Totals Projections

Totals Projections for an Individual Game

When making a totals line there are several things, we need to consider most of which are driven by the starting pitcher. We also need input from the bullpen as they will dictate how the game ends. I believe in the statistical approach.

I know that it takes roughly 2.1 hits to score one run, but I will use 2.0 hits as rule here and makes math much simpler. Any starting pitcher stats will show hits allowed so that is an easy number to get or calculate. We do really care about innings because there are times where games started will skewed and cannot give an accurate account of games (for example, SP who previously was a relief guy so he will show large innings but not many games started). We can provide ourselves with relatively accurate games started number to use by taking the innings pitched then divide them by 6. The result will give an estimated number of games started which will be accurate enough for our equations. We are going to use the formula of hits divided by games started so we can get hits per game number for the starting pitcher and we will do a very similar thing for the bullpen. Now we must decide on the influence and or types of game data to use to create this model.

It will be easier for me to show in a graph, so the following picture is my graphs for each pitcher. I use the last 30 days, L5 starts vs Opp., L3 years vs Opp. (vs OPP), Day if applicable, and home/away splits.

The example shows Mile Mikolas and Dallas Keuchel. The figures on the far right show the calculations that produce the run expected total for each pitcher. Miles Mikolas is expected to allow 3.2 runs and Dallas Keuchel 2.7 runs in this example. The data in the column on the left is the hits divided by 2 (using or data methodology from above) then added and averaged for a total. The second column is the innings pitched divided by 6 which produces our games started data. We divide the runs expected by the games started to get runs per game result. Keep in mind this is only half of the equation because the bullpen too will have allowed runs. It is a much simpler process. We are not singling out on players; we are using their collective data as a group. We will look at their last 14 days of work getting the number of games played in that span. We will then simply look at how many hits they have allowed. We will divide the hist by 2, then that total by the number of games they have played in the last 14 days. It will also give a total runs expected from the bullpen.

We will use the string pitcher expected runs, then add the bullpen expected runs to get total runs expected to allow for each team. We then add them together to get a total runs expected. There is also one other data point we must consider which is BB. BB on can create more opportunities for offenses to score runs but usually, they produce big innings of run-scoring as well. Looking at the data, I have learned that if we can project double-digit BBs between both teams, we need to add 2.5 runs to our projected total. When games get that many BBs in them, the scoring does go up, we can get an the edge here if we can predict it.

Using the same process as before, we already have the games started number, we need to total the BB categories then divide by the number of categories. In the case of the example, Mikolas has 4 categories and Keuchel only 3. Mikolas has 35 BB divided by 4 categories returns a value of 8.75 We then take that total our games started figure. IN Mikolas case we get 8.75 divided by 7.10 returning 1.23 BB per game. Keuchel has 24BBs divided by 3 categories for a value of 8.00. We divide that by 5.17 and returns a value of 1.55. Add them together and we get 2.78 BBs from the starters.  In the example, neither pitcher will have a high walk rate. Neither bullpen increases the numbers to double-digit BB expectations, so in this example, we are adding no additional runs.

The example will show that Mikolas should be expected to allow 3.2 runs and their bullpen 2.06 for a total of 5.26 runs allowed. Keuchel should be expected to allow 2.7 runs and their bullpen 1.54 for a total of 4.2 runs allowed. We have now an expectation that we should see 9.46 runs or our line of 9.5.

The actual line in this game was 7.5. The result was STL 7 and ATL 6.