.35 .250/.326/.411/.737 - 4.33
L 1.96 .303/.410/.515/.925 - 8.22
R 1.22 .237/.304/.385/.689 - 3.50 I am not sure how ERA is calculated in splits? what if a R got on, then scored with hit by a L? I believe it's the results of a plate appearance by a batter of that handedness. So when a RHB is up, anything that happens during that PA goes in the RHB bucket. For what it's worth, the LHB sample size is 39 PA at this point. Say you make the wind blow in harder on one of the two LHB home runs he's given up, turning it into an out, the line drops from .303/.410/.505 to .272/.384/.393. Still not great, but markedly different looking. There's still a lot of noise in that sample, at this point, before we should worry too much about it. It's backed up by 500+ PAs at the minor league level, though. jchang and I have posted about Owens' reverse splits before, and it increasingly looks like a thing.
|
|
|
Post by benogliviesbrother on Sept 22, 2015 15:54:41 GMT -5
It's an excellent resource. Thanks for posting the link. Is there a sticky somewhere on this board where various useful sites are linked? (methinks that would make for an amazing thread)
|
|
|
Post by sarasoxer on Sept 22, 2015 16:09:51 GMT -5
It's an excellent resource. Thanks for posting the link. Is there a sticky somewhere on this board where various useful sites are linked? (methinks that would make for an amazing thread) I was thinking the same thing...
|
|
|
Post by brianthetaoist on Sept 22, 2015 16:27:26 GMT -5
Whoa, posting this from James Dunne's tweet on the side of the page here:
That's a lot of baseball in between 1909 and today.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Sept 22, 2015 16:57:14 GMT -5
Thats all fine and dandy, I'm basically saying these numbers are so ridiculously low that its pretty silly to bring stats into this (2) and even sillier to be claiming statistical significance). I think if you polled the site 140 PA ago (1) and asked for estimates on Bogaerts HR over this period of time, the vast majority would have predicted 3 or more.I absolutely agree with ramireja here. Not choosing an alpha or power level, then conducting a test of statistical significance post-hoc is almost the same as saying, "Look, Xander is hitting more homeruns than he did before!" It doesn't really tell us anything about next season or about his next X number of PA's. (1) That was basically my point. Because we've watched the guy hit, we had an expectation that he would begin to hit more home runs, starting right then.
You objected to my posting the link to my scouting take on that expectation, thinking incorrectly that there was some kind of statistical rationale behind it, when there wasn't. (2) No, it's not silly at all. It's a necessary confirmation that the expected change is meaningful. We have reached the point where the evidence can now be said to be meaningfully consistent with our expectations. If he had hit one more home run after one third as much time had elapsed, we would have said "that means nothing yet." Whenever someone's performance has improved or declined, it is always helpful to know the probability of this happening at random. There's a huge epistemic difference* between saying ""Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 40%!" and "Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 3%!" And, strictly analytically, that means nothing more than what it means. But what it means is certainly useful. How can it not be? (One ubiquitous error of the Statistically Correct is thinking that it's not worth knowing if it only means what it means analytically.) Now, if you want to be strictly analytical, you can construct confidence intervals for his performance going forward, from the same data. I've done that in the past. But it's usually not worth the effort*. You can just take your scouting impression, the observed change in results and, always, and crucially, your knowledge of the likelihood of those results being merely random, and get a good sense of what is likely to happen in the future. *Sorry, reading too much philosophy! **The exception is when you get ridiculous results in a very small sample, which generally cause our mental intuition circuits to get cramps. I still think he could have hit ... but not playing him was relatively low on the list of Jimy Williams' mistakes.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,931
|
Post by ericmvan on Sept 22, 2015 17:17:14 GMT -5
Whoa, posting this from James Dunne's tweet on the side of the page here: That's a lot of baseball in between 1909 and today. Furthermore, based on WPA (from FG) relative to (bRef's Batting RAA + BaseRunning RAA + GDP RAA) / 9.0) ... Mookie has 0.5 wins of clutch hitting and Xander has 1.4. (The fact that Mookie has a -0.4 win "Clutch" metric at FG tells you how completely wrong that stat is.) That may have little predictive value for itself, but you can argue that it reflects a maturity and ability to handle pressure that has predictive value for their improvement in context-neutral measures. Or maybe not, and it's just good to know. Edit: oh yeah, b-Ref does not include advancing on hits and outs, and Xander and Mookie lead the team with 0.4 wins each. That should count as much as their regular WAR.
|
|
|
Post by jmei on Sept 22, 2015 17:37:09 GMT -5
Whenever someone's performance has improved or declined, it is always helpful to know the probability of this happening at random. There's a huge epistemic difference* between saying ""Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 40%!" and "Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 3%!" And, strictly analytically, that means nothing more than what it means. But what it means is certainly useful. How can it not be? (One ubiquitous error of the Statistically Correct is thinking that it's not worth knowing if it only means what it means analytically.) Now, if you want to be strictly analytical, you can construct confidence intervals for his performance going forward, from the same data. I've done that in the past. But it's usually not worth the effort*. You can just take your scouting impression, the observed change in results and, always, and crucially, your knowledge of the likelihood of those results being merely random, and get a good sense of what is likely to happen in the future. Its usefulness is sometimes (often) dwarfed by its potential to mislead. It's like how managers read too much into small sample pitcher-versus-batter matchup stats. Are those stats completely useless? No, but they're often misinterpreted, and we have more useful analytical tools.
|
|
|
Post by Nick Rabasco2 on Sept 22, 2015 18:37:07 GMT -5
Mookie's BABIP on May 17 (168 PA in) was .226. It's now at .309 for the season including that double.
|
|
|
Post by ramireja on Sept 22, 2015 19:00:29 GMT -5
I absolutely agree with ramireja here. Not choosing an alpha or power level, then conducting a test of statistical significance post-hoc is almost the same as saying, "Look, Xander is hitting more homeruns than he did before!" It doesn't really tell us anything about next season or about his next X number of PA's. (1) That was basically my point. Because we've watched the guy hit, we had an expectation that he would begin to hit more home runs, starting right then.
You objected to my posting the link to my scouting take on that expectation, thinking incorrectly that there was some kind of statistical rationale behind it, when there wasn't. (2) No, it's not silly at all. It's a necessary confirmation that the expected change is meaningful. We have reached the point where the evidence can now be said to be meaningfully consistent with our expectations. If he had hit one more home run after one third as much time had elapsed, we would have said "that means nothing yet." Whenever someone's performance has improved or declined, it is always helpful to know the probability of this happening at random. There's a huge epistemic difference* between saying ""Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 40%!" and "Look, Xander is hitting more homeruns than he did before, and the odds of this being random given his overall home run rate are 3%!" And, strictly analytically, that means nothing more than what it means. But what it means is certainly useful. How can it not be? (One ubiquitous error of the Statistically Correct is thinking that it's not worth knowing if it only means what it means analytically.) Now, if you want to be strictly analytical, you can construct confidence intervals for his performance going forward, from the same data. I've done that in the past. But it's usually not worth the effort*. You can just take your scouting impression, the observed change in results and, always, and crucially, your knowledge of the likelihood of those results being merely random, and get a good sense of what is likely to happen in the future. *Sorry, reading too much philosophy! **The exception is when you get ridiculous results in a very small sample, which generally cause our mental intuition circuits to get cramps. I still think he could have hit ... but not playing him was relatively low on the list of Jimy Williams' mistakes. I don't disagree with your general points here. I think my original annoyance has more to do with this particular case. He has hit all of 3 HR over 140 PA (so on pace for 12 hr over 560 PA), and suddenly we're to believe he's morphed into the power hitter that you predicted he would become? I mean the only reason he has hit significantly more home runs than he was expected to over this sample period, is because his baseline rate was astronomically low. But I don't think anyone else believed that rate was a true indicator of Bogaerts' power potential. Furthermore, when we're dealing with numbers this low (1.2 vs. 3 expected HR over 140 PA), the difference between 1 or 2 HR becomes huge in your analysis. A difference between 1 and 2 HR could easily be attributed to wind, park effects, etc. In fact, Bogey may have lost a home run last night due to the wind (although had he hit one, that would have only helped your analysis). Bottom line, these numbers are so low that I can't put much stock into your analysis....I wasn't trying to argue that you should give up on your analyses altogether. This one in particular is just a bit insane.
|
|
|
Post by beantown on Sept 22, 2015 19:17:37 GMT -5
Boy, Owens looks nasty tonight. FB ranging only 87-91 but the changeup is on another level. That velo still has me skeptical but hitters seem to struggle squaring him up, don't they? Tons of lazy fly balls..
|
|
gerry
Veteran
Enter your message here...
Posts: 1,667
|
Post by gerry on Sept 22, 2015 19:34:39 GMT -5
Boy, Owens looks nasty tonight. FB ranging only 87-91 but the changeup is on another level. That velo still has me skeptical but hitters seem to struggle squaring him up, don't they? Tons of lazy fly balls.. And as he gains experience, confidence and makes peace with his still filling out body, he should get better and better. He is already doing better than alot of long, gangly all stars did in their debuts. Nice!
|
|
|
Post by Legion of Bloom on Sept 22, 2015 19:36:40 GMT -5
And that is why Betts shouldn't play RF. He doesn't have the arm for it.
|
|
|
Post by benogliviesbrother on Sept 22, 2015 19:59:43 GMT -5
That was a big time escape.
|
|
gerry
Veteran
Enter your message here...
Posts: 1,667
|
Post by gerry on Sept 22, 2015 20:15:08 GMT -5
That was a big time escape. It isn't often that a fan gets a souvenier thrown in from RF. Mighty Mookie. Who knew. Am liking Shaw more every game. Savvy ballplayer with solid bat and glove.
|
|
|
Post by fenwaythehardway on Sept 22, 2015 20:17:13 GMT -5
Hey, remember when Owens had ace upside because of his deception and strikeout rate against double-A hitters? How all those scouts who called him a future #4 were just slaves to the radar gun? Fun times.
Don't get me wrong, I like the kid just fine. He'll be a valuable pitcher and that changeup makes him fun to watch. But he's a mid-to-back of the rotation guy.
|
|
|
Post by James Dunne on Sept 22, 2015 20:43:35 GMT -5
Whoa, posting this from James Dunne's tweet on the side of the page here: That's a lot of baseball in between 1909 and today. Yeah, I posted that in Bogaerts thread today. Basically either Bogaerts (or Betts, depending on what stat you like) is the best second-best 22-year-old a team has ever had.
|
|
|
Post by jmei on Sept 22, 2015 20:52:04 GMT -5
Hey, remember when Owens had ace upside because of his deception and strikeout rate against double-A hitters? How all those scouts who called him a future #4 were just slaves to the radar gun? Fun times. Don't get me wrong, I like the kid just fine. He'll be a valuable pitcher and that changeup makes him fun to watch. But he's a mid-to-back of the rotation guy. 51.0 major league innings seems a little too soon to be busting out the "I-told-you-so"s.
|
|
|