Many (if not all) those in Sabermetrics think the RBI is dead and that it should not be considered when evaluating a player. This is because RBI’s are dependent on players around you getting on base and in scoring position, thus not truly an individual statistic. The opponents of that thinking say there is something to be said about a player being ‘clutch’ and coming through in those situations because more runs lead to more wins. My thinking is that there is a case in the middle, which is looking into how a player moves runners on base. To score runs you need to get on base, then move the runners till they eventually score and there are many ways to move runners than just hits. This is where Runners Total Bases (RTB) comes in.
So what is RTB? Well, it is the number of bases that runner on base is moved by a player.
For example, let’s say Shohei Ohtani is on first base with Mike Trout at the plate. Trout singles and Ohtani moves to 3rd base. Then, Trout’s RTB for that at bat is 2.
Simple right? Naysayers would say that RTB alone is the same as the RBI because it is dependent on opportunities so let’s fix that with RTB%.
For example, let’s assume Trout had 4 RTB and 2 opportunities in one game (i.e. 2 runners were on base total during his PA’s) then RTB% = 2. This means that on average, Trout moves runners 2 bases.
From here we can get metrics like RTB%+, can add weights in for certain types of hits that contribute to RTB, among other things. I would remove plays that result in errors, as that is on the defense, and walks.
Is this metric perfect? Absolutely not and should be tested, but I think it is the start to improving on RBI to get a metric that evaluates a player fairly. The idea that is on my mind of how this helps is optimization: Get players with high OBP and high RTB% and you will score runs. Ideally putting high RTB% players hitting behind high OBP players.
You probably noticed that I don’t have data showing an example of RTB. The data is available via Baseball Savant, but would take a bit of time to code and apply the logic. With that in mind and having a full-time job, I wanted to put the idea out there (good or not) and apply it later. Maybe a future post I can take a small sample of data to show this in action.
Analytics In The Outfield 