If you saw Monday night's game with the Texas Rangers--and since the contest was televised nationally by ESPN, I did--you saw a capsule summary of the Orioles' play of late.  The pitching staff did an adequate job, holding one of the top performing offensive teams in all of baseball to four runs, one of which was unearned.  The offense, unfortunately, managed to score only twice.  In the first four innings of the game, the Orioles had nine baserunners and scored none of them.  For the evening, the Birds managed only one hit with runners in scoring position, and that was a weak groundball that went for an infield single, scoring no one.  As a result, the Orioles lost another winnable game.  

Since the All-Star Break, Baltimore has posted a 3-8 record.  In those 11 games, the Orioles received a truly poor performance from a starting pitcher only once.  A team putting up decent offensive numbers during this time would probably have managed a record of 7-4.  A hot offensive club would probably have gone 9-2 or 10-1.

It has recently been demonstrated that, until very, very recently, the Orioles have been creating plenty of runs.  However, anyone who's been paying attention can clearly see that the team hasn't been scoring very many runs on a regular basis for the better part of three months.  In the end, after all, it's how many runs a team scores, not how many it creates, that has a direct impact on winning and losing baseball games.

So, what the hell is going on?

Let's take a look.

I took a look at R/RC comparisons for the entire modern (post 1900) history of major league baseball and found the following:

1) On average, teams create more runs than they score.  The historical mean difference between runs and runs scored is -18.4.  In other words, the average major league team since 1900 has scored 18.4 runs fewer than it has created.  Why is this?  Well, one possible explanation is that the runs created formula isn't particularly accurate as a predictor for runs scored on a team level.   The runs created method, it should be noted, was developed as a measure of individual, not club, performance.  As an addendum to that point, it's worth noting that there are a number of baserunner-related elements of the game that aren't reflected in the runs created formula--outs made while running the bases (and extra bases taken or not taken) aren't part of the equation; neither are errors, wild pitches and passed balls, pickoffs (caught stealing is part of the measure) or balks.  So, runs created is an imperfect formula when assessing team offense.  But it's still a measure that one can work with, as long as allowances are made and the measure is adjusted.

2) The standard deviation for the R/RC discrepancy is 4.24.  

3)  The comparable mean and SD for the population when figured as R/RC discrepancy per game:  -0.12 and .03 respectively.  

4)  The largest negative discrepancy in history is -133 by the 2001 San Francisco Giants.  The Giants scored 799 runs that year while creating 932.  Only five teams in history have had triple digit negative discrepancies.  (Interestingly, the 2002 edition of the Giants barely missed being the sixth (-96).)  The other four:  the 1929 Philadelphia Phillies (-113), the 1937 St. Louis Browns (-110), the 2002 Phillies (-106) and, incredibly, the 1927 New York Yankees (-105).  (The '27 Yankees scored 976 runs in 155 games, but they created 1081 runs.  Adjusted for the context of the era, the '27 Yankees are probably the top sabermetric offensive club of all-time.)  The largest negative discrepancy compiled by an Orioles team was the -62 (-.40) of the 1954 version. 

5)  There have been no triple digit positive R/RC discrepancies.  The top mark of all-time was turned in by the 1914 Philadelphia Athletics, who scored 90 more runs than the created.  Only three other teams have had a positive discrepancy of at least 60 runs:  the 1915 Chicago White Sox (+79), the 1985 White Sox (+68) and the 1907 White Sox (+60).  Only nine teams in history (including the above four) have managed to score at least 1/3 of a run per game more than they have created.  The best positive mark by an Orioles club was +27 (+.17 per game) by the 1989 edition.

Well, that's all very interesting, but what does any of it have to do with the 2005 Orioles?

Here's a look at the American League, sorted by runs per game, through July 25:

Note that the Orioles rank seventh in the league in runs per game.  The AL average is 4.80 runs per contest, so the Orioles are now below the American League average in scoring.  Note that, despite the seventh place ranking in runs, the Orioles are fourth in the league in OPS--far ahead of the fifth place Blue Jays and only 7 points behind third place Texas.  That incongruity bears further investigation (see below).  

Since May 6, the Orioles have scored 314 runs in 71 games, an average of 4.42 runs per contest.  During that same stretch of time, the league as a whole has averaged 4.87 runs per game.  For the last 11-odd weeks, the Orioles have been scoring far below the league average.  Take out the first month or so of the season and the Orioles would rank 12th in the American League in scoring.  Obviously this team has a problem with offense.

But remember the discrepancy between the runs scored and OPS rankings as shown in the table above?  Consider the following:

In terms of runs created per 27 outs, the Orioles are a solid fourth in the American League.  But their runs scored ranking is much worse.  Let's take a look at the two measures side-by-side, in the same table, through games of July 24.

The above table, sorted alphabetically by team, shows that the Orioles have, easily, the largest negative R-RC discrepancy in the American League, as the earlier analysis hinted.  The current rate would yield a raw negative R-RC disparity of roughly 55 runs at year's end.  This would not be among the very largest of all-time as noted above, but it would be in the 90th percentile and it would rank with the second largest negative discrepancy (-54 in 1999) in the franchise's history since it moved to Baltimore prior to the 1954 season. 

What is causing this?  Why is the shortfall so great?

An examination of a couple of simple splits tells the tale.  Here are the aggregate American League batting statistics this season with the bases empty:

As you can see, the Orioles are a hair's breadth away from being the best offensive team in the American League with the bases empty.  In fact, the only thing that Boston is doing better than the Orioles with the bases empty is walking.  Only Texas has hit more home runs with the bases empty than the Orioles; no team has a higher slugging percentage (or is even close for that matter), no team has a higher OPS or a higher batting average.

There's obviously a "but" coming and it's a doozy.  Here are the comparable offensive numbers with runners on base:

Clearly, there's a problem here.  The Orioles have slipped to a rank of 11th in RC27 with runners on base, and are closer to 12th place than they are to 10th.  Only two  teams have actually scored fewer runs with men on base than the Orioles (Cleveland, nine fewer, and Seattle, one fewer).

Consider the following table, which reveals the disparity in stark relief:

RC27E - runs created per 27 outs with the bases empty
RC27R - runs created per 27 outs with runners on base
OPSE - On Base Plus Slugging with the bases empty
OPSR - On Base Plus Slugging with runners on base

Interepreting the above table, the Orioles are creating 1.87 fewer runs with men on base than they are with the bases empty and have an OPS mark 43 points lower with runners on base than with the bases empty.  Both numbers are the largest in their categories in the league, by a wide margin.  While there are some aspects of the runs created formula that would tend to depress RC marks with runners on base (it's impossible, for instance, to hit into a double play or have a runner caught stealing with the bases empty), and in fact nine of the league's 14 clubs create more runs with the bases empty than with men on base (the league average disparity is approximately 0.43 fewer runs created with runners on base), the Orioles 1.87 RC27 difference in favor of the bases empty split is simply breathtaking in its scope.

OPS has no built-in biases.  Given the fact that:  a) pitchers pitch from the stretch rather than the wind-up the vast majority of the time that runners are on base; and b) defenses are frequently not in their desired position with runners on base, one would assume that, at the very least, batting averages would go up with men on base compared to situations when the bases are empty.  Since batting averages are a tangible subcomponent of OPS (BA is a part of both OBP and SLG, the building blocks of OPS), one would certainly assume that OPS would be higher with runners on base than with the bases empty.  And, in fact, that is the case.  Only three AL teams have lower OPS marks with runners on base than with the bases empty and the average difference is roughly 43 points in favor of the runners on base split.

The Orioles, however, show precisely the opposite effect, and are 43 points of OPS worse with runners on base than with the bases empty.  Subtract the Orioles' performance from the rest of the league and the difference balloons to approximately 50 points of OPS in favor of the runners on base situation.  This means that the Orioles are between 90 and 100 points away from from the average AL team (sans Orioles) in terms of typical performance, a huge disparity.

I think it's fairly obvious, then, what the Orioles' problem is.  What's not so obvious is what to do about it.  The easy--and in my view, extraordinarily superficial--answer is that the Orioles are simply a bad clutch team, but that ignores two important points.  First, I've never seen any evidence that a true "clutch" effect exists, at the individual or team level.  Second, this really isn't a "clutch" analysis.  The meaningful delineation here is simply runners on base versus bases empty.  There are plenty of "clutch" situations that would fit into the latter category and lots of non-clutch scenarios that would be properly placed under the former expression.

For some reason, the Orioles' offense has performed far worse with runners on base than with the bases empty.  Double play grounders are a part of the problem.  The Orioles have grounded into 99 double plays on the season, more than any other team in the league (Oakland is next with 91), despite the fact that they are nowhere near the top of the league in on base percentage (the O's rank sixth in OBP).  So, this isn't a function of leading the league in GIDP because the team leads the league in potential DP opportunities.  No, the Orioles are simply highly prone to grounding into double plays.  But that can't come anywhere near explaining the entirety (or even the majority) of the problem.

Frankly, I'm not sure what the root cause of the problem is, or what to recommend doing in an attempt to solve it.  The only thing I am sure about is that if the Orioles don't find a way to reconcile this conundrum their proverbial goose is cooked.

Discuss this article on the Birds in the Belfry Message Board

Back to Birds in the Belfry