Jump to content

Providing independent coverage of the Minnesota Twins.
Subscribe to Twins Daily Email

The Forums

Done with the wrong! What has the FO done right? And...

Minnesota Twins Talk Today, 10:09 PM
I'm sorry, but I've tried to read each and every single 12 page post about everything done wrong with the Twins FO thus far. And I've jus...

Article: MIN 5, DET 3: Odorizzi Turns In Another Quality...

Minnesota Twins Talk Today, 09:57 PM
Jake Odorizzi followed up a one-hit effort against the Yankees by holding the Tigers to one hit through six innings before Detroit tacked...

Article: MIN 6, DET 1: Stewart Impresses, Rosario Exits D...

Minnesota Twins Talk Today, 09:57 PM
Kohl Stewart not only had his best outing so far with the Twins, but one of the better performances of any Twins rookie pitcher this seas...

Article: Playing Hurt

Minnesota Twins Talk Today, 10:09 PM
On Monday night, Eddie Rosario re-aggravated a quad injury while chasing down a ball in the outfield. Add him to the growing list of batt...

Week 3: Vikings vs. Bills

Minnesota Vikings Talk Today, 08:25 PM
The Bills are a horrible, horrible football team. I'm not worried one bit about their odds of winning. A successful game to me looks like...


What does one of the newest predictive measurements tell about the Twins' bats in 2018?

Posted by Thrylos , 19 January 2018 · 1,087 views

Originally published at The Tenth Inning Stretch

With the recent slew of raw data we have about hitters, based on Statcast, new metrics based on that raw data have been developed to predict hitter performance, based on objective measurements, such as bat exit speed and launch angle. One of these metrics is the expected weighted on base average, or xwOBA. It is meant to compare directly to wOBA, in a manner similar to that of xFIP to FIP comparison's for pitchers. And in the same manner if xFIP-FIP is a positive number for a player, as a baseline, you expect a player to improve next season, and the opposite if it is a negative numbers. Of course, this is one factor, and additional factors, like training, mechanics change, adding muscle etc, will affect future performance; however, unlike xwOBA-xOBA, they are hard to measure.

wOBAtries to measure a player's total offensive performance based on a series of weighted operations on offensive events on a players. The link will give you some basic information on the metric. xwOBA is a similar formula based on Statcast exit speed and launch angle. The link explains in detail the metric.

Here are the xwOBA for the 2017 Twins' batters in two groups:

The ones expected to improve in 2018:

Posted Image

The ones expected to decline in 2018:

Posted Image

As indicated only Joe Mauer, and in a lesser degree, Jason Castro are projected to improve, as far as the 2018 startling 9 of the Twins go. Pretty much everyone else is projected to decline.

If one looks at several projections about what the 2018 will do, which are based on xwOBA, expect them to show an overall decline in wins.

There is a silver lining: Other than Brian Dozier (and free agent Chris Gimenez,) most of the Twins' hitters expected to decline based on this formula are young, and the other factors like development, changes in mechanics, could easily trump these projections. If the 2017 Twins' hitters were an older bunch, things would have been different.

  • h2oface, dbminn, ToddlerHarmon and 1 other like this

That stat assumes stability in launch angle and speed off the bat, neither of which is mentioned as a stable trait in batters. The article states "Orioles third baseman Manny Machado produced a .335 wOBA in 2016. But based on the quality of his contact, his xwOBA was .357."Dis Machado regress because of poorer launch angles or exit velocity, better defense as his number of singles dropped or other factors? Predictive exercises can be entertaining, but might be of the same usefulness as the abduction/adductor machine

    • ToddlerHarmon likes this
Jan 20 2018 05:06 AM


That stat assumes stability in launch angle and speed off the bat, neither of which is mentioned as a stable trait in batters.


Zactly.In order to be meaningful, the stat would have to correlate launch angle/exit velocity with pitches, at the least (pops up four-seamers, chops curveballs, etc.) Then, you would also have to learn whether angle/exit stabilizes over a hitter's career.Do we predict off of a young player's second half, whole season, minor league record?


Also, is the stat a good predictor at all?Eyeballing the lists above against baserunning speed, it looks like it might underrate speed's effect on hitting performance, especially for those who hit grounders frequently.


Still, an interesting approach with potential.


Thanks for the post, Thrylos!

Jan 20 2018 10:11 AM
It does not factor in a batter’s speed as a skill.

xwOBA is concerned only with what happens just after the ball leaves the bat, and what the hitter does once he leaves the box doesn’t change it all. Of course, a speedy batter is much more likely to turn a gapper into a double or a triple than a slow batter is. Any difference as a result of speed should not change.

Another factor that might account for differences in xwOBA and wOBA is ballpark factors. xwOBA doesn’t care about the ballpark. Since the Twins are playing in Target Field any difference due to park should not change.

The last of course is luck. This is what xwOBA hopes to measure. Which batters hit the ball better than their numbers indicate? They might expect that their numbers did not match their ability and can expect that they would this year.

I also think a fourth factor might be a hitter’s spray chart. How easy is it to shift the defense in areas where the player consistently hits the ball hard?

What does it mean for those Twins?

Is Buxton difference mostly due to good luck or his baserunning skill? How about Mauer? How much is bad luck and how much is slow baserunning?

How about Mauer vs. Rosario? Rosario’s line drives are spread from foul line to foul line. Most of Joe’s are left field to center field. Which player is easier to shift against? Wouldn’t that effect this number?

It would be awesome to have a measure bad and good luck to help determine a player’s skill level? Baserunning is a skill. The ability to hit the ball to all fields is a skill. I don’t think this number factors those skills out.
    • ToddlerHarmon likes this
Good post Thrylos. Always more fun this time of year to attempt predictive stats than descriptive

Well, just based on xwOBA-wOBA, and just about the whole team declining, I guess it could be a long season. Damn. I might have to pretend I never read this.


It does not factor in a batter’s speed as a skill.


Likely why the first list has three slow runners, an extremely small sample size and the closest number to 0 of all the data points ... Meanwhile the top four negatives are arguably the four fastest players on the team that have a reasonable amount of data to draw from