Why are the Jays terrible in close games this century?

It should come as no surprise to anyone who's paid much attention to the Jays this year that they've got an awful record in close games. With yesterday's 10th inning loss, the Blue Jays fell to 5-15 in games decided by one or two runs. They're a further 3-6 in games decided by three runs, which are reasonably close games. But when the margin is comfortable, the Jays have been great, with a 14-6 record in games decided by four or more runs.

This state of affairs, to put it mildly, is spectacularly inefficient and the main reason why the Jays are mired near the bottom of the standings five games below .500 despite a very good run differential. If the Jays were even .500 in those close games, they'd instead be five games above and narrowly leading the division. And they wouldn't need to play .600 baseball the rest of the way just to get to 90 wins and have a reasonable chance to make the wild card play-in (which I personally don't consider making the playoffs, but that's a matter for another day).

While there's been a lot of fingers pointed in terms of assigning culpability, mostly at the bullpen as well as the GM who constructed it and the skipper who manages it, the problems certainly go well beyond. Just in the last homestand, tight losses have featured critical defensive blunders, outs on the basepath, and lack of timely hitting. More broadly, I cannot help but be struck with a sickening sense of déjà vu.

In 2008, like in 2015, the Jays had a veteran team with a franchise-record payroll, with high expectations in year three of a contention window that had failed to bear fruit. Then, as now, the team got off to a slow start and had a solid run differential but very poor record in close games, at 16-24 in one- or two-run games when John Gibbons was fired in late June with a 35-39 record. Had they just split those 40 games, they've have been 39-35. On its own, that might not have been enough to make the playoffs, but it probably would not have cost Gibbons his job.

Further, even the past couple years they've had troubles in close games, and I began to wonder if Gibbons is the common thread and perhaps the problems. Of course, he had solid reputation in his first stint for tactical decisions, so maybe I was constructing a narrative that didn't exist. So I went to Baseball-Reference's seasonal game-by-game results, and pulled the data for both of Gibby's tenures. It turns out I wasn't imagining things:

Throughout Gibbon's managerial career, the Jays have consistently been horrible in close games, equally in both one and two run games. A .440 winning percentage translates to 71 wins over a full season, which is obviously very bad. They've been okay in three run games, but downright excellent in games decided by four or more. A .549 winning percentage gives 89 wins over a full season, which would mean perennially contending.

So what's behind this? The high watermark was in 2006, when the Jays went 33-30 in close games on the strength of a 20-10 record in one-run games. They had a great bullpen that year, anchored by B.J. Ryan's 1.37 ERA and supported with a deep cast of good options. Overall, the bullpen ranked 7th in MLB with +3.99 Win Probability Added. But the next best year was 2007, when the pen ranked 7th from the bottom with a negative WPA. And the pen was very good in 2013, when their record in close games was characteristically bad. So it can't all be the 'pen.

Another frequent suggestion is that the Jays are too reliant on the home run in recent years, making them prone to big explosions but long droughts and leaving them vulnerable to not being able to score when it counts. The problem I see with this is that they had the same problems in 2004-08 (albeit to a lesser extent) when they were primarily oriented to preventing runs, and not home run reliant at all (18th in MLB from 2005-07 in HR).

What about pure chance? Since baseball games have only two outcomes, it's easy to model using the binomial distribution. As it turns out, even in a sample of 455 (total close games), the expectation is that in a 30 team league, about one team would have as extreme a deviation from average as the Jays. So we can't strictly reject the possibility it's just bad luck, though it's very unlikely to have just been that.

So is it Gibbons? I thought the best way would be to compare his record with the managers who directly preceded and proceeded him. They would have had very similar teams to work with, at least initially. If they had significantly different results collectively, then that would point to Gibby. So I pulled the same data for Tosca, Cito and Farrell:

The others show the same clear pattern, if not quite as extreme. Conveniently, this covers the last 13 seasons continuously, and the Jays have been bad in close games with all managers, and pretty good otherwise. And with a much bigger sample size, it's well outside the realm of chance, as it would be roughly a three sigma event. How can a franchise, across so many managers, different GMs, different strategies, be shooting themselves in the foot consistently? I wanted to see it if went back further, so I went back to Cito's first tenure starting in 1989:

From 1989 to 2001, winning close games wasn't a problem, as all four managers were above .500 in close games. Basically, there's a clear demarcation line when Buck Martinez was fired:

In the 13 years before Buck was fired as manager, the Jays were .522 in close games and .519 in games decided by four or more. Basically the same. In the 13 years since, it's been lopsided as the differential at the bottom shows (adjusted to account for the Jays being better in 1989-2002 than 2002-15).

In conclusion, for the better part of this century, the Jays have been awful in close games, to their great detriment. It's probably cost them a playoff appearance or two just versus being .500, nevermind if they were good. And there's really no obvious reason why. Maybe the answer is simply to put ole John Albert back in the manager's seat; he did just get a five year contract so that's already in order.

After all, something must be done. And that would certainly be something...