This winter Baseball Prospectus (BP) developed a new metric called “Deserved Runs Created” that they hope becomes the most advanced and accurate measure of a hitter’s contribution. The math behind it is pretty frightening, and they withhold the actual calculations to prevent another site from copying it, but BP has released a series of articles explaining how the statistic predicts hitter success better than any other blanket hitting stat. A comparison between it and other popular metrics is shown below.
Batting Metric Performance by Contribution Measures (teams, 1980–2018) | ||
Metric | Reliability | Predictiveness |
DRC+ | 0.62 | 0.42 |
True Average | 0.54 | 0.34 |
wRC+ | 0.54 | 0.35 |
OPS+ | 0.50 | 0.35 |
This data was taken from a Jonathan Judge article for BP (on their free version). https://www.baseballprospectus.com/news/article/45383/the-performance-case-for-drc/
Out of context these numbers might not mean much, but the bottom-line is that BP has backed up their claims with comparative analysis.
DRC+ may be the best metric available because it reduces the value given to events not directly correlated to player skill, includes strength of opponent measures, and offers improved ballpark adjustments. It attempts to minimize luck by giving less weight to all statistics overall assuming that most statistics from baseball include aspects that are out of the control of the hitter. This is especially true with stats like home runs. Events more related to a hitter’s skill like walk and strikeout rates are considered more, with strength of opponent adjustments.
The changed ballpark adjustments is perhaps where DRC+ stands out. Colorado hitters have been upset for years at the simple formula used in wOBA and wRC+. These metrics assume the player plays all their games in their home stadium, and uses a 20 year run rate to regulate their production. DRC+ uses data from just the previous year.
Wait, doesn’t a larger sample size mean more reliability? Not always, as Coors Field employees were being hammered for pre-humidor runs per game metrics, and on the reverse Mets players at Citi Field received more favorable conditions after the fences were brought in before 2012 and 2015. Additionally, the alleged (this writer says proven) physical changes made to the baseball during the 2015 season surely throw ballpark adjustments for a spin. To accommodate the loss in sample size ballpark adjustments are weighted less in the calculation of DRC+.
This all results in many Rockies players given more credit in DRC+ and Mets players less credit. A comparison of qualified 2018 hitters from both teams shows this.
Comparison between qualified Rockies and Mets hitters 2018 | |||
Player | Team | 2018 wRC+ rank | 2018 DRC+ rank |
Nolan Arenado | Rockies | 22 | 9 |
Trevor Story | Rockies | 33 | 21 |
Charlie Blackmon | Rockies | 61 | 30 |
DJ LeMahieu | Rockies | 123 | 99 |
Ian Desmond | Rockies | 130 | 121 |
Brandon Nimmo | Mets | 6 | 31 |
Michael Conforto | Mets | 53 | 64 |
Amed Rosario | Mets | 124 | 134 |
Yes the Mets only had three qualified hitters last season, however the MLB team average was only 4.7, and we did have Asdrubal Cabrera for much of the season. His story leads to the next reason DCR+ seems to be superior: how it measures players that switch teams year to year (last table, I promise).
Reliability of Team-Switchers wOBA (2010-2018) | ||
Metric | Predictiveness | Error |
DRC+ | 0.50 | 0.001 |
wOBA | 0.37 | 0.001 |
wRC+ | 0.37 | 0.002 |
OPS+ | 0.37 | 0.001 |
OPS | 0.35 | 0.002 |
True Average | 0.34 | 0.002 |
OPB | 0.30 | 0.002 |
AVG | 0.25 | 0.002 |
This table was in the same Jonathan Judge article. Nothing says more about effective ballpark adjustments than being the best at predicting a player’s success in a new ballpark. Fortunately for the Mets, offseason addition Jed Lowrie finished 22nd in DRC+ while only 46th in wRC+. Robinson Cano and Wilson Ramos failed to qualify in 2018, but their DRC+’s show that they were a little worse in 2018 than their wRC+ shows. However, the two of them rank about 20%-30% above league average in both statistics. Hopefully they help this dismal offense rebound in 2019.
My preference is for a site that creates a metric to make the calculation behind it publicly available.
I love this – great stuff Brendan. Thanks for bring this to light. It’s about time that luck is weighted appropriately. Something that comes to mind is a player scoring a run on a pop fly to shallow right field, the sun is in the OF’ers eyes and he never even makes a play on the ball, that pop-up likely has a 99% out-conversion rate. The hitter/runner’s statistics should not benefit from an error or lucky timing in my opinion.
I wonder if they’ll ever account for umpire strike-zone tendencies when assessing a pitcher’s performance..or wind factors