Let’s start with a tip of my hat to DBacksEurope, who wrote The Reliable Reliever. He applied process control charts to distinguish between reliable relievers and unreliable relievers. It inspired me to explain my ways to determine reliable relievers; there are two methods. One is ‘got the job done,’ and the other is ‘broken goose eggs.’ The goose egg method was developed by Nate Silver and he explained it here.
Got the job done. This method ignores the game situation when the reliever entered the game. This method ignores how many batters the reliever faced. It’s based on the idea that if no runs score while the reliever is pitching, he did his job. Even allowing inherited runners to score would count against the reliever. Given the three batter minimum rule, this idea holds water. Although I could imagine hypothetical situations such as walking three batters to load the bases and leaving the game, that did not happen for any reliever that I looked at. The strength of this method is that it allows for no debate that the reliever got the job done. The metric is calculated by a count of got the job done divided by games in relief.
Goose egg situations. When the reliever steps onto the mound to pitch, whether it is a goose egg situation depends on the inning (7th or later), and score (either tied, his team is ahead by 1 or 2 runs, or due to baserunners tying/winning run is at bat). There are more requirements explained in Nate Silver’s article.
Interestingly, every inning is a potential goose egg situation. In a single game, a reliever can have several goose egg situations (one for each inning he pitches).
There are three possible outcomes of a goose egg situation:
- Broken Egg: In the inning, at least one earned run scores (exemption in the last inning the reliever closes out a win while allowing an earned run).
- Goose Egg: In the inning, no run scores and the total of number of outs while reliever pitching plus inherited runners is at least three.
- Meh: This can happen several ways: An unearned run happens. Reliever leaves game before he completed the inning. Pitcher closed out a win while allowing an earned run.
Broken goose egg rate is calculated by count of broken eggs divided by (broken eggs plus goose eggs). The Meh outcomes are ignored in the calculation.
Looking at three methods.
Let’s look at the same relievers as DBacksEurope looked at to see how the methods compare to his process control method. The following table shows the comparison:
The most surprising result was Fernando Rodney’s 2017 season. Although his process was unreliable, his rate of got the job done was the highest of the five relievers and his broken goose eggs was the lowest of the five relievers. These results strongly indicate that he was an effective pitcher despite his unreliable process control chart.
The process control chart method held up Taylor Clarke as an example of a reliable reliever. His broken goose egg percentage (20%) was better than the two reliable relievers help up as examples (2019 Kenley Jansen and 2019 Greg Holland). Combining his results indicates he has a high ceiling.
For this season through 5 June, for the two Diamondbacks relievers, Taylor Clarke and Kevin Ginkel, the control chart process and goose egg results were consistent. Clarke had a reliable process and a low rate of broken goose eggs, while Ginkel had an unreliable process and a high rate of broken goose eggs.
Interestingly, the two relievers with a reliable control process(Jansen and Holland), had similar results for got the job done (67.7% & 70.0%) and broken goose eggs (26.3% and 30.0%). In future evaluations of Diamondback relievers, these numbers (67.7% and 30%) could be demarcation lines to separate the reliable relievers from the unreliable relievers.
Let’s end with a piece of sports wisdom.
“In any team sport, the best teams have consistency and chemistry.” — Roger Staubach