- FanGraphs Baseball - http://www.fangraphs.com/blogs -

Marlon Byrd, Mike Moustakas De-Luck’d

A refreshed look at the data.

Specificity is both delightful and dangerous. The guys at The Book Blog have previously remarked about how UZR and WAR would be better for mass consumption without the decimal because neither stat can show a true talent level within a single season, but the extra decimal can make it appear more certain or accurate than it is. At the same time, though, the difference between 1.0 and 1.9 WAR can be the difference of a starting job or a bench role (or the difference between 1.6 and 2.4, if rounding is your thing).

Well, today we will err on the side of specificity. In the past, when a player’s BABIP was .498 or .93278, we would just say, “Well, he will regress to the mean,” and then resume our toiling lives. Now, with Fielding Independent wOBA, we can whip their numbers into shape, we can thrust them into the De-Lucker and find out where a regressed BABIP will take them — which is good news for Marlon Byrd, but bad news for Mike Moustakas.

Let’s examine it.

Combining Fielding Independent wOBA (FI wOBA for shortsies) and slash12’s xBABIP, we can get a specific wOBA calculation that strips away unusual luck, whether good or bad. It is important to remember these are both regression-based calculators, so they are backward-looking, not forward looking. Please do not pester me in 3 months when “[X Player] didn’t suck like you said he would! lulzlulzlulz, ur dum!”

As in finance and in life, past performance is not necessarily an indication of future success (or failure), so these numbers simply tell us what their wOBA would have been given the BABIP that probably should have had.

Is BABIP the only stat that has a lot of luck in it? No. Stuff like home run rates can be wild early too. And moreover, BABIP is many parts skill, several parts luck. But there is more luck (or random variation) in BABIP than probably any other hitter stat out there. That is why it is worth focusing on it here.

No more ado; here’s this:

The De-Lucker!


Some assorted reactions to the De-Lucker results:

NOTE: I went ahead and made it for a minimum of 100 plate appearances. But let’s be honest, 100 is just as arbitrary as 50 or 72 or 0.1. If you have a hankering to see this applied to a different grouping of players — or an updated grouping of players (these numbers are current through Tuesday), then simply do the following:

Thing the first. Go to the leaderboards section and do a custom leaderboard like so:

This is just an image. Don’t freak out and click it.

Thing the second. Scroll back up the page and download Excel file:

Encircled and enarrow’d.

Thing the third. And then combine those two spreadsheets by your copy-paste method of choice. I recommend browsing over to the Excel tab entitled “FanGraphs Leaderboard (15)” before pasting, but as long as you unhide everything on the other tab, er’thing will be fine.