19 October 2015

The 2015 Orioles Offense Wasn't Good. It Wasn't Bad, But It Wasn't Good

The Baltimore Orioles in 2015 scored more runs than they did in 2014 - 713 to 705. The Orioles won the American League East Division in 2014, and so there is an intuitive belief that their struggles in 2015 could not have been caused by shortcomings in their offense. (For the rest of this article, I will be using "offense" as a substitute for "run production.") After all, if 705 runs were enough to win a division title in 2014, 713 runs should have been enough to win a division title in 2015.

And there's more. Many of the Orioles' starting pitchers in 2015 were worse than they were in 2014. Chris Tillman allowed 83 runs in 207 innings in 2014; he allowed 97 runs in 173 innings in 2015. Bud Norris allowed 68 runs in 165 innings in 2014; he allowed 57 runs in 66 innings in 2015. Miguel Gonzalez and even Kevin Gausman surrendered more runs in 2014 than in 2015. Isn't it obvious that the 2015 Orioles problems were on the defensive (shorthand for "run prevention") side?

But a further look at the 2015 Orioles offense calls that into question. The 2015 Orioles got substandard offensive production from their left fielders, their shortstops, their right fielders (except when Chris Davis played right field), and their designated hitters (except when one of their other regulars was taking a half-day off.) As good as they are, I can't quite believe a five-man offense of Manny Machado, Chris Davis, Adam Jones, Jonathan Schoop, and Joseph Wieters is an overpowering strength.

Can these apparently contradictory impressions be reconciled? How good, really, was the 2015 Orioles offense/run production? And, conversely, how bad was the 2015 Orioles defense/run prevention? How good would the 2015 Orioles have had to be at preventing runs in order to be a contender?

Before I continue, I will be using 88 wins as the target for "contender" in this article, although 86 wins would have gotten the Orioles into a 163rd game playoff with Houston. Using 88 wins as the target is more realistic and allows for unforeseeable and unpredictable differences had the Orioles been a different team with different playoff race dynamics.

Also, unless stated otherwise, all numbers come from baseball-reference.com.

Was the 2015 Orioles Offense Really Good?

No, it wasn't; despite the fact that they did score as many runs in their 2014 division-winning season. The primary reason is league context. In 2014, the average American League team scored 677 runs, so their 705 runs was 28 runs above the league average. In 2015, the average American League team scored 710 runs, so their 713 runs was three runs above the league average.

We do have to make one fairly significant adjustment. Some of the increase in the American League's runs scored was because the Orioles themselves allowed more runs in 2014 than in 2015. Because of interleague games, some of the increased runs belong to National League teams, but following that path gets us bogged down in complicated and unverifiable assumptions that are unnecessary for the general picture. The Orioles allowed 101 more runs in 2015 than in 2014; subtracting that from the league average decreases the team average to 703 runs, so the Orioles were ten runs better than the (adjusted) league average.

There's one other factor which I will mention but won't try to account for. In 2014, Camden Yards favored run prevention with a park factor of 97. Camden Yards favored run scoring with a park factor of 109. Some of the Orioles increase in runs scored can be attributed to the increased offense in 2015 Camden Yards.

Were the 2015 Orioles Really Bad At Run Prevention?

Again, no, they weren't; as with the offense, they were about average in run prevention. In 2015, the average American League team allowed 695 runs; the Orioles allowed 693 runs. The Orioles were actually two runs better than the league average in runs allowed. Adjusting this total for the increase in runs that the Orioles themselves scored wouldn't change this significantly.

What stands out is that the 2014 Orioles were really, really good at run prevention. The 2014 Orioles allowed 77 fewer runs than the league average; good enough for third best in the league. And it was the 2014 team that was the outlier - the 2012 Orioles were 8 runs better than league average; the 2013 Orioles were 13 runs worse than league average. So, the perceived weakness in run protection is really just perceived - the 2014 Orioles were (probably) playing at an unsustainable level.

One other point worth noting, although it has very little to do with the Orioles - the American League, in total, was slightly better than the National League. The American League teams combined to score more runs than they allowed. Since the major league totals must be equal, this means that in interleague play the American League outscored the National League.

How Good Did the 2015 Orioles Defense Need to Be?

In order to determine how likely it would be for the 2015 Orioles to contend given their offense, we can use the Pythagorean relationship between runs scored, runs allowed, and winning percentage to determine how few runs the Orioles would need to allow to get to 88 wins. The Pythagorean relationship is:

[runs scored]2 / ([runs scored]2 + [runs allowed]2) = winning percentage

Given that relationship, and that the Orioles scored 713 runs, we can determine how few runs the Orioles would have had to allow to get to 88 wins. 88 wins is a .543 winning percentage, so plugging our known quantities into the formula we get:

508369 / (508369 +  [runs allowed]2) = 88 / 162

Cross-multiplying (a technique from high-school algebra) gives us

44736472 + 88 [runs allowed]2 = 82355778

Solving for [runs allowed]:

88 [runs allowed]2 =37619306
[runs allowed]2 = 427492.1136
[runs allowed] = 653.828

So, for the Orioles to have won 88 games in 2015 with their offense, they would have had to allow only 654 runs, 39 fewer than they actually did. They needed to be fifth-best in the American League in run prevention.

How Good Did the 2015 Orioles Offense Need to Be?

Just as we did above, we can determine how many runs the 2015 Orioles would have needed to score to achieve 88 wins given that they allowed 693 runs. While the level of math involved is the same as above, it's somewhat harder to write each step clearly, so I'll just include the important steps:

[runs scored]2 / ([runs scored]2 + 480249) = 88 / 162
74 [runs scored]= 42261912
[runs scored]2 = 571106.919
[runs scored] = 755.716

The 2015 Orioles actually scored 713 runs, so in order to have won 88 games, they would have had to have scored 43 more runs. Essentially, they were just as / only as good at scoring runs as they were at preventing runs.


The 2015 Orioles finished at .500 for two reasons. The minor reason is that they underperformed, slightly - their runs scored and runs allowed were such that they could have expected to finish 83-79; they underperformed that by two games. The big reason was that they were an average team in scoring runs and an average team in preventing runs. It shouldn't be surprising that they were, overall, an average team.


Anonymous said...

This isn't particularly surprising to me, good to see the evidence to back it up though. I have a few questions, curious about your guys thoughts.
1. Could Ben Zobrist & Daniel Murphy match (from a WAR perspective) Davis's production? I don't mean power, of course, though one could make an argument that the two could approach 20HR each hitting LH in OPACY
2. What would Cahill and Hill cost? I just watched every Hill start from last year, and man, he was impressive. He made his rounds dominating every team in the AL East. Fangraphs did an interesting article on him. How I see it, he has a chance to start next year. If not, it seems like he could be quite a good LOOGY, better than Matusz probably. Sign me up if the price is right (1-3 MM?). I noticed Jon tweeted something about Cahill's success in the playoffs probably being driven by higher velo out of the bullpen. Fair enough. But, again, if he could be got for 2-4MM, wouldn't that be a smart investment considering that he would have a chance to start.
When it comes down to it, I would much rather have Cahill (2-4MM) and Hill (1-3MM) than Gonzalez (4.9MM) and Matusz (3.4MM)
Can anyone present a compelling argument for why the Orioles woudln't non-tender the latter two in favor of the former two? Groundball rate is also (easily) on the side of the (Ca)Hills.

Anonymous said...

Also, I've been looking at interesting comparison's between four 23yr old 1B/OF prospects playing in the same leagues in 2015
Greg Bird (NYY)
MLB: .261/.343/.529
AAA: .301/.353/.500
AA: .258/.358/.445
Josh Bell (PIT):
AAA: .347/.441/.504
AA: .307/.376/.427
Trey Mancini (BAL):
AA: .359/.395/.586
A+: .314/.341/.527
Aaron Judge (NYY):
AAA: .224/.308/.373
AA: .284/.350/.516
I know the disentangling of minor league stats, the causes of different results difficult, if not impossible, to disseminate.
First of all, according to Baseball America, the Eastern League was historically good this year. This would provide some level of explanation for why Bird and Bell improved as they ascended. Yet, while it is feasible that Eastern League pitching was better than International League pitching, it isn't feasible that EL pitching is better than MLB pitching. Bird's stats are strange in that way.
Mancini, the case I know best, attributes his improvements at Bowie to a ST approach that he wasn't comfortable with for his first month in Frederick (before he caught fire for the rest of the year).
Judge is also an interesting case, ironically because his case makes the most 'sense'. Poor strikeout rate in the EL but enough power to compensate. With the large IL ballparks and more veteran pitchers, the holes in his swing were exposed and fewer HRs hit.
I'm fascinated to see where each of these players start in 2016, and how they each perform. One would tentatively expect to see: Bird (MLB), Mancini (AAA), Bell (AAA), and Judge (AAA). I think Judge is the least likely to see significant time with the Yankees. It would be surprising if Bird spends much time in the minors. Mancini and Bell statuses mostly depend on what kind of offseason 1B signings their respective teams make, and how they perform in spring training.
Any thoughts on this? Interestingly enough, one of the Baseball America guys told me that he would consider Mancini to be a safer bet and Bell to have more upside. I found this surprising-- I would've thought it to be the other way around.

Anonymous said...

Cahill is just "NO". Even ATL dumped him when he was cheaply dumped on them by AZ. He was horrific. Bud Norris was better. If he somehow luckily performed well for a short period then that's what it was - luck. Rich Hill on the other hand was tremendous. I can't imagine BOS not keeping him especially as he is from there and he may only have one more good year in him.

Anonymous said...

So there we are. Either we can add 40 runs on offense or subtract 40 runs on defense. Is a pitching run worth the same as a hitting run? Which makes the most sense to address. How many runs could Mancini/Walker add to the team we have right now? Say by subtracting Parra. How many runs can Bundy/Wilson/Gausman/Drake/Cabral save in a full season? Say by subtracting Matusz and/or Gonzalez? People are down on Paredes but he did have a spectacular first half. He was a darn good DH when the Orioles were in 1st place. Thinking about just the 1st half, how many runs do you lose by not playing Paredes next year? You were down on Estrada and Happ because they outperform their peripherals but I see them as improvements over Miguel.

Unknown said...

Or a combination - add some runs on offense and subtract some on defense

Unknown said...

Or a combination - add some runs on offense and subtract some on defense