Were the 2016 Angels actually better than 2015’s? The untold story of run differential

Run differential has been seen, for the most part, as a reliable indicator of predicting winning percentage. By no means is this stat perfect, as there are always a few outliers every year. Were the Angels one of them, and were they actually better than what their winning percentage says they are?

If you’re not familiar with run differential, that’s ok! It’s the difference in runs scored and runs given up over a full season.

Let’s take a look at this season’s winning percentages and run differential.

Do you see a correlation? Well, a higher run differential means more games won, right?Let’s take put team run differential versus team winning percentage, just to double check.

Any early observations?? Well, that’s ok. This bar graph was surprisingly more difficult to draw an inference from than I anticipated. That’s alright, let’s put this into a line graph instead. This time, I’ll put a linear trend line in, just to see where the Angels were in relative to this projection.

[You might need to zoom in to see this more closely]

Hey, team run differential and winning percentage are related! There are statistical anomalies, but a higher run differential correlates with a higher team winning percentage.

If you can tell by the graph, the Angels had a -10 run differential on the season with a .457 winning percentage. However, the trend line shows the Angels’ record severely underperformed its projection; the trend line shows a little under .500 (let’s call it 81 wins since that is .500), meaning the Angels should have performed 7 wins better than their record says they did.

But this also brings us to another important point: how much run differential is equivalent to one win?

A little digging on the internet led me to find that with every run differential increase of 10 correlates to an increase of 1 win. But this was just an approximation.

Luckily a little more scouring the interwebs churned out something else. The answer is analyzing RPW, or runs/win! There’s this beautiful formula that we use to calculate RPW, actually. This number fluctuates every year, since the ratio of runs scored to innings pitched constantly differs.

RPW = 9*(MLB Runs Scored / MLB Innings Pitched)*1.5 + 3

Thanks to Fangraphs, we know that this season’s run/win was...9.778.

So how do we use this number to calculate what their run differential should have been, not relying solely on the linear trend line of an Excel spreadsheet?

(Team Run Differential/Season RPW) + 81

I add this amount to 81 wins because it shows the pure effect of run differential on the total wins, and therefore the winning percentage. Let’s try it out with the Angels!

(-10/9.778) + 81 = 79.977 wins, which means we underperformed by 6 wins.

Is this stuff fake? What if the Angels just sucked the whole time?

Let’s take a look at the Dodgers, a team that performed near its trend line in the line graph shown earlier.

(+87/9.778) + 81 = 89.897 wins...Well what do you know?? The Dodgers finished at 91 wins. I’ll do another one if you want.

Let’s take the Atlanta Braves. (-130/9.778) + 81 = 67.704...the Braves won 68 games this season. Pretty accurate, eh? I think you get the picture, let’s move on.

I could do this many more times, but the data shows that the Angels were one of the biggest statistical underperformers of run differential in the entire league this year. But why? Were there any specific reasons for this?

Thanks to Baseball Reference, we can see the run differential broken down by game (pictured below); This can help us analyze what went wrong this season.

A few takeaways from this image.

(1) This team was streaky as heck! This is a team that went on multiple five game winning and losing streaks, with the losses being for a longer period of time.

(2) When we won, we won big. When we lost, it was often close (within 3 runs). In watching the games this season, I cannot recall the number of times where the Angels have the lead only to give it up late in the game OR the Angels’ starter pitched badly (gives up 5 runs) and the Angels’ offense keeps it close, only to lose by a close margin in the end — the latter of which occurred consistently throughout the season.

My Conclusion?

It’s hard to tell. I’d venture to guess that the Angels’ poor luck and extreme streakiness would make it difficult to accurately predict their winning percentage through run differential. There are always outliers in every data set, and this year’s Angels look like one of them.