Marcel is Tom Tango's projection system.  It's freely available at fangraphs.  It's based on a weighted average of prior seasons, aging factors, and regression to the mean.  I've combined the projected wOBA with a weighted average of UZR regressed to the fan scouting report, positional adjustments, and replacement level to calculate each players projected WAR.  I don't include a baserunning component and catcher defense is unregressed.  

The defense line is a combination of UZR and the positional adjustment.  For example, Morneau is rated as a +4 at 1B so his total defensive rating is -8 to account for the -12.5 positional adjustment for a 1B.  

I used a league average batting and defensive projection for Nishioka since a marcel projection was not available.  I think this is conservative as I've seen WAR projections as high as 3.3 from the Oliver system at THT.  I projected Morneau for 600 AB and Thome for 300.  I've got Slowey in the bullpen as a swingman.  Duensing and Slowey's projections are close enough to be interchangeable.  We'd pick up a little more than 0.5 WAR if Blackburn is the one to go to the pen. 

So does 88ish wins and a run at a 3rd straight central division title seem right?  Link to full spreadsheet and numbers below the jump.   

Continue reading this post »

Based on work by Eric Seidman at Baseball Prospectus an average pitcher who pitched both out of the rotation and pen gained 2% on their strikeout rate without walking many more.  I thought it would be interesting to project what our starting pitchers would do out of the bullpen.  I used each players career rates, adjusting their K/9 up by 2% to calculate a FIP as a reliever.  Eric found that ERA decreased an average of 0.75.  His article didn't address whether or not relievers were able to improve their HR/9 and this may account for the difference.  Results are below the jump.   

Continue reading this post »

Joe Nathan thinks he's going to be ready to contribute in 2011.  In an article I read as the off season began, he said he thought relievers could recover from TJ surgery faster than starters because they didn't have to worry about fatigue as much.  Liriano apparently told Nathan he'd just suddenly lose his command and control about 3-4 innings into each start the first year back.  Nathan then pointed to Billy Wagner's return after 11 months in the article as an example of a player who'd returned quickly and to a high level.   

To try and put an estimate on how quickly Nathan might come back I found all relievers with at least 30 saves from 2004 to 2009 who had TJS according to http://www.baseballinjurytool.com/index.php.  I then estimated their true talent pre-surgery using a weighted average of FIP with regression to the mean.  I excluded the year of injury because there was a FIP spike for most relievers the year they had their surgery.  I then compared the true talent to the post-surgery first year results to get an estimate of talent lost.  There were only 8 relievers in this group so it's not statistically significant but I found it interesting nonetheless.  Results below the jump.      

Continue reading this post »

I've taken 3 years of UZR/150 weighted by games played and weighting recent seasons more heavily as described by MGL on the Book blog.  I then regressed that to Tango's fan scouting report.  The amount of regression is inversely related to the sample size, ie big sample = less regression.  1 standard deviation on the fan scouting reports is equal to 10 runs.  I rounded to the nearest half run because I don't pretend the numbers are exact.  Catcher numbers are strictly from the Fan Scouting Report.  All numbers below the jump are in runs per 150 games and are applicable to their labeled position.     

Continue reading this post »

I used the linear log5 method outlined here, http://web.archive.org/web/20080622145043/http://www.diamond-mind.com/articles/playoff2002.htm.  To estimate each teams baseline winning percentage I used BTB Power Ranking which can be found here, http://www.beyondtheboxscore.com/2010/9/4/1670477/btb-power-rankings-week-22-were#storyjump.  The Win% are based on team wOBA, xFIP, baserunning and defense.  Yankees have a estimated win percentage of 0.593 and the Twins are at 0.578.  The historical 0.040 home field advantage was then added to the baseline win percentage.  Using those numbers the Twins have a 52.5% chance of winning home games and a 44.5% chance of winning road games. 

The probability of the Twins winning the series can then be figured by calculating the probability of each sequence of wins and losses that would result in a series victory and then adding those probabilities.  Results and limitations are below the jump.   

Continue reading this post »

At fangraphs, Jack Moore has done a series looking at projecting a players wOBA based on his BB%, K%, BABIP, and XB/H.  It relies on the fact that K% stabilizes after 150 PA,  BB% stabilizes after 200 PA, and ISO after 550.  I've decided to run the numbers twice using both Young's career BABIP and a marcel of his last 4 years BABIP.   The final assumption is a league average distribution of 2B/3B/HR per extrabase hit.   

Using his current BB%+HBP% of 5.6%, K% rate of 13.8%, XB% of 38% and career BABIP yields an expected wOBA of .364 compared to his current wOBA of .344.  The difference comes from his lower than career average BABIP and higher than average number of 2B compared to HR.  Substituting the marcel BABIP, gives a projected wOBA of .359. 

I've included some of the background material below the jump.  Interestingly, wOBA was least sensitive to changing walk rates and most sensitive to BABIP.  Delmon Young just needs to go back to hitting em where they aint and stop hitting balls off the top of the wall instead of over it.  

Continue reading this post »

Platoon skills vary from player to player.  However, yearly splits are a poor way to estimate the platoon skill due to small sample sizes.  One way of getting around this is regression to the mean.  The amount of regression depends on the number of plate appearances.  I used each players statistics from 2002 till now, so Jim Thome gets much less regression than Drew Butera.  The average platoon split is 17 points of wOBA for RHB and 27 points for LHB.  Switch hitters don't have a platoon split on average.  I am using a spreadsheet I picked up on the Book Blog.  Each players platoon split is centered around their ZIPS projection where available.

 http://spreadsheets.google.com/ccc?key=0Ag5QJfW00sh8dFhMak5ta2h0Z0k0VEtSUnJLRDZFWHc&hl=en&authkey=CJOikbkN#gid=0

I started this little project after Jim Tracy was commenting that the Twins are an offensive machine after losing a series at Target field.  He's right against RHP.  We have 8 hitters who are at least average, and 2 hitters with estimated wOBA over 400.  I also quickly realized the Thome is a platoon DH,  mashing RHP and Puntoesque agains LHP.  Hopefully, others find this interesting as well.     

1/2 season of Cliff Lee $17M, 3 seasons of Scott Baker $23M, 3 seasons of Kevin Slowey $18M, 3 seasons of Nick Blackburn $10M, Wilson Ramos $14.2M as a hitter ranked 51-75.  Kyle Gibson is probably ranked in the 26-50 range and worth $15.9.  Aaron Hicks is a hitter ranked 26-50 and worth $23.4M.  Revere is worth between 6-12M depending on whether you rate him 76-100 or a B prospect.  B hitters are worth $5.5M.  B Pitchers are worth $7.3M.  Young C pitchers are worth $2M.  Young C hitters are worth 0.7M.  Methodology after the jump.

Continue reading this post »

Thome as a pH would have a value of about 2 runs above the expected output of Punto/Harris over 60 AB.  The methodology is below the jump.  The rates below are per AB and take into account platoon splits and a pH penalty.  If he only gets 60 pH ABs he would be almost worth his contract.  0.4M + 0.2WAR x 4M/WAR = 1.2M.  If he gets 200 AB vs RHP he'd add another 0.7 WAR and would be a steal.  I honestly can't see why the White Sox would prefer a rotating DH.

Punto .04 runs/AB, Harris .03 runs/AB, Young .02 runs/AB, Hardy  .015 runs/AB.  He would have negative value pH for Mauer, Morneau, Span, Cuddyer, Kubel, although this probably goes without saying.  It is assumed that he'd be facing a RHP who would not be lifted for a Loogy. 

Continue reading this post »

In Nate Silver's chapter, Is Alex Rodriguez Overpaid? from Baseball Between the Numbers he did regression analysis to determine the magrinal economic value of 1 additional win.  The value was approximately linear between 60 and 84 wins and 94+ at $750K.  Wins increased in value by roughly $750K until the 89th win, which was the most valuable at $4.5M.  I currently project the Twins as an 88 win team.  So as long as we don't pay more than $4.5M/win the players salary should be paid for by increased revenue.  The 2 biggest issues with this study are that it's from 2005 and based on WARP1. 

Both defensive and offensive projections were obtained at baseballprojections.com.  CHONE is the name of the projection system.  I then added in the team baserunning totals found at BP from 2009 to the WAR for position players.  I gave starters 600 AB and reserves 100.  I used innings projections from CHONE for starting pitchers.  

This is only an estimate of each teams true talent.  Teams are usually within 7 games, plus or minus, of their true talent.  I also realize that the offseason has just begun so every team has the opportunity to improve.  I'd have posted the excel documents but I can never get them to format correctly.      

Continue reading this post »

Fangraphs is allowing fans to project players this year.  The composite projections for players with at least 30 entries are then posted on the site.  It is a very simple system that lets you pick a range for BA, 2B, HR, BB%, K%, SB, UZR for each player.  You must set up an account before entering any projections.  I'd love to get 30 or more projections for all the Twins.  Follow this link to participate.  http://www.fangraphs.com/blogs/index.php/wisdom-of-the-crowd

This is the time of year where everyone hopes to trade unwanted Twins for unwanted all stars.  How much are some Twins worth?  More than I thought.  How much are some stars worth?  Less than I thought.  The trade value calculator is based on the idea that a team should be willing to trade assets equal to a players value minus his salary.  This is why Joe Nathan is going nowhere.  He is a good player being paid exactly what he's worth. 

Prospect valuations courtesy of Victor Wang and 2009 Hardball Times.  Prospect Rankings courtesy of John Sickles and Baseball America.  Trade Value spreadsheet courtesy of Sky Kalkman at Beyond the Boxscore

Continue reading this post »

Taken from combining wRAA and UZR along with positional adjustments and replacement level for all games at each position.  For example, catcher consists of a combination of the work of Mauer, Redmond, and Morales.  I thought it would be an interesting exercise as we think about areas to improve for 2010.  0 WAR is replacement level.  2 WAR is league average.    

Continue reading this post »

Using David Pinto's lineup analysis tool at Baseball Musings I calculated runs per game based on OBA/SLG projections from CHONE, PECOTA, 2008 stats.  I used Span, Casilla, Mauer, Morneau, Young/Cuddyer, Kubel, 3B, Punto, Gomez as the lineup.  The results were:

CHONE:  4.76 runs/game

PECOTA 4.52 runs/game

2008 stats 5.04 runs/game

PECOTA is not high on the 2009 twins, with everyone regressing except small improvements from Gomez and Young.  CHONE has significant regression from Span, Kubel and Punto with Mauer and Morneau holding the line and improvement from Young.  If everyone hits like last year the offense will be fine.  Getting rid of Monroe, Lamb and Everett helps quite a bit. 

Substituting in Crede at 3B did not make a significant difference in scoring due to his low OBA.  His glove may still make him a worthwhile addition.  Adding Wiggington increased scoring by 0.07 runs/game.  If Gardy could stomach his glove is a different question.

Best lineups all have Mauer leading off, increasing scoring by 0.1 runs per game.  Personally, I'd like to see him hitting second behing Span.

Taken from baseball analysts great website, you can see the original at http://baseballanalysts.com/

Continue reading this post »