On Sunday, an opponent of the St. Louis Cardinals scored, which was starting to seem like a lost art. Between a Todd Frazier RBI groundout in the fourth inning the previous Wednesday and Elian Herrera’s homer in the third inning Sunday, Cardinals pitchers went 38 innings without allowing a run, the franchise’s second-longest scoreless streak of the divisional era. On the strength of that near-perfect run prevention, St. Louis won four consecutive games, padding baseball’s best record, which now stands at 73-40, five games better than any other team’s.

By any estimation, the Cardinals pitched well during their string of shutout innings. But they didn’t pitch perfectly: Eleven Reds or Brewers reached scoring position, and between them, the teams left 29 runners on base. It’s not that there were no rallies; it’s that no rallies led to runs. Any jam that a Cardinals pitcher created, he (or a helpful reliever) almost immediately escaped, dressing every mistake with a soothing balm of strikeouts and well-placed batted balls.

The Cardinals have held hitters down all season. I’ll cite some new-school stats in a second, but let’s start with one of the oldest and most maligned: earned run average. ERA is a Frankenstein stat, an awkward synthesis of objective events and debatable rulings that isn’t adjusted for era or ballpark and doesn’t predict future results well. Flawed or not, though, it captures the staff’s outlier status to date: The Cardinals’ ERA is 2.60, almost six-tenths of a run lower than the second-place Mets’. Speaking of Mets, Matt Harvey’s ERA is 2.61: Imagine a team that always has Harvey pitching, on full rest and without any worries about workload, and you’ll get some sense of how stingy St. Louis’s pitching has been. The last time a team maintained an ERA as low as the Cardinals’ over a full season was 1972, the last pre-DH year, when teams averaged 3.69 runs per game, the second-lowest total since the dead ball era. And the Cardinals have done this despite losing ace Adam Wainwright to a season-ending Achilles injury on April 25.

Team-level stats that estimate run prevention from pitchers’ peripherals aren’t as kind to the Cardinals. St. Louis has the best FIP, but only barely: The Cardinals’ FIP is 0.70 runs higher than their ERA, tied for the fifth-largest difference in that direction since World War II. More complex run estimators — xFIP, SIERA, DRA — are even less impressed, placing the Cardinals inside the top 10 but outside the top five. The Cardinals don’t strike out batters more often than everyone else, or get grounders more often, or allow fewer walks. They haven’t acted like a team with a surplus of pitching, acquiring two former closers at the trading deadline. And it’s not as if St. Louis has a Royals-esque defense propping up its staff: Its fielders are good, but not great.

But there is one thing the Cardinals staff has done better than anyone else’s: strand baserunners. Cardinals pitchers have a collective 81.5 percent left-on-base percentage, a rate that hasn’t been seen since — well, ever, actually.

Again, we’ll get to the fancy stats, but let’s start with something simpler: slash lines. Leaguewide, pitchers have allowed an OPS about 40 points higher with men on and in scoring position than they have with the bases empty. The Cardinals have turned that trend on its head, then tackled it, kicked it into unconsciousness, and buried it alive.

With the bases empty, the Cardinals have held hitters to a league-average line. Once runners have reached base, though, they’ve clamped down, limiting them to a sub-.600 OPS. With runners in scoring position, St. Louis has been even more miserly. With men on second and/or third, the Cardinals’ opponents have hit like Billy Hamilton, minus the stolen-base ability. So, Cardinals pitchers are like Matt Harvey, and Cardinals opponents with runners in scoring position are like Billy Hamilton without blazing speed. You could call that a mismatch.

The Cardinals’ success with runners in scoring position is a convincing illustration of the concept of “cluster luck,” first described by author Joe Peta and calculated daily by Grantland contributor Ed Feng. The crux of the concept: Given the same number of hits — and even extra-base hits — two teams’ run totals can vary dramatically depending on the way those hits are distributed. A team that bunches its hits in close succession will score more runs than a team that parcels them out sparingly from inning to inning, as the Reds and Brewers did against St. Louis last week. This seems like something that should even out over the course of a season, and often it does, but it hasn’t come close to equalizing in the Cardinals’ case. According to Feng, the Cardinals’ crackdown with runners in scoring position has saved them 84 runs relative to the total their underlying stats say they should have allowed, which would outstrip the all-time second-place team in Feng’s records (which go back to 2001) by more than 16 runs. BaseRuns, another method of calculating the number of runs a team “deserved” to allow, yields a slightly more conservative estimate of 61 runs below the team’s expected tab. Either way, it’s an enormous, potentially playoff-race-altering number.

I’m not the first to notice something strange in the Cardinals’ stats. Early on, though, it was easier to dismiss as a small-sample fluke that would regress over the rest of the season. But as the sample has grown larger, the split has become even more obvious. The Cardinals’ strand rate has been even higher in the second half of the season than it was before the break, defying the usual tendency of extreme performances to become more conventional over time.

This isn’t the first time in recent memory that high-leverage heroics have helped the Cardinals. In 2013, when the Cardinals staff strung together a post-World War II franchise-record 39 consecutive scoreless innings, it was actually the offense that benefited the most from good fortune. That season, Cardinals batters produced a .653 OPS with the bases empty but upped that figure to .839 with men on and .865 with runners in scoring position, thanks to a .330/.402/.463 line over 1,621 RISP plate appearances. At the time, it was temping to “post hoc, ergo propter hoc” that performance and give credit to the Cardinals Way, but in 2014, the Cardinals hit almost exactly as well with the bases empty as they did with men on. The old Devil Magic didn’t last, and the current incarnation will probably prove just as fleeting.

Before we dismiss the Cardinals’ 2015 sequencing as completely random, though, we should do our due diligence to determine why they have been better in crucial situations.

With runners in scoring position, the Cardinals have struck out batters more often than any other team. But most of the split stems from a 75-point gap in BABIP, which in turn is attributable to a difference in quality of contact. With runners in scoring position, St. Louis has allowed fewer hard batted balls, made evident by a lower line drive rate, a higher percentage of popups, and (in part because of the popups) fewer fly balls going over the fence.

Really, though, we’re just kicking the question down the road. The Cardinals have been better with runners in scoring position because they’ve allowed weaker contact. So where did the weaker contact come from? There are several possible explanations:

They’re Clutch

Cardinals pitchers have, unquestionably, been clutch. There’s a stat for that, and, unsurprisingly, St. Louis is leading the league. While it would be accurate to call the Cardinals’ past performance clutch, though, it’s another thing to expect their clutchness to continue. Over the past five seasons, the year-to-year correlation of pitching Clutch scores is 0.09, where 1.0 is a perfect one-to-one relationship and 0.0 indicates no connection. In other words, there’s little evidence that clutchitude is consistent across seasons, at least on a leaguewide level, although we can’t prove that it is or isn’t present on any particular team. Even if the Cardinals were inherently superior in pressure situations, though, that ability would have to manifest itself in some other stats.

They Turn Two

The Cardinals still get ground balls, in accordance with the “keep the ball down” philosophy carried over from their Dave Duncan years. St. Louis ranks fourth in overall ground ball rate and fifth with runners in scoring position. In theory, more ground balls should translate to better odds of salvation in double-play situations. In practice, things haven’t worked out that way. St. Louis ranks 15th and 23rd in DRS and UZR Double Plays Runs, respectively, and 22nd in Double Play Percentage, turning two in 10.5 percent of its opportunities (slightly below league average).

They Don’t Hold Anything Back

It could be that the Cardinals, through some combination of coaching and character, are able to dig deeper — and throw harder — than anyone else in the moments that matter most. With the bases empty, the Cardinals have averaged 92.6 mph on their two- and four-seamers. With runners in scoring position, they’ve pumped up the average to 92.9. Supporting evidence!

The problem: This phenomenon isn’t limited to St. Louis. All pitchers tend to reach back for a little extra when they’re in trouble: Leaguewide, teams are averaging 92.1 with the bases empty and 92.5 with runners in scoring position, so if anything, St. Louis shows less of a speed differential than the typical team. Theory rejected.

They Optimize Their Pitch Selection and Location

If there’s one thing that’s different about batters with runners in scoring position, it’s their eagerness to swing. With the bases empty, hitters have swung at 46.0 percent of pitches and 25.5 percent of first pitches this season. With runners in scoring position, primed to provide an RBI endorphin rush to any hitter who drives them in, those rates rise to 48.7 and 33.5, respectively.

Early last season, my colleague Jonah Keri noted that under Duncan, the Cardinals tried to take advantage of their opponents’ eagerness by throwing fastballs and sinkers on the first pitch with men on second and/or third much less often than normal, relative to their overall fastball rate. However, that tendency no longer holds true.

The graph below shows the difference between the rate of four-seamers plus sinkers on the first pitches of plate appearances with the bases empty and runners in scoring position, both for the Cardinals and for the other 29 teams. From 2008 to 2011, the Cardinals threw fastballs far less often than the rest of the league on first pitches with runners in scoring position, compared to their bases-empty baseline. After 2011, though — perhaps not coincidentally, the last year of Duncan’s and Tony La Russa’s tenure with the team — they became much more willing to start hitters off with hard stuff, to the point that they now do so much more often than the typical team.

None of the Cardinals’ pitch-usage rates changes dramatically with and without runners in scoring position, with the biggest difference being a slight uptick in slider rate. Location-wise, all that stands out is that the Cardinals decrease their percentage of pitches in the strike zone with runners in scoring position slightly more than the league as a whole, maybe because of their confidence in Yadier Molina’s blocking ability.

They Keep Runners Glued to Their Bags

Molina might not be the hitter or framer he once was, but he still has a strong arm. The Cardinals lead the NL in Takeoff Rate Above Average, a Baseball Prospectus stat that measures a team’s impact on the likelihood that opposing runners will try to steal. Both the Cardinals’ catchers and pitchers also rank above average in Swipe Rate Above Average, which measures their impact on the likelihood that a given steal attempt will succeed. The result: an MLB-low 37 stolen bases allowed, at a lousy 64 percent success rate. Discouraging steals might also translate to smaller leads on the basepaths, which would mean fewer runners scoring from second on singles, although according to BP’s Baserunning Runs, the Cardinals have been about average at restricting the running game overall.

It’s possible that some of what we’re seeing is the residue of design. For instance, we know that the Cardinals have done well in the draft, and while draft success is largely dependent on luck, we also know that St. Louis (at least for a time) was evaluating amateur prospects in a way that other clubs weren’t, which makes it more plausible that there was at least some signal amid the noise. If there’s a similar method to the madness on the mound that we’ve seen this season, it’s tougher to detect. Even if St. Louis has discovered the secret to RISP success, it’s unlikely the edge is this large. But even if the pitching staff’s incredible clutchness doesn’t continue, the wins it’s helped St. Louis secure could help the Cardinals hold off the Pirates (who’ve enjoyed a more modest form of cluster luck) and the Cubs (who haven’t) down the stretch. Unsustainable success with runners in scoring position, combined with a quality roster, helped propel the Cardinals to a division title over Pittsburgh and a pennant in 2013. Don’t be surprised if this race resolves in a similar way.

Thanks to Rob McQuown of Baseball Prospectus for research assistance.