Too Close for Comfort

In my series preview on Monday, I brought up this interesting fact:

Excluding the blowout loss to the Athletics on April 10, every single game the Mariners have played has been decided by three runs or less.

That trend has continued over the first two games of this series. The closer the game, the smaller the margin for error is and the Mariners have been walking that fine line all year long. When games are being decided by one or two runs, the outcome of the game gets closer to a coin flip. Success in these close games isn’t necessarily a skill but it can’t only be attributed to "luck"—the reality lies somewhere in between. Let’s take a deeper look into the Mariners’ performance in these situations to see if we can find anything of note.

Last year, the Mariners’ record in games decided by three runs or less was 56-54, the seventh best record in the American League. That also means the Mariners’ win percentage in games decided by four runs or more was a whopping .596! Like I stated above, every single game the Mariners have played this year, except one, has been decided by three runs or less which means they’re 9-10 in those games.

Because of all of these close games, the Mariners lead the league in high leverage at-bats and high leverage innings pitched. Their offense has actually performed better in high leverage situations than would be expected in a context-neutral environment. In other words, the Mariners have benefited from some clutch hitting this year.

They’ve been able produce some clutch hits despite being mediocre when it comes to run sequencing. Using Batting Runs (the context-neutral hitting component of WAR) and RE24 (a metric that shows how efficient a team has sequenced their hits), we can see which teams are out performing their context-neutral performance. This year, the Mariners have scored 6.4 more runs than we would have expected given their context neutral performance. That may sound good but that mark ranks just eighth in the AL.

What about the Mariners’ bullpen? They lead the league in high leverage innings and it isn’t even close (the average Leverage Index for our relievers is 1.49, the next highest is just 1.15). In these situations, the bullpen has performed worse than would be expected in a context-neutral environment. The bullpen’s FIP is just 3.32, the fifth best mark in the league. That may give you the impression that the bullpen’s performance is bound to regress towards the mean, that they’ll eventually start performing better in high leverage situations. That may not be the case.

Despite running a decent FIP, much of that is driven by an incredibly low home run rate. In fact, the Mariners’ bullpen leads the league in walk rate. If you remove home runs from the equation by using xFIP, the Mariners’ bullpen is 12^th in the league. On top of it all, the Mariners’ defensive efficiency (as measured by opponent BABIP) has been pretty terrible as well. The bullpen’s opposing BABIP is a whopping .313, second worst in the league. That’s not just a function of opposing batters getting a few lucky hits, it’s also tied to our lousy fielding so far this year.

There is very little correlation from year to year in team performance in close games. The randomness of the game pushes most teams towards a .500 record in close games and there are few outliers. The Mariners’ offense has performed well in high leverage situations so far but that’s no guarantee of future clutch performance. More worrying is the performance of our bullpen. The surface level stats don’t tell the whole picture. Giving up a huge amount of walks while pitching in front of a poor defense is not a recipe for success.

Last year, the Mariners we’re very efficient defensively and the bullpen wasn’t walking nearly as many batters. The Mariners probably won’t see much improvement on the defensive front, so the bullpen will have to get their walks under control before we see an improvement. The bullpen was also able to strand 80% of their runners last year, a mark that lead the Majors by a wide margin. This year, the bullpen’s strand rate is well below that mark and is slightly below average. I’m actually surprised the bullpen hasn’t seen more meltdowns and we may see things get even worse if something doesn’t give.