How badly do the Sox need another bat for 2010?
by Buzzy on Dec 19, 2009 11:11 AM EST
44 comments
Here on OTM there has been a recent flurry of statements concerning the Sox offense for 2010. Much of the speculation centers on the obvious fact that the Sox have made efforts to improve their run prevention ability for next year. The signings of Cameron, Scutaro and Lackey make that clear. At the same time, the Sox will lose perhaps their most potent bat from 2009, Jason Bay. In addition Mike Lowell, a decent offensive player, will also move on. Scutaro and Cameron are not spectacular offensive players. Given these facts, many have cried for the signing of, or trading for, a player or players that could provide the needed offensive upgrade. The quantitative question is, how much of an upgrade would be needed to put the Sox in the ballpark of the offense they had in 2009?
I am assuming that the Sox had a very good offense in 09. The team was 3rd in the league in runs scored, and trailed the Yankees (number 1) by a mere 43 runs. No Sox offense since 2005 scored as many runs as the team did last year. While there have been complaints that this was a mirage, a reasonable analysis (that I will not provide here) shows otherwise. One should always be suspicious of the claim that certain offenses only hit well against poor pitching. Generally speaking, teams that score a lot of runs are good offenses.
The basis of this analysis will be the statistic wRC, an estimate of how many runs a player contributed or contributes to total offense. If you go over to Fangraphs, you can check that a sum of wRC for a season is close to the total number of runs scored, which speaks to accuracy. If you prefer James' "runs created" stat, the picture I present here will be unchanged. You can find RC on baseball-reference.com. What I will do is breakdown the Sox 2009 by wRC, and compare that to the James projections for wRC for 2010. I do this to compare "apples to apples." Note also that this attempts to guess what the Sox offense will look like next year.
Here is the breakdown in terms of "runs contributed."
2009
Youkilis-111.7, VMart/Tek-(42.3/43.1), Bay-112.5, Drew-91.3, Ellsbury-97.3, Pedroia-103.9, AGonz/Lugo/Green-(20.1/14.8/27.2), lowell-65.1, Ortiz-80.9.
This totals to 810.2. Note that this is close to the team's total runs scored last year, and does not include the contributions of Kotchman, LaRoche,Baldelli, Kotsay, Lowrie, etc. We should perhaps include the estimated 10RC that Kotchman, LaRoche and Kotsay contributed at 1b. This brings the total to 820.9, without other bench play.
The James estimates are
2010
Youkilis-105.4, VMart-95.8, Cameron-76.3, Drew-93.5, Ellsbury-97.1, Pedroia-105.6, Scutaro-72.5, Ortiz-91.9. This does not include a 3b at all (or a extra 1b if Youkilis moves over). The total is 738.1. Now lets add the additional player-Kotchman is projected to be 49.8 wRC in 408 PA. Lowrie is predicted to be 58.9 wRC in 437 PA, and for fun, Beltre is predicted to be 61.7. All of these work out to be about 60 in a full season. This puts the estimate to about 800 runs even. Thus, a reasonable evaluation of the numbers places the 2010 Red Sox at a mere 10-20 runs behind the 2009 Sox, despite losing Bay who contributed more "runs" in 09 than any single player.
Where does this come from? Well, Cameron is actually decent. Scutaro and VMart are more than 10 run upgrades each. Also, Lowell was not offensively effective last year. Of course this evaluation is crude. For example the benches have been left untouched, I have not factored in adjusting players like Scutaro, Cameron and Beltre (if he were to sign with the Sox) for new parks. In fact that fact alone will move the evaluation even closer to 2009's number.
The bottom line is that the Sox offense, even without someone like Adrian Gonzalez, will be close to what it was last year with Bay and Lowell. On the other hand, a full season of Lackey, Dice-K and Buchholz combined with the improved defense (potentially at 3-4 positions, eg LF,CF,3b and ss) will save a lot of runs. I am excited for 2010 even if the Sox don't pull of any more deals.
5 recs |
44 comments
|
Email
Print
Comments
Rec'd
I Still want Beltre though :)
"That was a lot of fun… You just keep pounding balls into the gap. The one thing you don’t want to do is hit a home run. That’s a rally-killer." Jeff Francoeur
Thanks.
Beltre would probably save 10-20 runs on defense. That is where is realy value is (as you know).
Question.
We all know runs and RBIs are a useless way of evaluating individual batters, because they can’t help who hits in front of them or behind them. Right?
So then how can we correlate a couple of inches/feet of range with saving runs? Wouldn’t it be easier and more accurate to correlate defensive advantage to “outs”?
Manny ain't the only bad man.
Hi
Runs and RBIs are the most important stats there are (especially runs, as that is the very definition of offense). They are not so useful in evaluation of a player because they cannot be taken out of the context of other players. Therein lies also part of the problem with any defensive metric, since defense is an inherently more collective endeavor than offense in baseball. However I should say that if you do accept part of the concept UZR (even if uou don’t accept it fully) then you should note it is not all based on range, and that outs can be (approximately) valued in runs anyway.
I also did see your back and forth on UZR a couple of days ago. I don’t want to argue that UZR is perfect or even as reliable as your favorite offensive metric. As I said, defense may never be quantifiable in that way, and even if it is, UZR will likely not be the number that is used. However I wanted to point out that:
a)You can make statements about how much UZR fluctuates compared to offensive metrics. The bottom line is that it does fluctuate (year to year) more, but perhaps not as much as you would think. Part of the fluctuations are not even connected to UZRs flawas, but just the number of chances. For example UZR fluctuates more for OFs than IFs, which makes sense since they get fewer chances.
b)Although a deep study has not been done, it is clear that team UZR correlates with team earned runs minus team FIP (converted to runs), while team fielding percentage correlates with nothing. I also think that team BABIP correlates with team UZR (but this I have to check).
c)As I have mentioned team WAR (with a large UZR component) correlates more with W-L than does team runs or team runs allowed.
Given these things (and I know this was not your point) I am happy enough with the UZR coversion to runs, especially for a guy like Beltre who has a very consistent UZR, and is one of those guys for which your eyes also tell you he is awesome.
I should also add
on the topic of your original point-it is likely more useful to try to convert everything to a runs scale (created/saved) becuause as a predictor of W-L some variant of Pythagorean is best, and it is based on run differential.
But as UZR is context-neutral
how can it account for runs? Does it take the average number of hits that would equate into a run, then show how many hits were saved by say Beltre over the course of a season?
And how many hits exactly would a Beltre-type save over the league average?
Manny ain't the only bad man.
Hi Tommy
there is a difference between the stat “runs scored” by an individual player and a context-neutral measure of the runs created by a player (eg wRC). The remarkable empirical fact is that one can estimate the runs a team score via essentially context-neutral measures. Hence, the total of a teams wRC or RC is very very close (often within 5%) of the total runs scored by the team. Such metrics only use ingredients that are not context specific (or only indirectly so). Indeed, that this is true is clear from the remarkable empirical fact that something like OPS or wOBA (on the team level) correlates remarkably strongly with total runs (the target metric of an offense), eg R=0.95 or so.
While not as accurate, UZR aims to do the same thing for defense.
Hi Buzzy
If the correlation is based on team UZR, I’m wondering how you can project that Beltre alone would save 10-20 runs over X.
On average, how often would he make a play that Youk wouldn’t make? Once a game? Every five games? And then what is the ratio between that extra hit to runs?
Or does UZR not measure that?
Manny ain't the only bad man.
UZR
team UZR is just the aggregate UZR of the individual players (and their positions). As for your question, the “runs” factors of UZR are broken down by range and error components. UZR as you know is based on measuring the number of balls in play a player is able to convert to outs compared to average. Outs are converted to runs (saved/lost) essentially via the algorithm discussed here (this is old and has been updated)
http://www.baseballthinkfactory.org/files/primate_studies/discussion/lichtman_2003-03-14_0/
Thus, you have an estimate of runs saved/lost. Last year, Beltre was estimated to have saved 20 runs over an average 3b if he played 150 games. I would bet this algorithm is not nearly as good as the one used to estimate runs scored (eg the one used in this post) but it is the best I know. If you believe it, Beltre’s defense alone would be the difference between the offensive runs in 09 and projected ’10 without Bay.
One other thing-
the only way to test these things is at the team level, but that does not mean it cannot be estimated at the individual level. For example, for runs you can estimate each player’s contribution and, as a test of the accuracy, see how well that corresponds with how many runs the team actually scored. One can try to do the same for run prevention. It is more complicated of course. On can try to separate pitching from fielding via, say, FIP or tRA, and then use UZR for fielding. The correlation is there (on the tem level where it can be tested) but not as strong as for offense.
Also,
I believe that UZR measures the way a ball is hit (by velocity, direction, trajectory, etc.) and converts those numbers to an estimated context-neutral run value. So when a 3B saves a scorcher down the line (that would be a 2B), it has a different value than when a 3B charges a makes a great play on a ball that would usually be an IF hit.
So looking at Beltre compared to Youk at 3B, he would save something like 15 runs over the course of a season which is probably about 20 plays over the course of the year.
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
From memory,
a double is worth something like 1.1 runs and a single .75 runs. Most infield plays would be preventing one of those two, so 20 plays sounds about right.
Offensively, players are defined by just as few plays. The difference between a replacement level player and a solid player could very well be 20 HR’s.
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
I think wRC is derived directly from wOBA.
Speaking of that, FG’s now has wRC+ which adjusts for park and league and scales it like OPS+, ERA+ etc. Didn’t you have a bet with somebody about league-adjusted wOBA or something?
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
ERA-FIP and UZR are more closely connected the higher you get.
Basically, there’s a lot of fluctuation around the middle—+10 UZR teams ending up with positive ERA-FIPs and vice versa—but when you hit the top of the list—the +50, +60 UZRs—you start seeing a notable difference where, just based on ERA-FIP, you’re saving 50 runs a season.
USG
by Ben Buchanan on Dec 19, 2009 3:19 PM EST up reply actions
This is based on exactly 2 seasons worth of research, mind you, so I'm not preaching gospel.
But it seems to me like that would make sense. A truly strong defensive team will see better returns per UZR on its investment than a mediocre team, likely because glaring holes in the defense are glaring holes.
USG
by Ben Buchanan on Dec 19, 2009 3:20 PM EST up reply actions
Wouldn't a +10 UZR equate to a negative ERA-FIP,
since the above-average defense would lead to an ERA lower than it should be?
Brett Anderson is the Truth. Brett Anderson is divine presence. Brett Anderson is eternal life. Brett Anderson is within you. Brett Anderson is here. Brett Anderson is Now.
by Frederick0220 on Dec 21, 2009 1:26 AM EST up reply actions
Yes
Which is why there being contrary numbers at middling levels of UZR shows that there’s a lot of fluctuation.
USG
by Ben Buchanan on Dec 21, 2009 5:16 AM EST up reply actions
I feel like if we aren’t first or second in runs scored for a season, then people will complain that we don’t have enough offense. With the pitching we have in place now, we could wind up 10th in offense produced and still have plenty of wins.
Rec'd
Not to come of like I was arguing you or anything either, I was just agreeing with the last part of the article.
by Justin_Bobo on Dec 19, 2009 12:49 PM EST up reply actions
Yup.
We scored 1000 runs a few years ago. That does not happen too often, people.
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
The Sox scored 961 runs in 2003.
I gotta go 'cause I'm probably definitely gonna nod out again.
by Drugs Delaney on Dec 20, 2009 8:05 PM EST up reply actions
Well argued (with only a couple of teeeeeensie kvetches)
Agreed with general theme and almost every single item cited; small quibbles only: I would note that on the measure of runs scored, the most-valuable-player for the Red Sox might have been Fenway Park, as in most years. Not saying that the Sox wouldn’t have been a very good team playing in, say, Petco, just that on that particular measure, a tiny caveat is in order.
Second, I’m not entirely sure what measure “runs contributed” is supposed to be, being myself more familiar with “runs produced” (R+RBI-HR) and “runs created” (essentially the VORP measurement). The decimal points in your figures suggests it can’t be the former, which is always an integer, but then I don’t understand how you get Bay ahead of Youkilis, because the only measure I can think of that says Bay’s the team’s most potent bat is the simple runs plus RBIs one. Mostly curious, but also to add the really big point, that the Red Sox’ offense received a tremendous contribution from Youkilis, Pedroia, and Ellsbury that exceeded (on a VORP of 134) the top three players from, say, the Angels or Rangers, or Oakland’s top six. Bay’s numbers tell the story of a better season from him than I remember, and he actually comes quite close, but I have to disagree with any measure of him as the Sox’ best hitter.
But again: Nicely argued; compelling, and more importantly, correct.
Tony the Pony
Hi Tony
Agree on the park. For most of the players, projections are just based on overall numbers. This should be accurate for most of the players here as they were Sox last year. For the additions this year, as I mentioned, they are likely underestimations (Cameron and Scutaro both played in parks not nearly friendly for hitters as the will this year). Of course this does not say if X runs for a Sox offense means they are good or bad. Either way they should not be that different than last year.
wRC is closer to runs created. However RC overestimates things the same way OPS does. wRC is like runs created but using the (likely) better stat wOBA as a scale/variable. Bay was not a better hitter than Youkilis, but he was close and played more, and thus contributed more runs. I think by RC Bay was 117-close to the 112.5 above. I used wRC not only because I am somewhat biased to think it is a better measure than RC, but also because Fangraphs gives projections for 2010. I should say that I think RC might correlate better by linear regression with total runs than wRC (much the same way that OPS does a tad more than wOBA) but it is not really significant.
I know we’ve discussed this, but Tango had a lot of issues with the way that the OPS v. wOBA study was done, from what I remember.
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
I actually have to say
that from my investigations I don’t agree with him. There is not reason that a stat like wOBA should just be a player specific metric that need not correlate at the team level. That is a copout that is always used, and it is not scientific. The fact that OPS and wOBA are both as context neutral as possible means that the team OPS/wOBA should be the metric you compare with total runs. While wOBA has a more satisfying rationale, it is not derived any more than a totally random number like OPS is (nothing is derived in baseball stats as there is no underlying fundamental theory). All of this should be viewed as empirical, and correlations are correlations. What may well be possible is that in a more complete statistical investigation, wOBA is actually shown to be empirically superior by looking at non-linear correlations. What this means is that it is possible that by looking at more subtle correlations one can show that at the team level wOBA actually has better correlation beyond the linear regression correlation coefficient, but those studies have not been done. I should say, however, that the differences in R with total runs is small anyway.
His disagreement was in the way that wOBA was being converted to runs, not in the correlation. Tango said:
I do not know what Colin did in order to convert wOBA to total runs created. Seeing that wOBA is Linear Weights, then I expect an r = .999 of any results coming out of wOBA or Linear Weights. If you don’t see that, then there’s a gap somewhere in the translation.
It gets a little too technical for me at a point. But wOBA is derived from past empirical results. So it is not based on a fundamental theory, but is much less arbitrary than OPS. To me, OPS lacks any conceptual sense, so I prefer wOBA, but they are close enough that I really do not care too much.
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
Well done.
Maybe a part 2 showing the estimated runs scored against for next year?
"Ninety percent [of my salary] I'll spend on good times, women, and Irish whiskey. The other ten percent I'll probably waste."
-Tug McGraw
There is this theory rolling around the blogs that the Red Sox somehow NEED
Adrian Gonzales. I dont agree with this theory. Now after seeing Buchholz turn around his career and become a solid rotation arm, the signing of Lackey and hopefully Dice-K’s pride kicking in, the Red Sox have potentially a 90s ATL Brave rotation.
Adrian Gonzales would be a luxury, there are other players out in FA or on other teams the Sox could get to replace Mike Lowell at 3rd and help hit in the lineup.
If it was just Buchholz and some prospects Id say do it, but Hoyer knows that Kelly and Westmoreland are the gems of the system and will not give up AGon without them.
So its time to move on and find another solution, but I agree with Buzzy, the offense will be better this season even without AGon.
One argument that hasn't been addressed
is the Sox lost in the ALDS largely because of a lack of offense. Since that was with a squad that put up the 3rd most runs, you could argue that the team needs players like AGon to improve the offense and give them a better chance to score runs in the playoffs.
I’m hopeful that the offensive outage was just some freak occurrence, and they’ll roar into October in 2010.
"It's just a tiny little nick, but it hurts when I get champagne in there."
- Jason Bay, on getting spiked scoring the winning run in ALDS Game Four.
it has been addressed
and shot down. It was a three game series. There is not a team out there that will not occasionally slump for 3 games.
I would argue that because the Sox put up the 3rd most runs, offense should be one of the lowest priorities this offseason.
Defense should be number one- defensive metrics are sketchy at best, but we looked awful by all of them.
I don't know
when I watched the ALDS I saw something different-defensive falures. You could really argue that games 2 and 3 were lost on defense. In game 2 Beckett has given up one run (Abreu) that got on on a very weak single to 3rd that probably a decent 3b converts to an out. Then he gives up basically the rest on a 2 out triple that, somehow I still can’t figure out, Ellsbury can’t come close to, even though the ball was not hit hard or far and landed feet in front of the warning track in CF. In game 3, Youkilis makes a horrible play that somehow is ruled a double to start the 8th (it was a clear error). Then in the same inning Pedroia botched a clear DP with Kendry Morales running. Sure, it would have been really nice to score more runs (1 in 2 games ain’t good) but that can always happen in the playoffs where the games are tight and pitching good. Games 2 and 3 really could have gone completely the other way if the Sox didn’t f-up repeatedly (nmuch like the 08 Angles did no baserunning and fielding gaffes).
As a hated Yankee fan..
The thing I noticed about the Red Sox batting order from the 2003-2005 teams to the current teams is how less scary that lineup is.
As I Yankee fan during the 2003 season, I remember how scary the Sox lineup was from top to bottom.
Now looking at the lineup…unless Bay resigns…who really scares you in the Red Sox lineup. Youkilis is a great hitter, but the problem is that he is like Derek Jeter on the Yankees. He’s not the bat that can take over a lineup. Same with Pedroia.
I as a Yankee fan cheered the day that Manny Ramirex left It turned what was a frightening 1-2 punch in the order to a mediocre lineup with an aging Ortiz.
Now it appears the Yankees have the scary 1-2 punch in Tex and A-rod.
Now, Adrian Gonzalez would scare me.
It would be very nice to improve the offense
But for now, the offense was very good last year. The defense was not, so our number 1 priority NEEDS to be shoring up the defense to make the team well-rounded. All the home runs in the world won’t save us if we can’t field the ball.
Gonzalez really would be nice. He’s great defensively too, don’t get me wrong, just the worry is that the blow our pitching would take would be too much.
The offense last year was better than it was in 2007
When the Sox won the WS. As long as the Sox’ pitching and defense improves, they can take a slight hit to what was a very good offense last year.
Speaking of A-Rod, he was pretty scary in the field last year. He’ll be 35-years old next year and is coming off a hip injury ….
I gotta go 'cause I'm probably definitely gonna nod out again.
by Drugs Delaney on Dec 23, 2009 4:33 PM EST up reply actions
Unfortunately
“Fear” is a statistic like “Clutch”. Made up and worthless. If you can assemble a team that only hits singles but with extreme patience that takes walks and gets on base at a .400 clip, none of them will scare you but that offense is going to score a lot of runs.
"We are not normal, We are Legends. People will tell their kids about us." - Deon Butler before Ohio State Game 2008.
measuring Red Sox production
The analysis is worthless because it doesn’t split out home and away games.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
In Association With
Big League Stew - MLB - Yahoo! Sports
- The Juice: Swisher ends Orioles sweep dreams with timely homer
- Create-a-Caption: Dusty Baker wants YOU to buy playoff tickets
- Padres' Heath Bell wistful about Trevor Hoffman's 600th save
- Want your first career home run? Face a pitcher from Chicago
- Slumpbot .200: Phil Hughes runs into New York's Joba Rules
Certain photos copyright © 2010 by Associated Press or Getty Images. Any commercial use or distribution without the express written consent of Associated Press and Getty Images is strictly prohibited.
