Twinkie Town: An SB Nation Community

Navigation: Jump to content areas:


Sports blogs for fans, by fans.
Around SBN: SB Nation NFL Power Rankings for Week 11

2009 AL Central Win Projections - version 1

A bunch of us at Beyond the Boxscore are collaborating on developing community Win Above Replacement (WAR) projections for all 30 MLB teams.

BTB WAR Projections

Using the recommended (draft) spreadsheet (thanks Sky!), I've taken a crack at projections for the Twins, as well as the rest of the AL Central (to give these numbers some context). I'm hoping as a group, we can collaborate on an "official" Twins projection and compare to other teams in the division. I'm very interested in your input and suggestions, hopefully we can refine this as we move toward Opening Day. The spreadsheet is posted here.

2009 AL Central Win Projections

There are separate tabs for each team, as well as a rollup "standings" for the entire division. If I feel froggy, I may complete a projection for other divisions and post as well.

Details on how I came to these initial projections, as well as some interesting findings follow the jump.

 

Star-divide

The general method to calculate WAR is documented in an eight part series over at Fangraphs:

Win Values Explained: Part Eight

For hitters, we basically use wOBA, total PA and an adjustment (in wins) for baserunning (basically +/-0.25, does not include SB/CS which are included in wOBA) and for fielding. For pitchers, we use ERA, IP and a leverage factor (closers higher, long relief lower) for the bullpen. WAR is calculated for each player, totaled for the team, and added to a 45 win replacement baseline to get total wins.

A couple notes and disclaimers:

  • This projection is probably biased somewhat toward the Twins, as I follow throughout the year. I don't know nearly as much about other teams, but can adjust as needed if anyone has a recommendation.
  • All of these projections are going to be optimistically biased. Each assumes that no major injuries occur, although individual IP/PA are reduced a bit to account for potential injuries. These projections are more useful in assessing teams relative to one another and should not be considered absolute until I can apply a calibration factor. I can't calculate this calibration factor until I've finished projections for all teams using the same method...

Method:

  1. For the Twins, I used my projected PA/IP allocations across each position. Feel free to weigh in if you disagree. For other teams, I started with the ESPN depth charts and made a few mods of my own.
  2. Projected wOBA is based on a subjective combination of Bill James and CHONE, using Marcel to bump or down a bit.
  3. Projected baserunning is also subjective, but I've got data from 2008 to justify most of the +/- 0.25 values.
  4. Projected defense is based on UZR/150 over the past few years for the player at the given position. I tried not to be too extreme one way or the other with the defensive ratings.
  5. ERA is generally based on Bill James and Marcel projected xFIP, adjusted up or down using CHONE projections.

Findings:

  • Based on these projections, Cleveland is the early favorite to win the division. There's a long way to go before Spring Training or Opening Day though.
    1. Cleveland Indians 91.5 wins
    2. Minnesota Twins 88.0 wins
    3. Detroit Tigers 87.4 wins
    4. Chicago White Sox 80.0 wins
    5. Kansas City Royals 79.9 wins
  • Cleveland (+28.6 WAR) and Detroit (+27.9 WAR) are the two strongest projected offenses in the division. Minnesota is in the second tier (+22.5 WAR), just ahead of KC and Chicago.
  • The Twins have by far the best projected pitching staff (+20.5), followed by Cleveland (+17.9). The rest of the division is bunched around +13-14. The projections expect Chicago to take a major step back on the mound after losing Vasquez and having Gavin Floyd regress.
  • Minnesota, Detroit and KC are all helped by about +1 win by their baserunning and fielding. Chicago (-3.8 fielding) and Cleveland (-0.9 fielding) were hurt. Based on the no-field 1B/DH types KC has signed, I expected them to be much worse. They grade out real well in LF/CF (DeJesus/Crisp) and SS (Aviles/Pena) though.

Thoughts?

 

1 recs  |  Comment 39 comments

Story-email Email Printer Print

Comments

Display:

As a sox fan,

I’m skeptical of Detroit’s success. I think the Sox, as currently constructed, are about an 80 win team. So, this looks pretty accurate- for now.

www.stallingtheman.blogspot.com

by Raf on Jan 8, 2009 1:26 PM EST reply actions   0 recs

I'm skeptical too

I just don’t think Detroit has nearly enough pitching to compete for the division. I was only able to come up with 310 non-replacement IP from their bullpen. Chances are, of the 195 innings I’ve allocated to “Replacement RP” at a 4.70 ERA, those innings may actually come in below replacement, further hurting the Tigers. I am also skeptical of their Edwin Jackson trade, he could very well go above 5.00 again, and who knows what they will get out of Dontrelle. Then again, if their hitters stay healthy (at least healthier than 2008), they may not need much pitching.

Also, I suspect the inherent bias here is something around 10%, or 3-4 wins off the WAR total. That means the projections actually think Chicago is a 76-77 win team. I’m inclined to agree with you, as I have a hard time believing the White Sox rotation will in fact have a 4.65 ERA next year.

by Adam Peterson on Jan 8, 2009 2:02 PM EST up reply actions   0 recs

Since ERA is defense-dependent

Should one maybe use RA or FIP (or maybe tRA) for the pitchers, since the defense is already accounted for with the individual non-pitchers?

Aaron King is still my homeboy... iffy mechanics and all

by baetown415 on Jan 8, 2009 2:46 PM EST reply actions   0 recs

ERA in label only

I agree. It’s a bad label on the spreadsheet. I used xFIP projections from Fangraphs to provide a defense-independent measure of runs for pitchers. I adjusted up or down where there were real differences with CHONE projections.

by Adam Peterson on Jan 8, 2009 3:00 PM EST up reply actions   0 recs

Oh ok, gotcha.

Aaron King is still my homeboy... iffy mechanics and all

by baetown415 on Jan 8, 2009 5:53 PM EST up reply actions   0 recs

Any suggestion for a better label? This has come up a lot.

Email me or post in the BtB thread.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 8, 2009 6:38 PM EST up reply actions   0 recs

Great post

I have 2 minor critiques. I think that the wOBA used for Harris should be higher presuming he’s platooned. I asked Tom Tango on his blog about Marcel projections which assume the same splits going forward. However, in 2008 harris faced roughly 65% RHP and 35LHP. Assuming he gets 90 of his at bats against LHP in 2009 the projection should be adjusted to a wOBA of 0.340. Buscher doesn’t get much of a bump, due to the fact he had the platoon advantage 80% of the time in 2008.

Secondly, while throwing arms usually make minimal difference the twins have two 10 run arms in Cuddyer and Young based on 2007 hardball times stats. Their defensive value should probably be bumped up to account for this.

Thanks for the great analysis.

by Jon Kammerer on Jan 8, 2009 3:02 PM EST reply actions   0 recs

Harris and Arms

Thanks for the feedback. With Harris, I’ve got it as a semi-platoon, if that makes any sense. Not really a strict LHP/RHP, that would be more of a 2 to 1 difference in AB in favor of Buscher. Even though the projections don’t show it (wOBA and defense are pretty similar), I’m expecting Harris to take the job over from Buscher same as he did at the end of 2008. Gardenhire could not stomach Buscher’s defense down the stretch and Harris really picked up his defense after settling in at 3B. I’m projecting to start the season, Harris plays 2-3 times as often as Buscher at 3B, with Buscher playing a couple times a week against the tougher RHP. If this happens, it would put Harris at about 50-50 R-L to start out (2 games/week each), so there should be a little bump there. Then, I’m assuming Punto ends up with at least one stint on the DL (probably sliding into 1B…) and Harris starts at SS with Tolbert providing the same type of cover against some RHP. In the end, 500 PA for Harris feels about right to me. I expect the LHP-RHP split to be closer to 50-50, so perhaps a bump up from .327. Where would you put his wOBA with a 50-50 split? In 2008, Harris’ L-R splits were virtually identical:
2008 vs LHP: .265/.323/.391 (151 AB)
2008 vs RHP: .265/.329/.396 (283 AB)

As for throwing arms, yes, I should probably account for Cuddyer and Young’s arms. I didn’t really take OF arms into consideration. If I bump up Young and Cuddyer an additional +0.25 in the field, it bumps Minnesota up to 88.3 wins. Does that seem right? I’ll need to do for the rest of the division…

by Adam Peterson on Jan 8, 2009 3:29 PM EST up reply actions   0 recs

Harris' splits

There’s a decent change last year was an aberration in Harris’ platoon splits. His career splits are about 90 points of OPS in favor of the platoon advantage (.800 vs. LHP, .711 vs. RHP), and that’s including last year’s neutral split.

"There are only two things that are infinite, the universe and human stupidity, and I'm not sure about the former." - Albert Einstein

by BeefMaster on Jan 8, 2009 3:37 PM EST up reply actions   0 recs

You're right

Yes, last year was likely an aberration, LHP vs RHP. Out of line with his career average. The only reason (other than a statistical aberration, of course) I can think of for this is that with the Twins philosophy of going the other way, Harris improved his hitting against RHP while messing up against LHP (not doing what felt natural, pulling the ball).

by Adam Peterson on Jan 8, 2009 6:26 PM EST up reply actions   0 recs

Arms

I’ve never read anything too definitive on outfield arms and how much value they really add. Just looking at what Cuddyer does in right, I think any statistical measure of his arm would undervalue him. At home, with the baggie, he held so many runners at first on balls off the wall. He has to have the most valuable arm in the league. How many runs/wins this is actually worth is beyond me.

by lookatthosetwins on Jan 9, 2009 12:58 PM EST up reply actions   0 recs

THT has outfield arms ratings.

The best and worst arms are in the 8-10 run range per season. I would project the best and worst arms as plus or minus 5 runs maximum.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 9, 2009 4:48 PM EST up reply actions   0 recs

platoons and arms

Here’s the math.

Marcel wOBA = %vsRHP(x )+ %vsLHP (x + 0.020) where x = his wOBA against RHP. I got this from Tom Tango on his blog. I’m using the average platoon advantage. You can adjust the 0.020 constant if you feel Harris has an above average advantage against LHP. So with a 50/50 split his wOBA is 0.332. 90/10 gives the above 0.340.

I don’t think you need to adjust the rest of the division to reflect outfield arms. Sizemore was the only central division player to make the 2007 leader board at -2.5 runs. I would be tempted to give Cuddyer or Young a 7-10 run bump when playing RF, instead of 2-3 runs. So 0.75 increase would be my recomendation.

Interestingly I calculated the twins as a 86 win team, not counting baserunning or outfield arms. So 88 seems about right. I don’t see many spots for a big improvement, so holding tight and waiting for bargains seems like the right plan.

by Jon Kammerer on Jan 8, 2009 4:19 PM EST reply actions   0 recs

Since my projection is relatively 50-50

I’ll use the .332 number. Thanks.

Given my projections (no Young playing RF, even though I think Young should play more RF and Span more LF), that means pretty much only Cuddyer would be getting a bump for his arm in RF. A 0.25 bump for Young in LF makes sense as well (IMO), considering throws to 2B and Home.

The only player out there that I see being a real improvement, other than a marginal upgrade by adding a bullpen arm like Lyon, is Adrian Beltre. However, we would probably have to give up too much in return to have a net impact worth the cost in salary and players.

by Adam Peterson on Jan 8, 2009 6:34 PM EST up reply actions   0 recs

Well Detriot did get Edwin Jackson....

If you factor Jeremy Bonderman’s return the additon of Edwin Jackson and a healthy Fully Rested Justin Verlander the Detroit Tigers should return to above average League Pitching

a Starting Staff of:

Justin Verlander
Armando Gallaragra
Jeremy Bonderman
Edwin Jackson
Willis/ Vasquez/ Minor

isn’t that bad IMO.

by SteveHoffmanSlowey on Jan 8, 2009 5:53 PM EST reply actions   0 recs

Bullpen

Will Rodney and Zumaya be healthy?

"You're thinking too much. Just have fun." -- Bennie "The Jet" Rodriguez in Sandlot

by cmathewson on Jan 8, 2009 6:05 PM EST up reply actions   0 recs

I'm not sold on Edwin Jackson

I think last year was the aberration for Jackson, and that’s why TB was so eager to trade him (they made out well on that trade). Verlander and Gallaraga are nice at the top, but Bonderman can’t seem to stay healthy. I expect him to be hurt again.

Overall, I see their rotation being a bit below average, with their bullpen well below average. IF everyone stays healthy AND Willis rounds back into form, they have potential to be a bit above average. That’s how they could get back to a league average staff, IMO.

by Adam Peterson on Jan 8, 2009 6:37 PM EST up reply actions   0 recs

I'm also not high on Edwin.

But it’s relative. Some see him as a 4.00 ERA guy as he “continues to grow”. I see him as a 4.75 to 5.00 ERA guy, same as he’s always been, skill-wise. But even some one like that has decent value, maybe 1 WAR. And the Tigers had some awful pitching performances last year.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 8, 2009 6:40 PM EST up reply actions   0 recs

Jackson

His only real improvement last year was about a 1.2 reduction in BB/9. This was largely cancelled out by a 1.8 reduction in SO/9. What’s interesting is that his FIP was 4.90 in 207 and 4.88 in 2008. Virtually the same, yet his ERA went from 5.76 to 4.42. This was primarily due to his BABIP dropping from a sky high .351 to a more sensible .301. Considering that his unearned runs pretty much fell off the map (13 in 2007, 1 in 2008), I suspect much of this BABIP/ERA drop was due to the improved defense. Part was also probably due to luck. Now he moves to Detroit, where Everett-Polanco will be very good up the middle as well. In order to be a pitcher with a FIP in the low 4s, he needs to get that SO/9 back up near 7 and reduce the BB/9 even further to around 3, IMO.

by Adam Peterson on Jan 8, 2009 6:55 PM EST up reply actions   0 recs

Cleveland

don’t know what formulation you used exactly, but if you worked off Cleveland’s ESPN depth chart you should probably have the Indians aligned defensively as Cabrera (2B), Peralta (SS), DeRosa (3B) as that is how Shapiro has talked about it. For some reason they leave out Jensen Lewis on the Indians bullpen. The order should probably go Wood, Lewis, Perez, Betancourt, Smith. Also, the starting rotation looks, for at least the start of the season, to be Lee, Carmona, Pavano, Reyes, and then one of the following group; Aaron Laffey (I think the favorite), Zach Jackson, Scott Lewis, or David Huff. Westbrook should be back by mid-season and neither Pavano or Reyes are locks for the entire season.

I am surprised to see the Tigers score out so well, but that might reflect my bias towards thinking injuries will continue to be huge for them.

by APV on Jan 8, 2009 7:06 PM EST reply actions   0 recs

Thanks

Yes, I started with the ESPN depth chart. I’ll make a few updates to the Cleveland projections. Thought I read somewhere that the DeRosa signing was going to move Peralta to 3B and Cabrera to SS. No worries.

As far as the rotation is concerned, I’ll update to reflect the Pavano signing, but I can’t in good conscience assign him any more than 50 innings. And even that’s a bit high. Agree that Reyes isn’t a lock for the entire season either, this should probably be reflected in an IP projection even lower than 140, possibly 120 or so.

by Adam Peterson on Jan 8, 2009 9:16 PM EST up reply actions   0 recs

Updates have been made for the Indians

I updated the Indians projections as you note above. With those changes, Cleveland improved by +0.5 WAR, 0.3 offense and 0.2 pitching. I didn’t include Westbrook in the rotation, spots were pretty full as it was. Perhaps he would replace Zach Jackson’s innings…

by Adam Peterson on Jan 9, 2009 12:41 PM EST up reply actions   0 recs

Nice work

From a Tigers fans perspective I don’t view this as to Twins biased at all. In fact you have the Tigers 4 wins better than I did. The bullpen is at best uncertain and probably closer to atrocious. They have Rodney, Seay, and the rest I have penciled in as replacement level. A healthy Zumaya would help but I didn’t even include him at this point. The Tigers have a half dozen rookies they could turn to in the pen, the hope is that 1-2 of them are decent enough to stick.

Here are my projections if you wanted to compare notes.

http://www.detroittigersweblog.com/2008/12/tigers-2009-war-projections-pitchers/

by billfer1 on Jan 8, 2009 9:22 PM EST reply actions   0 recs

I suspect I would have each team 3-4 wins better

across the entire league, using these projections.

I think Lopez will provide a bit of RP value above replacement, but agree that Rapada, Dolsi and the rest may end up at around replacement. As you say, hopefully (for Detroit), one or two of the rookies pan out in the pen. Zumaya also needs to be healthy.

Looks like offensively, we’re three wins apart (27.9 vs 24.9), not sure where the major differences are, will have to look deeper.

Thanks!

by Adam Peterson on Jan 8, 2009 9:38 PM EST reply actions   0 recs

Updated Replacement WAR

Folks at BTB have been discussing replacement WAR, and I updated from 2.5 for AL (2.0 NL) to 2.25 across the board. This lowers each team’s offensive WAR by about 2.0, which brings the numbers to a more realistic level. Cleveland moves to a shade under 90 wins, and KC/Chicago end up just below 78. Feels more sensible to me, given the setup of current rosters. Perhaps Detroit a little higher than I expected, but I’ll take a look at other projections and attempt to refine.

by Adam Peterson on Jan 9, 2009 12:44 PM EST reply actions   0 recs

Individual values seemed a bit inflated

as well to me, figured it would work just fine to adjust and use a straight 2.25 across the board for the AL. I’ll see how it holds up as I add teams outside the division. I can set the win baseline accordingly to calibrate.

Also, considering calculation of total wins for a team, I’m wondering if it might be more effective to calculate hitting/running and pitching/fielding totals independently in terms of RAA, then apply the Pythagorean to calculate W-L.

by Adam Peterson on Jan 9, 2009 1:43 PM EST up reply actions   0 recs

Won't it come out the same?

The standard 10 runs per win is based on the pythagorean, so it shouldn’t matter if you use WAR or RAR.

by lookatthosetwins on Jan 9, 2009 2:01 PM EST up reply actions   0 recs

Pythagorean is not linear

the 10 runs per win is calculation of the number of additional runs that must be scored (or prevented) to go from 81 to 82 wins. Going from 82 to 83 wins, it takes a little more runs, etc.

But that’s not my real concern with the method. IMO, WAR is most effective when used to assess an individual player’s value. When one ties WAR to total wins for a team, one needs to determine the replacement level of wins that is added to WAR to get the final result. Every aspect of this calculation (hitting, pitching, running, fielding) is relative to league average at some point (until replacement values are applied). Performing the win calculation using these RAA values allows total runs scored and allowed to be projected without having to define a replacement level.

My guess is that the calculation will not come out the same for teams as they vary from league average offense or pitching, but I have not confirmed this.

by Adam Peterson on Jan 9, 2009 3:52 PM EST up reply actions   0 recs

Exactly

It’s the uncertainty of replacement values that I would want to replace. Then again, once every AL team is done, the league replacement wins takes care of itself.

by Adam Peterson on Jan 9, 2009 9:02 PM EST up reply actions   0 recs

If I can calculate RAA for offense and defense

then I’ll just calculate absolute runs scored and allowed and let the Pythagorean method calculate wins and losses…

by Adam Peterson on Jan 9, 2009 9:03 PM EST up reply actions   0 recs

It's already in the spreadsheets...

I think maybe I’ll add the Pythag method to the next iteration of the spreadsheet. At least a piece that removes replacement level on the team level (leave it in for players).

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 10, 2009 6:17 PM EST up reply actions   0 recs

Sensible idea

I’d be curious to see how this works out

by Adam Peterson on Jan 11, 2009 9:51 AM EST up reply actions   0 recs

trade rumors

It would be interesting to see how some of the rumored trades would affect these win projections, just for fun. Slowey for Hardy, ??? for Yunel Escobar, free agent signing of Blake, etc. You have to think that even the unsabrest (new word!) teams do something simlar when looking at trades. Well, maybe not the twins.

by lookatthosetwins on Jan 9, 2009 1:06 PM EST reply actions   0 recs

So has the market come down for Ty Wigginton at all?

By now, I would suspect that Wigginton knows he can’t get the exact contract he may have been looking for say just 1 and a half months ago.

I would still be intrigued by him as he would fit rather nicely Batting 7th behind Mauer, Morneau, Cuddyer, Kubel

by SteveHoffmanSlowey on Jan 10, 2009 12:54 PM EST reply actions   0 recs

I'd have to think so

especially since Cleveland now has their third baseman. The basic question is whether Wigginton is truly an upgrade at 3B over Buscher/Harris. Defensively, I suspect it’s a wash, Harris a little better and Buscher a little worse in the field. Offensively, I think Wigginton can be an upgrade, but probably on the order of 10 runs / 1 win. As long as we’re talking about a 1 year deal in the $5M range, perhaps a team option in 2010, I’m fine with signing Wigginton. Having Wigginton, Punto, Harris and their flexibility in the infield would give us a lot of versatility when inevitable injuries occur.

All said though, I think I’d rather have Joe Crede on a 1 year contract. Considering defense alone, he’s over 10 runs better than Wigginton or any of our current 3B options. Yes, the injury is a risk, but if he gives us a solid half year with a wOBA in the low .330s and great defense, could be an improvement of over 10 runs, worth the likely one year cost. More than a half year is icing on the cake. Less than a half year, oh well, we’re not obligated beyond 2009.

by Adam Peterson on Jan 11, 2009 1:07 PM EST up reply actions   0 recs

Yes

I’ve been hoping for a Crede deal all off season, if we could get him for a reasonable one year deal.

by lookatthosetwins on Jan 12, 2009 1:05 PM EST up reply actions   0 recs

Comments For This Post Are Closed


User Tools

TT is an SB Nation blog of, by and for the fans. We strive to be the best Minnesota Twins blog by providing quality content and analysis, as well as daily news and notes on the team. We hope you'll make Twinkie Town your home for all things Twins!
Start posting about the Twins »

Join SB Nation and dive into communities focused on all your favorite teams.

FanPosts

Community blog posts and discussion.

Recommended FanPosts

Small
Organization Review (Relief Pitchers)
Small
On Roy, Ramos, and RISK

Recent FanPosts

Small
Anybody want to talk revenues?
Joel87bw5_small
Signing up for the Minors
Small
Roy, Ramos, and RISK, Part II
Small
30 Cents on the Dollar = 2B Indifference
P1060527_small
New Uni Thoughts
Small
Minor League Report...November 14, 2009
Pose_small
Prediction Time (My Guess at 2010 Organizational teams)
Minnesota_twins_vinyl_baseball_small
New Uniforms on Monday

+ New FanPost All FanPosts >

Twinkie Town On Twitter

SPONSORS


Editor-In-Chief

Twinkietown_small Jesse

Senior Writer

Hrbek_small Jon Marthaler

The_jet_small cmathewson

Gladdentwins_small Adam Peterson

Hosken_powell_autograph_small RandBall's Stu