One way that analysts determine whether a team has overachieved is by looking at their run differential and determining their pythagorean expectation. If a team is above .500 but has allowed more runs than it has scored then it safe to presume that a given team has been lucky. If so, it should be possible to use runs scored and runs allowed to determine whether a team has been "lucky" in a specific type of game. For example, does a team win a large percentage of games where they only scores 2 runs?
I decided to build a dataset (accurate as of 8/3/2014) consisting of the 2014 results (W/L, Runs Scored and Runs Allowed) for each game played by the Orioles, Blue Jays, Yankees and Rays. Then I determined how many wins or losses they had when they scored a given number of runs or allowed a given number of runs. This should show us whether a given team excels or struggles in certain types of games.
Using Retrosheet data, I was able to build a control dataset with the results for every team from 2011 to 2013. The Orioles/Blue Jays/Yankees/Rays have won 52.5% of their games so I built a factor into the dataset to account for this. Retrosheet won't release data for 2014 until after the season. I include control percentages where relevant but I'm not sure whether they're helpful for analyzing 2014 results.
This first chart shows how each of the four teams has performed based on runs scored.
This chart shows that these four contenders lose most of their games when they score two or fewer runs, win about half of their games when they score three runs and win most of their games when they score four or more runs. The exception to this rule is Tampa Bay which struggles when they score 3 or 4 runs.
The next chart shows how these teams have performed based on runs allowed.
This chart shows that teams win most of their games when allowing 2 or fewer runs. Toronto, Baltimore and the Yankees win a large number of their games when allowing 3 runs while the Rays have won fewer than half of theirs. These teams win about half of their games when allowing four runs.
It is possible to form groups using these results. This is how the numbers looked grouped based on runs scored. HIS stands for historical results.
The Orioles have either gotten lucky or have shown an ability to win games even when their offense doesn't score. This potentially could be because of the fact that Britton, O'Day and now Miller have been dominant so far this season. It makes sense that a team with a dominant bullpen could win a larger percentage of low scoring games than the average club. While the Orioles have had disappointing results when scoring 5 or 6 runs compared to the 2011-2013 historical data, they have had good results compared to the other three 2014 teams in the sample. The Orioles have struggled when scoring 7 or more runs and should probably have gone 21-2 instead of 20-3 in those games.
The Blue Jays have either been unlucky or have shown an inability to win games when their offense doesn't score 4 or more runs. If they won the expected amount of games when scoring three or fewer runs then they would have eight more wins and would be in first place. This could indicate that they are stronger than they appear or it could indicate that they have a fatal flaw. They have slightly over performed when scoring 4 or more runs.
The Yankees have been about as successful in each of these categories as one should expect.
The Rays have won a larger than expected number of games when scoring between 0 and 2 runs but a lot fewer than expected when scoring between 3 and 4 runs. This was mentioned above and still remains true. If they had performed as expected then they would five more wins. While they would still be in fourth place their position would be much stronger then it is at the present.
This is the same chart but for runs allowed.
The Orioles have underachieved when they have allowed 0 to 2 runs. We've been shut out eight times this season and three were 1-0 losses. This is especially confusing because we've won a higher than expected amount of games when our offense has scored between 0 to 2 runs. We've overachieved when allowing 5 to 6 runs. Luck is another possible explanation.
The Blue Jays have won more games than expected when allowing 2 or fewer runs. This indicates that their offense is pretty powerful and can win games given a strong pitching performance.
The Yankees have underachieved slightly when they have allowed 0 to 2 runs. But they've really struggled when allowing 5 to 6 runs. While they've overachieved when allowing 7+ runs this indicates that they don't have a very powerful offense and are unable to win if their pitching gets blown out.
The Rays have struggled when allowing 3 to 6 runs. While they haven't struggled when allowing 7+ runs this is because teams nearly always lose when allowing 7+ runs. The Rays have had 40 games where they've scored between 0 to 2 runs or about six more then average and 23 games where they've scored between 3 to 4 runs or ten fewer than average. This may have something to do with their problems.
I'm not sure whether this analysis actually is meaningful or whether it is just measuring noise. If it is meaningful then it indicates that the Orioles have had reasonably predictable results this season and that our pitching can win despite some poor offensive performances but that our offense has lost some very winable games. It indicates that the Blue Jays have had major struggles when scoring between 2 and 4 runs and that they likely are fatally flawed and need a significant amount of pitching help. It indicates that the Yankees offense is unable to win when their pitching gives up a large number of runs although this is almost definitely noise. Finally, it indicates that the Rays struggle when they don't receive strong pitching performances.
It will be interesting to see whether these trends continue for the second half of the season.