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 Spring Training News/Discussion
|
Post by jimed14 on Mar 23, 2016 12:30:51 GMT -5
What is seen, cannot be unseen.
|
|
redsox04071318champs
Veteran
Always hoping to make my handle even longer...
Posts: 16,483
Member is Online
|
Post by redsox04071318champs on Mar 23, 2016 12:48:42 GMT -5
Looking over the roster it would seem to me it's pretty well set now with the only mystery being will the Sox go with a 12th pitcher, Elias or with a fifth bench player, Murphy?
What does everything think? Does Murphy have a role on this team? Does anybody really think the Sox will go with a Murphy/Young platoon in LF rather than handing LF over to Castillo?
I think that Murphy will begin the season with the Sox and they will go with 11 pitchers (including Matt Barnes in the bullpen in place of Carson Smith and including Wright in the rotation.)
|
|
|
Post by telson13 on Mar 23, 2016 13:32:30 GMT -5
Looking over the roster it would seem to me it's pretty well set now with the only mystery being will the Sox go with a 12th pitcher, Elias or with a fifth bench player, Murphy? What does everything think? Does Murphy have a role on this team? Does anybody really think the Sox will go with a Murphy/Young platoon in LF rather than handing LF over to Castillo? I think that Murphy will begin the season with the Sox and they will go with 11 pitchers (including Matt Barnes in the bullpen in place of Carson Smith and including Wright in the rotation.) Yeah, I'm not sure they really need a 12th pitcher, and I can see them holding on to Murphy just to have a little insurance in case JBJ or Castillo struggle. Plus, they'll probably want to keep Elias stretched out for now.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 9,015
|
Post by ericmvan on Mar 23, 2016 13:42:17 GMT -5
Looking over the roster it would seem to me it's pretty well set now with the only mystery being will the Sox go with a 12th pitcher, Elias or with a fifth bench player, Murphy? What does everything think? Does Murphy have a role on this team? Does anybody really think the Sox will go with a Murphy/Young platoon in LF rather than handing LF over to Castillo? I think that Murphy will begin the season with the Sox and they will go with 11 pitchers (including Matt Barnes in the bullpen in place of Carson Smith and including Wright in the rotation.) Murphy was signed in large part as catastrophe protection. He has no positive value as any kind of everyday player, because he's reached the age where he can't play the OF adequately. He would be the team's 7th OF if he were on the roster, and would get in a game only if there were two pairs of OF collisions where both guy had to leave the game. Rusney Castillo's defense, on the other hand, has been absolutely elite in his brief MLB PT, which indicates that it's at the very least well above average. He can be just as disappointing at the plate as we fear, and still be a win or two better versus RHP than Murphy. As a matter of fact, it's unclear that Young will be better versus LHP than he will be -- suggesting that we may seen some young at DH, especially early in the season if Papi gets off to one of his frequent slow starts. There is one thing Murphy can do, though, and that's pinch-hit. He'd be a nice bench piece if you had room on the roster for him. With him and Young on the bench, you can pinch-hit for Castillo with one of them, and if the other team brings in a same-handed pitcher, you come back with the other guy. I'm in favor of going with a 5th bench guy for the Indians series, and it would be either Murphy in that role, or Marrero as a defensive replacement for Sandoval. I have no strong feelings either way ... but either guy is off the roster after they leave town for Toronto (and maybe even for the 3rd game of the series).
|
|
|
Post by sarasoxer on Mar 23, 2016 18:03:04 GMT -5
What is seen, cannot be unseen. Master Po...Kung Fu
|
|
|
Post by soxfan06 on Mar 23, 2016 19:21:45 GMT -5
So Cleveland is shopping for an outfielder? How can we make a Rusney for Salazar happen?
I know, I'm just dreaming. That would be awesome though.
|
|
|
Post by pokeyreesespieces on Mar 24, 2016 1:01:45 GMT -5
So Cleveland is shopping for an outfielder? How can we make a Rusney for Salazar happen? I know, I'm just dreaming. That would be awesome though. I would throw in some pretty sizeable prospects to make that happen. Rusney, Owens, Travis... ship em out!
|
|
|
Post by dirtywater43 on Mar 24, 2016 3:21:06 GMT -5
So Cleveland is shopping for an outfielder? How can we make a Rusney for Salazar happen? I know, I'm just dreaming. That would be awesome though. I would throw in some pretty sizeable prospects to make that happen. Rusney, Owens, Travis... ship em out! That'll get you a Trevor Bauer probably with that package, maybe another throw in. I hope you weren't expecting Carrasco with that set of players. Carrasco is starting with a JBJ package followed with probably a lot more. Ohh and Salazar isn't being traded for anything less than a Mookie Betts, which is why we can stop dreaming about Salazar right now.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 9,015
|
Post by ericmvan on Mar 24, 2016 9:23:43 GMT -5
Performance (R/150) and Rank of various Sox OFers in the three things that DRS measures, among 150 players with 550 or more innings, 2014-2015. Performance in RF and LF has been translated to CF based on the performance of everyone who played an inning from 2004-2015 (see the end for the translation table).
The figures are not precisely accurate because DRS reports in whole numbers only. But, close enough.
Plays Made 24.9 Holt (2) 9.6 Bradley (15) 5.1 Betts (24) 4.0 Castillo (27) -5.6 Craig (58) -9.5 Young (77) -18.7 Murphy (126) -35.7 Ramirez (149)
Arm 14.7 Castillo (1) 4.8 Bradley (15) 4.2 Betts (19) 1.3 Craig (49) -2.9 Young (106) -7.0 Murphy (142) -10.9 Ramirez (149) -11.0 Holt (150)
Good Fielding Plays 3.9 Holt (11) 3.5 Castillo (14) 2.5 Betts (23) 1.6 Bradley (36) -1.3 Murphy (97) -2.0 Craig (114) -3.2 Young (135) -8.0 Ramirez (150)
Guys who beat JBJ in all three measures: nobody. Guys who beat Castillo in all three measures: nobody (obviously). Guys who beat Betts in all three measures: Kiermaier.
That's 4 of the 8 guys who were top 40 in all three measures. The others: Leonys Martin, Kevin Pillar, Billy Hamilton, and Daniel Nava (!).
Other guy who Ramirez was able to beat in one measure: Michael Morse, -37.4 PM.
Note that Castillo's crazy Arm figure will regress to the mean a ton once guys stop running on him. But his performance at all three positions translated to +20 in LF even with 0.0 Arm.
Here are the runs you subtract from a RF or LF to get a figure for CF:
Metric RF LF PM 8.0 9.6 Arm 0.2 1.9 GFP 0.2 0.8 For total DRS, it's 8.4 and 12.4 (instead of the 12.3 total of the above, a rounding fix).
You'll note that when CF play RF, their arm plays a tiny bit better. Whereas, as expected, the move from RF to LF is slightly more costly to arm than to range (PM), 1.7 to 1.6.
(GFP for an individual player tends to correlate more strongly with Arm than to PM, or PM does to Arm. I think that hitting or missing the cutoff man is a big part of it.)
This was based on everyone who played two positions in the same season from 2004 (when they introduced GFP) to 2015. You take the lesser of the innings played and prorate the greater down to that. For instance, if a guy played 500 innings in CF and 100 innings in RF, his CF play would go into the database as 100 innings with 100/500 * each of his metrics as his performance. You then just take the totals for each pair of positions, over the 12 years. It's a minimum of 70,000 effective innings for each pair of positions (100,000+ for RF / LF).
(The actual change in PM for guys who played both CF and LF in a season was 9.0, while if you add the CF-RF and RF-LF changes together, you get 9.9. That's because of selection bias in who gets to play what position. The 9.6 figure just combines the two sets together, weighted.)
The last time I did this, about 10 years ago using UZR, CF to LF was c. 12 runs, but CF to RF was about 6. (The numbers were approximate because I used a less good methodology.) So, since then, there has been little or no change in CF versus LF, while RF performance has shifted about 2 runs per 150 in the direction of LF.
|
|
|
Post by grandsalami on Mar 24, 2016 10:54:36 GMT -5
Sean McAdam @sean_McAdam 3m3 minutes ago With Henry Owens optioned, fifth starter down to Steven Wright, Roenis Elias - bit.ly/1MEucXD
|
|
|
Post by sox fan in nc on Mar 24, 2016 12:38:05 GMT -5
I would throw in some pretty sizeable prospects to make that happen. Rusney, Owens, Travis... ship em out! That'll get you a Trevor Bauer probably with that package, maybe another throw in. I hope you weren't expecting Carrasco with that set of players. Carrasco is starting with a JBJ package followed with probably a lot more. Ohh and Salazar isn't being traded for anything less than a Mookie Betts, which is why we can stop dreaming about Salazar right now. OK.....JBJ + Owens + Travis for Carrasco....I think I'd have to pull that trigger. OF will get crowded next year when AB/Moncada being close to ready....Carrasco is signed for another 5 years on Miley dollars (until age 34) last 2 of those years team options @ 9 mil. I'd go with Mookie CF, Young LF, Castillo RF...Murphy & Holt/Shaw on the bench. Decent 4th OF'ers are not the hardest thing to find if injuries occur.
|
|
|
Post by jimed14 on Mar 24, 2016 13:08:28 GMT -5
They just are not going to create a huge hole in the outfield and decimate depth to upgrade a starting pitcher. Take it to the trade forum.
|
|
|
Post by telson13 on Mar 24, 2016 15:32:08 GMT -5
Performance (R/150) and Rank of various Sox OFers in the three things that DRS measures, among 150 players with 550 or more innings, 2014-2015. Performance in RF and LF has been translated to CF based on the performance of everyone who played an inning from 2004-2015 (see the end for the translation table). The figures are not precisely accurate because DRS reports in whole numbers only. But, close enough. Plays Made
24.9 Holt (2) 9.6 Bradley (15) 5.1 Betts (24) 4.0 Castillo (27) -5.6 Craig (58) -9.5 Young (77) -18.7 Murphy (126) -35.7 Ramirez (149) Arm
14.7 Castillo (1) 4.8 Bradley (15) 4.2 Betts (19) 1.3 Craig (49) -2.9 Young (106) -7.0 Murphy (142) -10.9 Ramirez (149) -11.0 Holt (150)
Good Fielding Plays
3.9 Holt (11) 3.5 Castillo (14) 2.5 Betts (23) 1.6 Bradley (36) -1.3 Murphy (97) -2.0 Craig (114) -3.2 Young (135) -8.0 Ramirez (150) Guys who beat JBJ in all three measures: nobody. Guys who beat Castillo in all three measures: nobody (obviously). Guys who beat Betts in all three measures: Kiermaier. That's 4 of the 8 guys who were top 40 in all three measures. The others: Leonys Martin, Kevin Pillar, Billy Hamilton, and Daniel Nava (!). Other guy who Ramirez was able to beat in one measure: Michael Morse, -37.4 PM. Note that Castillo's crazy Arm figure will regress to the mean a ton once guys stop running on him. But his performance at all three positions translated to +20 in LF even with 0.0 Arm. Here are the runs you subtract from a RF or LF to get a figure for CF: Metric RF LF PM 8.0 9.6 Arm 0.2 1.9 GFP 0.2 0.8 For total DRS, it's 8.4 and 12.4 (instead of the 12.3 total of the above, a rounding fix). You'll note that when CF play RF, their arm plays a tiny bit better. Whereas, as expected, the move from RF to LF is slightly more costly to arm than to range (PM), 1.7 to 1.6. (GFP for an individual player tends to correlate more strongly with Arm than to PM, or PM does to Arm. I think that hitting or missing the cutoff man is a big part of it.) This was based on everyone who played two positions in the same season from 2004 (when they introduced GFP) to 2015. You take the lesser of the innings played and prorate the greater down to that. For instance, if a guy played 500 innings in CF and 100 innings in RF, his CF play would go into the database as 100 innings with 100/500 * each of his metrics as his performance. You then just take the totals for each pair of positions, over the 12 years. It's a minimum of 70,000 effective innings for each pair of positions (100,000+ for RF / LF). (The actual change in PM for guys who played both CF and LF in a season was 9.0, while if you add the CF-RF and RF-LF changes together, you get 9.9. That's because of selection bias in who gets to play what position. The 9.6 figure just combines the two sets together, weighted.) The last time I did this, about 10 years ago using UZR, CF to LF was c. 12 runs, but CF to RF was about 6. (The numbers were approximate because I used a less good methodology.) So, since then, there has been little or no change in CF versus LF, while RF performance has shifted about 2 runs per 150 in the direction of LF. So, basically, Castillo is a 2 WAR (league average) player even if he hits at replacement level and you don't account for his arm, and 3 WAR player with fringy production and the arm.
|
|
|
Post by jimed14 on Mar 24, 2016 16:01:19 GMT -5
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 9,015
|
Post by ericmvan on Mar 24, 2016 16:11:14 GMT -5
Note that Castillo's crazy Arm figure will regress to the mean a ton once guys stop running on him. But his performance at all three positions translated to +20 in LF even with 0.0 Arm. So, basically, Castillo is a 2 WAR (league average) player even if he hits at replacement level and you don't account for his arm, and 3 WAR player with fringy production and the arm. Not quite, because there's some regression to the mean that'll happen on the range as well. I have almost all of the data I need to derive a kick-ass DRS projection system for OFers, including the crucial adjustment factors for the three factors, which allows me to treat the three positions as one. I think I will try to do that by the weekend. [For geeks only!] Doing it right is a lot of work. But I recently developed a method for regressing small samples just for their smallness. The standard Bayesian way, adding a fixed number of innings or PAs at average level, does not work. Your goal is to eliminate the inverse correlation between the sample size and the effect size, for the positive outcomes (there are fewer extreme negative outcomes in SSS because of selection bias, so they will be reduced even further). So, the steps are: 1) Combine everyone each year into one OFer with the three metrics. Do all three metrics separately. 2) Grab the ages for everyone in the report, plus the names of players who played any OF in 2001-2003. 3) Calculate the career totals for every player going into each season, for the players who began their OF careers in 2004 and later. I can do that with a bunch of cut-and-pastes from a pivot table, but maybe there's a better way! 4) Regress everything to the mean for SSS, as above. 5) Look at the year-to-year and year-to-previous-career correlations for each metric, as a function of sample sizes. Use that to try to determine the minimum number of innings for the regressions. 6) Find the best set of regression formulas for predicting Year from Year-1. Do not assume that it's a straight linear function; I've previously found that negative values regress more than positive. Throwing in the square of the metric as a possible factor should be able to account for most of the non-linearity but I'll still graph the residuals versus the underlying numbers to see if any patterns emerge. 7) Derive regression formulas for the cases where more than one year of data is available (most of them, that is). One of the things I'll throw into the mix is the career to that point. What I've found previously is that data more than 3 years old is meaningless, but I'd like to check that. Once you have these regression formulas, you can look into questions about learning curves and differences between positions.
|
|
|
Post by thelavarnwayguy on Mar 24, 2016 17:28:30 GMT -5
With 3 guys who are plus defensive Center fielders, and guys like Benintendi and Moncada for long term depth, I think we should absolutely maximize our utility for gain by trading a centerfielder not named Betts. Preferably Castillo or Bradley. That would seem to be a top priority to me right now. We are probably not going to get optimum value out of three bonafide plus defenders in CF unless one of them is traded.
|
|
|
Post by Oregon Norm on Mar 24, 2016 18:13:49 GMT -5
With 3 guys who are plus defensive Center fielders, and guys like Benintendi and Moncada for long term depth, I think we should absolutely maximize our utility for gain by trading a centerfielder not named Betts. Preferably Castillo or Bradley. That would seem to be a top priority to me right now. We are probably not going to get optimum value out of three bonafide plus defenders in CF unless one of them is traded. I do believe there's a fairly significant spread in perceived value in any trade market for those two guys. Castillo is almost the obverse of Bradley, at least with the bat. On one side, a right-handed bat who's not had much time to establish himself, who seems to be largely averse to walks, and who has a fairly pronounced platoon split. On the other a left-handed bat who finally had a small breakthrough, one of who's strengths is OBP, someone who flashed the kind of power that Castillo has yet to display, but with an interesting reverse split (so far). The glove, the on-base stuff, and those extra-base hits, probably put Bradley well ahead of Castillo in value. I'm not sure that the FO is going to be trading either, however. Ever since Dombrowski arrived, Bradley's been a player, a good one. And as Eric has pointed out, Castillo has some serious defensive chops. I thought he was exceptional in left field. I don't know that he'll be sent off. You did mention long-term. Right now, from what we're hearing, the team is in it for the short-term, as in winning this year. Maybe someone gets traded, but I'd put it at less than 50-50.
|
|
|
Post by telson13 on Mar 24, 2016 19:41:50 GMT -5
So, basically, Castillo is a 2 WAR (league average) player even if he hits at replacement level and you don't account for his arm, and 3 WAR player with fringy production and the arm. Not quite, because there's some regression to the mean that'll happen on the range as well. I have almost all of the data I need to derive a kick-ass DRS projection system for OFers, including the crucial adjustment factors for the three factors, which allows me to treat the three positions as one. I think I will try to do that by the weekend. [For geeks only!] Doing it right is a lot of work. But I recently developed a method for regressing small samples just for their smallness. The standard Bayesian way, adding a fixed number of innings or PAs at average level, does not work. Your goal is to eliminate the inverse correlation between the sample size and the effect size, for the positive outcomes (there are fewer extreme negative outcomes in SSS because of selection bias, so they will be reduced even further). So, the steps are: 1) Combine everyone each year into one OFer with the three metrics. Do all three metrics separately. 2) Grab the ages for everyone in the report, plus the names of players who played any OF in 2001-2003. 3) Calculate the career totals for every player going into each season, for the players who began their OF careers in 2004 and later. I can do that with a bunch of cut-and-pastes from a pivot table, but maybe there's a better way! 4) Regress everything to the mean for SSS, as above. 5) Look at the year-to-year and year-to-previous-career correlations for each metric, as a function of sample sizes. Use that to try to determine the minimum number of innings for the regressions. 6) Find the best set of regression formulas for predicting Year from Year-1. Do not assume that it's a straight linear function; I've previously found that negative values regress more than positive. Throwing in the square of the metric as a possible factor should be able to account for most of the non-linearity but I'll still graph the residuals versus the underlying numbers to see if any patterns emerge. 7) Derive regression formulas for the cases where more than one year of data is available (most of them, that is). One of the things I'll throw into the mix is the career to that point. What I've found previously is that data more than 3 years old is meaningless, but I'd like to check that. Once you have these regression formulas, you can look into questions about learning curves and differences between positions. That went over my head like an SR-71 Blackbird. I still appreciate the explanation though. I know enough statistical analysis to not embarrass myself I this board. This is two standard deviations above my mean. FWIW, it sounds like you may be able to extrapolate some interesting conclusions here. Curious to see how it all jibes with scouting.
|
|
|
Post by telson13 on Mar 24, 2016 19:48:04 GMT -5
With 3 guys who are plus defensive Center fielders, and guys like Benintendi and Moncada for long term depth, I think we should absolutely maximize our utility for gain by trading a centerfielder not named Betts. Preferably Castillo or Bradley. That would seem to be a top priority to me right now. We are probably not going to get optimum value out of three bonafide plus defenders in CF unless one of them is traded. I do believe there's a fairly significant spread in perceived value in any trade market for those two guys. Castillo is almost the obverse of Bradley, at least with the bat. On one side, a right-handed bat who's not had much time to establish himself, who seems to be largely averse to walks, and who has a fairly pronounced platoon split. On the other a left-handed bat who finally had a small breakthrough, one of who's strengths is OBP, someone who flashed the kind of power that Castillo has yet to display, but with an interesting reverse split (so far). The glove, the on-base stuff, and those extra-base hits, probably put Bradley well ahead of Castillo in value. I'm not sure that the FO is going to be trading either, however. Ever since Dombrowski arrived, Bradley's been a player, a good one. And as Eric has pointed out, Castillo has some serious defensive chops. I thought he was exceptional in left field. I don't know that he'll be sent off. You did mention long-term. Right now, from what we're hearing, the team is in it for the short-term, as in winning this year. Maybe someone gets traded, but I'd put it at less than 50-50. Agreed. Trading from depth without Benintendi or Moncada ready to step in and contribute is potentially disastrous. If Shaw can play some left, such that he, Holt, and Young can hold it down, maybe Castillo is tradeable. But he's probably not bringing adequate return...his skill set just isn't valued enough. JBJ is still in a career upswing, and although he has more value, I still doubt they get reasonable value in return. Better to wait until they a) get an offer they can't refuse, or b) get a legitimate contribution from their minor league stable.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 9,015
|
Post by ericmvan on Mar 24, 2016 21:40:34 GMT -5
With 3 guys who are plus defensive Center fielders, and guys like Benintendi and Moncada for long term depth, I think we should absolutely maximize our utility for gain by trading a centerfielder not named Betts. Preferably Castillo or Bradley. That would seem to be a top priority to me right now. We are probably not going to get optimum value out of three bonafide plus defenders in CF unless one of them is traded. The hope is that Castillo has a good enough year to establish some trade value, while Benintnedi has a good enough year to project as his replacement beginning in May of next year. Then you have a very nice trade chip for next winter. If Castillo isn't good enough for contenders to pursue as a CF solution, then it's likely that he'll still be a very good or great 4th OFer, a guy who can do some damage against LHP and play plus defense in CF -- a bench guy like Holt, Hanigan, and Shaw who is absolutely good enough to start for second division teams. In which case, you keep him, even if he's not going to get sufficient PT to earn his salary. In the meantime, if Castillo is still struggling versus RHP come mid-season this year, you can start platooning him with Holt, who is also a terrific defender. Worst case scenario is that you need to get a LF at the deadline, but I think a LF platoon with Holt, Shaw, Castillo, and Young is plenty good enough. The rationale for adding a bat would be if there had been major disappointments elsewhere in the lineup and you felt you needed someone else to hit in the middle of the lineup. But right now, LF is just a complementary piece, not a key one, and as long as it stays that way, we have multiple viable options.
|
|
|
Post by philsbosoxfan on Mar 24, 2016 21:51:20 GMT -5
Put me in the camp that says unless we are totally blown away by an offer, we shouldn't be trading anybody for at least 2-3 months into the season. That includes excess outfielders and catchers.
|
|
ericmvan
Veteran
Supposed to be working on something more important
Posts: 9,015
|
Post by ericmvan on Mar 25, 2016 2:25:15 GMT -5
Not quite, because there's some regression to the mean that'll happen on the range as well. I have almost all of the data I need to derive a kick-ass DRS projection system for OFers, including the crucial adjustment factors for the three factors, which allows me to treat the three positions as one. I think I will try to do that by the weekend. [For geeks only!] Doing it right is a lot of work. But I recently developed a method for regressing small samples just for their smallness. The standard Bayesian way, adding a fixed number of innings or PAs at average level, does not work. Your goal is to eliminate the inverse correlation between the sample size and the effect size, for the positive outcomes (there are fewer extreme negative outcomes in SSS because of selection bias, so they will be reduced even further). So, the steps are: 1) Combine everyone each year into one OFer with the three metrics. Do all three metrics separately. 2) Grab the ages for everyone in the report, plus the names of players who played any OF in 2001-2003. 3) Calculate the career totals for every player going into each season, for the players who began their OF careers in 2004 and later. I can do that with a bunch of cut-and-pastes from a pivot table, but maybe there's a better way! 4) Regress everything to the mean for SSS, as above. 5) Look at the year-to-year and year-to-previous-career correlations for each metric, as a function of sample sizes. Use that to try to determine the minimum number of innings for the regressions. 6) Find the best set of regression formulas for predicting Year from Year-1. Do not assume that it's a straight linear function; I've previously found that negative values regress more than positive. Throwing in the square of the metric as a possible factor should be able to account for most of the non-linearity but I'll still graph the residuals versus the underlying numbers to see if any patterns emerge. 7) Derive regression formulas for the cases where more than one year of data is available (most of them, that is). One of the things I'll throw into the mix is the career to that point. What I've found previously is that data more than 3 years old is meaningless, but I'd like to check that. Once you have these regression formulas, you can look into questions about learning curves and differences between positions. That went over my head like an SR-71 Blackbird. I still appreciate the explanation though. I know enough statistical analysis to not embarrass myself I this board. This is two standard deviations above my mean. FWIW, it sounds like you may be able to extrapolate some interesting conclusions here. Curious to see how it all jibes with scouting. Well, I've got some very interesting results already, which I'll put in non-stat terms. We all know that small samples of fielding need to be taken with a big grain of salt. One rule of thumb is that you want three years of data before you decide how good a guy is. But is this because it takes that many innings for the luck to even out, or because of actual performance variation? IOW, how many innings does a guy have to play before we look at his DRS figure and say, that's how good he actually played? He may play a lot better or worse in his next that-many innings, but that's a real number, as far as we can tell. It's not too high or low because of measurement error, or the luck that always comes with too-small sample sizes. (I believe there's a pretty good argument that measurement error is going to behave exactly like on-the-field luck. It's still a form of luck.) For OF range (Plays Made), the answer appears to be 380 -390 innings. Which is roughly a quarter of a season.
Here's the argument: you don't see any increase in the frequency of unusually good PM/150 scores until you get that low. Below that, you start to see higher scores, and the lower you go, the more and more you see, exactly as you would expect if guys were getting lucky in small samples. But you can set your lower limit basically anywhere above that and still get the same standard deviation of performance of the good players in your sample. There's no greater variation of measured performance among all guys who were better than average over 390 or more innings than among all guys who played 1180. You wouldn't see that if it were possible to have your 400-inning sample inflated by good luck.
Of course, you see bigger and bigger bad numbers as samples get smaller, because the worse you are, the less they let you play out there. And because there's a limit to how good you can be (some balls nobody can catch) but almost no practical limit to how awful you can be (the "Hanley effect"), there is actually more variation among below-average players than above-average ones when you stretch the samples to go below 175 innings. So the first effect is that there's more possible variation among below-average players than above-average ones, and that's what you see in the whole sample; but as you raise the innings limit, you weed out more and more bad players, until they end up having a lot less variation than the good ones.
You would expect that the limit for Good Fielding Plays would be higher, since they are rarer events. And you'd be right. Positive GFP/150 scores start to inflate, apparently from sheer luck, below 800 innings. There's only a slightly greater variance for guys with good GFP than for bad. IOW, it's barely selected for.
Arm, now. Arm is really interesting. For one thing, no matter where you put the innings cutoff, you will see considerably more variation among good arms than bad ones. But that's to be expected. There's a practical lower limit -- not being able to throw out your grandma ("The Damon effect") -- and guys like that do hold down jobs because they can hit and catch the ball. While the good arms range from solid to spectacular. But what's really interesting is the frequency of high scores as a function of PT. There is a big boost in them at almost precisely the point where PM gets a boost, below 380 innings. That would appear to be luck again. But there's a shallow rise pretty much the whole way: the more innings you play, the less likely you are to have a really high Arm/150 score. The reason for that seems clear: at some point, they stop running on you. And that point appears to have a lot of variation. I would guess it happens much more quickly with guys with cannons than guys with average but accurate arms with quick releases ("the JBJ versus Mookie effect"). The data suggests that it can take more than a season before guys stop trying to take the extra base (to at least some degree) on a guy who's throwing guys out by stealth. BTW, that means Rusney's 14.7 Arm/150 was actually 12.7 ... which is still ridiculous. As I expected, a look at guys who had similar seasons shows a usually big drop-off the next year. I'm guessing he'll project to about +5 ... but that's still half a win just from the arm.
|
|
|
Post by umassgrad2005 on Mar 25, 2016 2:58:10 GMT -5
Put me in the camp that says unless we are totally blown away by an offer, we shouldn't be trading anybody for at least 2-3 months into the season. That includes excess outfielders and catchers. I wouldn't make a trade just to make a trade, but if we have a need we should fill it. 2-3 months is a long time if you have a bunch of injuries early on.
|
|
|
Post by umassgrad2005 on Mar 25, 2016 3:10:14 GMT -5
I would throw in some pretty sizeable prospects to make that happen. Rusney, Owens, Travis... ship em out! That'll get you a Trevor Bauer probably with that package, maybe another throw in. I hope you weren't expecting Carrasco with that set of players. Carrasco is starting with a JBJ package followed with probably a lot more. Ohh and Salazar isn't being traded for anything less than a Mookie Betts, which is why we can stop dreaming about Salazar right now. In what world is Salazar worth Betts?? I just don't see it. I get Salazar would cost a bunch of prospects, but one of the best young players in the game? So far Betts has shown he's just a notch above Salazar. Now Castillo, Owens and Travis for Bauer is only a good trade if you think Castillo is a bust and you want to just dump his large contract. I just think Owens alone if given a chance is very close to the pitcher Bauer is.
|
|
|
Post by dirtywater43 on Mar 25, 2016 4:21:45 GMT -5
That'll get you a Trevor Bauer probably with that package, maybe another throw in. I hope you weren't expecting Carrasco with that set of players. Carrasco is starting with a JBJ package followed with probably a lot more. Ohh and Salazar isn't being traded for anything less than a Mookie Betts, which is why we can stop dreaming about Salazar right now. In what world is Salazar worth Betts?? I just don't see it. I get Salazar would cost a bunch of prospects, but one of the best young players in the game? So far Betts has shown he's just a notch above Salazar. Now Castillo, Owens and Travis for Bauer is only a good trade if you think Castillo is a bust and you want to just dump his large contract. I just think Owens alone if given a chance is very close to the pitcher Bauer is. He isn't worth a Mookie Betts to the Red Sox but the Indians don't have to trade him and that's what they would ask for. That's the piece you're missing here.
|
|
|