"On Paper" Playoff Leaders
This Week's Breakdown
I made several changes under the hood that affect the power rankings this week.
1. The National League fared better against the AL this year than they have in years, and improved for the second consecutive year. Consequently, I've decided to decrease the league adjustment to reflect the interleague record over the past two years. So, instead of giving a 40 run/season boost to AL teams' run differentials (and a similar penalty to NL teams), I'm now only applying a 28 run/season adjustment. The AL still is getting a considerable boost, but it's 30% less than it was a week ago. We can re-evaluate as time goes on whether this is appropriate.
The consequence of this change was roughly a boost of ~0.008 TPI to NL teams, and a drop of ~0.008 TPI to AL teams.
2. Thanks to a suggestion by Nick Steiner (vivaelpujols), I've modified the way I'm accounting for fielding. Previously this year, I was just taking an average of UZR and DRS as reported at FanGraphs. I was never entirely comfortable doing this, however, because both are based on the same dataset, and we've shown many times that there can be significant differences between datasets that can massively affect how we rate players. Last year I had used THT's team fielding stat as a check against UZR, but they are no longer publishing it. Nick suggested that I instead use the difference between xFIPruns (which I use to estimate pitcher runs allowed) and Base Runs (which I've opted to park adjust) as an additional measure of fielding (let's call it BsRFld). There are many ways to drive differences between xFIPruns and Base Runs, but a big one is fielding ("luck" and unreconciled park effects are the others).
Here's a correlation matrix showing how the three fielding measures relate to one another:
DRS and UZR correlate the best to one another, which is no surprise given that they are similar in construction and use the same BIS input data. But I was pleased to see how comparably BsRFld correlated to DRS, even if the relationship to UZR is a bit weaker.
So, I'm comfortable using this. I've decided to now calculate my overall fielding stat as:
( 2*( (UZR+DRS)/2 + Catcher) + BsRFld)/3
As you an see, it's 2/3 "measured fielding" (DRS & UZR) and 1/3 BsRFld (or, if you like, 1/3 UZR, 1/3 DRS, and 1/3 BsRFld). "Catcher" is my home brew catcher stat, very much like what Matt posted the other day.
One of the things that I like about incorporating this is that it helps soften some of the rare, large disparities between estimated runs and actual runs allowed in the power rankings. The Diamondbacks are one of the teams most affected by this: last week, we estimated that they've allowed 74 more runs than expected based on xFIP and UZR/DRS. Part of that could be (and probably is) some bad luck, but some of it could also be bad fielding. We estimated their overall fielding at +4 runs (including catcher), but perhaps those estimates are off and they are instead a below-average fielding team (their park-adjusted defensive efficiency is 4th-worst in baseball). This week, using our new approach to fielding, the power rankings have their fielding at -12 runs, and now they've "only" allowed 35 more runs than expected. That's a number I can stomach a lot more than the 74 run total.
3. In looking at and comparing Matt's catcher data to my catcher data, I discovered an error in my spreadsheet. I was vlookuping the wrong column from the catcher data page, and the result is that all of the catcher data that was being reported and used in the power rankings were wrong (or, at best, incomplete). I don't know how long this has been going on (April?), but it's now corrected. Sorry, I'm a dork. At least, most of the time, it didn't make a big difference.
From all of that, the Rockies seem to be the team that benefited the most. They surged 11 spots in the rankings and gained 38 points in TPI. The Giants got a similar boost from the rule changes, but played so badly this past week that it doesn't show up in the rankings. The team with the biggest dropoff from the methods changes were the Diamondbacks, who fell 5 spots and 44 points this week.
Hopefully you agree that the changes are improvements!
Under The Hood
Converting Runs to Wins
RS = Actual Runs Scored, after a park adjustment
eRS = Estimated Runs Scored, after park adjustment (see "Offense" table below)
RA = Actual Runs Allowed, after a park adjustment
eRA = Estimated Runs Allowed, after park adjustments (see "Defense" table below)
W% = Actual Winning Percentage
pW% = PythagenPat Winning Percentage, based on actual runs scored and run allowed totals
cW% = Component Winning Percentage, using estimated runs scored and estimated runs allowed totals. If you don't like the league adjustment, click in the header and sort by this column to get an "unsullied" ranking.
LgAdj = League adjustment, based on differences in league quality (justification here and here, and modified most recently here). The number shown is the number of runs credited to both the offense and defense of AL teams, as well as the number penalized to both the offenses and defenses of NL teams. By season's end, each team's run differential will be altered by 28 runs.
TPI = Team Performance Index, a hypothetical winning % based on component estimates of runs scored and runs allowed after the league adjustment.
Team Offenses and Defenses
RS = Actual Runs Scored
eRS = Estimated Runs Scored: wRC + EqBRR
wOBA = The Book's statistic, but park adjusted, and using data from both wRC and EqBRR
OBP = On Base Percentage (Times on Base / Plate Appearances)
SLG = Slugging Percentage (Total Bases / At Bats)
wRC = From FanGraphs, with baserunning removed, after park adjustments
EqBRR = Dan Fox's composite baserunning statistics from Baseball Prospectus, minus stolen bases since they are included in wRC.
Clutch = "Clutchiness" measure from fangraphs; difference between actual WPA and expected WPA based on component statistics. We report this in runs.
RA = Actual Runs Allowed, after park adjustment
eRA = Estimated Runs Allowed: xFIPRns - Field
ERA = Straight-up Earned Run Average
FIP = Fielding-Independent Runs, based strictly on K-, BB-, and HR-rates.
xFIP = Experimental Fielding-Independent Runs from FanGraphs. Like FIP, but with HR/Outfield Fly Ball rates regressed completely to league average. xFIP is as predictive as any other DIPS-like stat.
xFIPrns = Pitching Runs Allowed, based on xFIP
Field = Described in this post. It is essentially an average of team UZR, DRS (minus rSB since I calculate catcher fielding separately), and BsRFld. BsRFld is just difference between xFIPRns and park-adjusted Base Runs, and is a less direct approach of measuring fielding. The fielding number also includes a catcher fielding statistic, based on SB's, CS's, WP's, PB's, E's, and this year catcher interference. The methods are essentially those described here. But I'm using B-Ref data this year, and so there are slight tweaks to the methodology, generally in ways that should lead to greater accuracy.
BABIP = Batting Average on Balls In Play. Fluctuates at the team level with fielding, although park effects and chance events can have effects as well.