# What do you think of WAR? Topic

I guess I do.
2/4/2013 4:05 PM
So luck, should we start giving the player with the highest MLB BA every season a BA+ of 1.000?
2/4/2013 4:30 PM
Posted by all3 on 2/4/2013 4:30:00 PM (view original):
So luck, should we start giving the player with the highest MLB BA every season a BA+ of 1.000?
?
2/4/2013 5:09 PM
2/4/2013 5:10 PM
I have to admit I'm biased against stats I can't calculate myself. I like to see the assumptions being made.

OPS is a simple (well, OK, it's got a lot of terms) mathematical formula.

OPS+ includes park and league factors that are somewhat opaque.

Nobody can even seem to agree on what WAR and Win Shares and Runs Created are exactly.

So I also prefer OPS. But I'm not willing to tell anyone that should be their preference.  It's just mine.

2/5/2013 7:48 AM
Posted by dahsdebater on 2/3/2013 5:32:00 PM (view original):
OBP is the most important individual offensive stat.  You could argue that WAR encompasses more, and if I could only look at one stat when trying to pick players for my team it would probably be something cumulative like WAR, or at least RAR, or TBA, or RC/27.  But in the real world, in which I can look at any number of stats as well as watch a player, the first stat I'd look at would be OBP.  If you look at baseball history, OBP has always been the most critical stat in terms of scoring runs.  More often than not the team that leads the league in runs also leads the league in OBP, and when they haven't they're almost always 2nd.  Last year was only the 2nd time since 2000 in which the AL OBP leader (team) hasn't scored the most runs.  Only reason I know that off the top of my head is because I've had this argument on the Orioles website before, and they happen to be a modern AL team.  But the trend is universal.  The more guys you get on base, the more chances you have to score runs, and the more runs you tend to score.  It's not just on WIS that Tony Phillips was a better offensive player than Dave Kingman or Rob Deer.  Power is nice, but if you can draw walks and get yourself on base you're contributing to your team's offense.
Actually, slg has a higher correlation to runs scored in real life baseball than obp. Using last season as an example, multiply a team's slg by 1725 and a team's obp by 2187. Then dividing predicted runs by actual runs, the cod for obp is 5% while the cod for slg is 3%.
2/5/2013 9:46 AM
That's one season...

Trust me, I have done this math, it won't bear out over the long term that SLG outweighs OBP.  Particularly on the high end.

2/5/2013 4:21 PM
Posted by dahsdebater on 2/5/2013 4:21:00 PM (view original):
That's one season...

Trust me, I have done this math, it won't bear out over the long term that SLG outweighs OBP.  Particularly on the high end.

I used the last 5 seasons and slg still predicted runs better than obp using just a multiplicative calculation.  Runs increase approximately linearly as slg increases whereas runs increase somewhat exponentially as obp increases.
2/5/2013 7:13 PM
Then maybe you're running the wrong regression on OBP...  It should increase exponentially.  If your team OBP increases to 1, you score infinite runs.  Technically it shouldn't run linearly for either stat.  I didn't say it had to be a linear regression...

If you want to REALLY dumb down the calculation, but come up with a very elegant result, one easy way to do it is to compare league rankings in OBP/SLG with league ranking in runs scored.  That is much closer to a linear fit.
2/5/2013 7:15 PM
This guys used 1996-2000. OBP was slighly better than SLG

http://www.eg.bucknell.edu/~bvollmay/baseball/runs1.html

2/5/2013 7:28 PM
But he's treating everything as linear as well, which is frankly ridiculous.  The data isn't and shouldn't be linear.  OBP in particular doesn't even look linear even over the relatively narrow range of the real values.
2/5/2013 8:28 PM
Stan Musial could flat out hit the ball.
2/5/2013 8:49 PM
Just out of curiousity, do you think a walk is as valuable as a single?   A double?
2/5/2013 9:16 PM
Of course not.  Although if you calculate all the probable outcomes of a walk and a single, it's closer than most people think...

I didn't say OBP is the only thing that matters.  I said it's the most important determinate of run scoring, at least on a team-wide basis.  A guy with a .290-.300 OBP who can hit 35 or 40 homers is moderately valuable, but only if there are guys in front of him who can get on base for him to drive in.  A team of that guy would be a fairly average offense in spite of leading the league in HRs by a mile and slugging by a healthy margin as well.  Alfonso Soriano has quietly slugged .489 over the last 3 years - one of the better slugging percentages in the NL over that time (17th) - but nobody wants him at even a reduced pricetag.  The Cubs can't move him.  You don't think his .312 OBP has something to do with it?  Here are a few NL players with lower SLG numbers over the past 3 years: Ryan Zimmerman, David Wright, Andrew McCutcheon, Pablo Sandoval, Hunter Pence, Adam LaRoche, and Jason Heyward.  You think those guys aren't valued head and shoulders above Soriano?  Of course they are.  And their OBPs all exceed his; other than Sandoval, they really all blow him away.

Please note that I never said I would throw SLG away.  But OBP is vastly more important.  That's all.
2/5/2013 10:31 PM
Sez you.
2/5/2013 10:35 PM
◂ Prev 123456 Next ▸
What do you think of WAR? Topic