A sport like baseball isn’t easy to design. It isn’t even easy to replicate. Take the 1984 Nintendo game Baseball, one of the many early attempts to render the sport onscreen. If the catcher threw the ball to second and a runner on third broke for home, he would always score: the speed of the ball thrown across the diamond and back couldn’t match the 20 mph of the baserunner heading ninety feet. The most famous Nintendo baseball game of the era, RBI Baseball, faced a similar set of troublesome physics, and solved it a different way: the AI never allowed the catcher to throw the ball if there was a runner who could take advantage. RBI Baseball tracked many stats during the game, a fun feature; defensive indifference was not one of them.

In 2016, there is still no baseball video game where you can do what Melvin Upton, Jr. did on June 3rd.

With the bases loaded and Drew Pomeranz at the plate, Upton took a walking lead against Rockies starter Chris Rusin, and then broke for home. Nolan Arenado, playing at normal depth, immediately recognized it and covered the bag for the fake. Everyone else was caught off guard, including the Padres runners, who failed to capitalize on the distraction by taking their own bases.

It caught everyone off guard because it was a terrible idea. To be fair to Upton, in order for a steal of home to be a good idea, it has to be a terrible idea; that’s the beauty of game theory. With the Padres ahead and the need for a single run obviated, we can go straight to the run expectancy tables. Bases loaded with one out has a run expectancy of 1.52. In order for Upton’s madness to be even a break-even proposition, he’d have to succeed at a rate that balanced the expectancy of failure (two outs, runners on first and second, or 0.343) against the results of success (0.908 run expectancy plus the one in the bank). The algebra states that Upton needs to steal home with about a 75 percent success rate. (RE24 is based on averages, so having Drew Pomeranz wielding the bat softens that number significantly.) The success rate of all of baseball, when it comes to situations where the catcher has to throw the ball to someone else first, is 69 percent.

So when Chris Rusin winds up to pitch, plotting the downfall of his opponents, a steal of home was not among his contingencies. By the time he throws the ball, Upton is more than halfway to the plate.

The rest is magic. The throw is clean, the catcher Nick Hundley receives it in time to lunge out and put his glove between Upton and home plate. But in a beautiful and improvised act, Upton slides not to the plate but to its left, contorting himself around the outstretched glove and slapping the plate with his left hand. In retrospect, the terrible idea turns out to be, in retrospect, the right one; Pomeranz strikes out and Jon Jay grounds out to first. But that doesn’t matter; that worldview projects us to the point of looking at an insurance run in a Padres-Rockies game on a planet still full of conflict and sorrow. Only those five seconds matter.

Stealing home is one of the rarest accomplishments in major-league baseball, and like inside-the-park home runs, they’re surprisingly difficult to track down. There is no independent stat, no fan-driven homemade leaderboard. They dominate the GIF viewership for a single news cycle and then get stuffed into the record books with all the other stolen bases. It’s a sad fate. The league is actually 4 for 11 stealing home this year, but that’s because baseball has a hard time knowing what to call a stolen base attempt. In reality a lot of different things can happen:

Francisco Cervelli, 4/3/16: The failed squeeze. What begins as a perfectly honest squeeze play is transformed, by a clumsy batter, into a mark of failed deviousness. This is another situation where our statistics aren’t refined enough; Cervelli’s march into No Man’s Land is recorded as a caught stealing, as would a pickoff throw that forces a runner into a hopeless rundown. We’re seeking something comparable to Upton’s daring, but in reality, the vast majority of our results are mere incompetence.

Matt Szczur, 5/2/16: The failed double steal. With two outs, a runner on first gets caught in a rundown; the runner on third, realizing that if he can sneak home while the fielders are enjoying their pickle, the run will count. Two of the four stolen bases of home were scored this way: Andres Blanco and Travis Shaw. In a strange outcome, Czszur is credited with the rare caught stealing/run scored, as the credit for his achievement is bestowed instead on Josh Harrison’s exuberance.

True Steals of home. We’ve had two this year: Upton and Jacoby Ellsbury, who did it on 4/22 and who didn’t even get that advantageous of a head start, nor as tricky a slide, but with pure, simple speed. Last year, there were none.

Stealing home is not a dying art: It is a dead art. Ty Cobb stole home 54 times in his career. Rickey Henderson did it four times, all of them in his first three seasons playing under Billy Martin. Eleven players have stolen home twice in a game; only Vic Power, in 1958, did it after desegregation. One of the enjoyable things about baseball is its adaptation, the delicate balance of strategy and game theory. But in this case, the defense didn’t just respond to the steal of home; they essentially solved it. Pitchers tightened up their deliveries; catchers and infielders caught on to the delayed double steal. (Sam Crawford, in The Glory of Their Times, told of how, when taking a walk with Cobb on third, the latter would signal him. In this case Crawford would near first and then immediately steal second, to try to draw a throw and allow Cobb to steal home. “Most of the time they were too paralyzed to do anything,” he notes, “and I’d wind up at second on a base on balls.”) Even beyond the defensive improvement, increased offense pushed the break-even point of success high enough to make the play not worth the risk.

It’s a shame, because from an aesthetic standpoint, the steal of home is one of the most exciting plays in the game. It’s not an act that happens in a blink of an eye, like the slider that dips below the bat of a hitter thinking fastball 3-1. It’s not the upper-deck shot, the display of raw power built over ten thousand weight room sessions but inflicted with the unconsciousness of muscle memory. It’s the purest form of true cunning, a mental battle played out in real time, a runner whose audacity is laid bare in front of every person in the crowd.

And without a moving walkway on the third base line, it won’t be coming back. There aren’t any obvious ways to tinker with the rules to allow us more of these. So all we can do is appreciate the hubris when we can.

Special thanks to Bryan Grosnick for hand-collating the data for steals of home in 2015-2016.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
You forgot Rod Carew, 1969. If the umpire hadn't blown the call on the eighth, he'd have the single season record. Ibid Billy Martin.
I don't think one can assume that Pomeranz strikes out and Jay flies out if Upton remains on third. This isn't Strat-O-Matic. All sorts of factors are different. (Or, as a Strat-playing friend once told me, "It's a different time-space continuum.)

Also, the outfielder's name is Matt Szczur.

Thanks for an interesting piece!
Sorry, meant that Jay grounded out.
Fixed, thank you. I should have left in my first draft of his name, Szzzczzzr.

These GIFs are amazing.