Three-and-uh-oh! How Astros pitchers stared down three terrifying 3-0 counts

For theAstrosto come back and even up the American League Championship Series at 3-3 on Friday night, two Houston pitchers had to come back from three terrifying 3-0 counts against very good hitters in very big situations. They each did, and it's worth looking at how important those moments were against theNew York Yankees.

First, Justin Verlander was facingGary Sanchez with a three-run lead, two men on and two outs in the sixth inning; his Astros were roughly 88 percent likely to win, according to FanGraphs' win expectancies. Verlander's first pitch -- his 73rd of the night -- came in at 98 mph, the fourth-hardest pitch he had thrown all night; the second, at 97 mph, was 10th fastest, and the third, again at 98 mph, was sixth, according to PITCHf/x. All three were balls, though. The Yankees' cleanup hitter was way ahead in the count.

On FanGraphs, the win expectancy hadn't budged, but in actuality, the Yankees had just gotten a pretty big boost. Across the majors, hitters are close to twice as destructive after reaching a 3-0 count. The league's wOBA -- an all-in-one offensive metric that is scaled to on-base percentage and can be converted into runs -- is about .330. After 3-0 counts, though, the league's wOBA is .549.

A huge part of that damage comes from walks, of course; nearly two-thirds of batters who reach 3-0 eventually will walk. If Sanchez drew that walk, the bases would have been loaded for the dangerous left-handed hitter Greg Bird. Even ignoring Bird's talent, the win-expectancy tables would boost the Yankees by about 4 percentage points.

But it's not just the walks. Batters hit the ball much harder after reaching 3-0, able as they are to zero in on their pitch and swing big, knowing the cost of a single strike is relatively small. Batting average on balls put in play goes up 15 points after 3-0, while home runs per ball in play go up almost 50 percent.

Put all these great outcomes together and just by working to a 3-0 count, Sanchez had increased the Yankees' chances to win by about 4 percent. In other words, just by getting to the 3-0 count, Sanchez had already improved the Yankees' chances by about as much as a walk itself would. The odds were that he would walk, but if he didn't, the odds were that he would hit the ball, and hit it hard. He had greatly increased his chances of tying the game with one mighty swing.

Instead, Sanchez got fooled by a slider in the strike zone, an audacious pitch selection in the moment. This is what the at-bat looked like to poor Gary Sanchez:

Sanchez started to swing, then saw that it wasn't the fastball he'd hoped to get and tried to halt his attempt. Rather than missing the pitch for a strike one, he barely tapped the ball, on a checked-swing, for an easy 6-3 putout.

It was, according to Statcast, the 11th-weakest 3-0 contact by any hitter this year, at just 60.9 mph. About one in 10 grounders hit at 60 mph goes for a base hit; Sanchez is not the sort of runner who can represent the one in 10. It was an extraordinarily weak outcome in an extraordinary moment.

So Verlander escaped the sixth, only to put the first two hitters on in the seventh. Facing Aaron Hicks -- a switch-hitter -- Verlander again fell behind 3-0, this time on an extremely close first-pitch curveball and then a pair of wayward fastballs. The Yankees had again, for the moment, improved their chances of winning without making a mark on the box score: When Hicks stepped in, they were 21 percent likely to win; but with the 3-0 count on the very good hitter in a very high-leverage situation those odds had grown to about 27 percent, and changing every 25 seconds.

Hicks, unlike Sanchez, rarely swings at 3-0 -- only twice in his career, in fact. He took a 96 mph fastball down the middle for the first strike. Verlander got a generous call on a fastball that was at the bottom of the strike zone but, apparently, off the plate away, for strike two.

The two reasons we gave for why 3-0 counts are so deadly -- lots of walks, lots of hard contact -- only cover two-thirds of the story. The final leg is that it takes a lot to strike a batter out after 3-0. About 7 percent of batters whiff after getting that far ahead; it's 21 percent at the start of an at-bat. But Verlander was two-thirds of the way against Hicks, which is when the at-bat got wild.

After working back to the full count, Verlander threw three fastballs on three different edges of the zone: 96 mph on the black away; 97 mph on the black inside; then 97 mph at the top edge of the strike zone.

He finally tried to surprise Hicks with a slider, but it was a slow and ugly one, curling in just enough to catch the outer part of the strike zone but not enough to fool Hicks, who fouled it off. Verlander's 10th pitch was perfect: A slider at the bottom lower corner of the zone, which Hicks missed by a literal foot:

Verlander is a great pitcher, but like all pitchers, he is vulnerable when he falls behind 3-0. But as his career has progressed, he has become more and more in control of those at-bats. Through 2012, batters had hit .299/.742/.425 against him after 3-0 counts; since then, even as his stuff overall has occasionally been diminished, he has held hitters to a .189/.618/.351 line after 3-0. In three-ball counts, in general, batters hit just .210 and slugged just .340 against him this year. He's far more likely now to throw something other than a fastball on three-balls -- as he did to Sanchez -- and he's far more likely now to throw pitches in the zone with three balls, as he did repeatedly to Hicks.

We don't know what would have happened had Hicks, say, walked. Maybe the next three batters all would have struck out; we'll never know. But the two perfectly normal outcomes that followed his strikeout show just how important an out there could figure: Todd Frazier flied out to deep-center field, and instead of being a sacrifice fly to drive in a run and move the other runners up, it was the second out of the inning. Then Chase Headley grounded out toJose Altuve's left, and instead of driving in a second run, it was the third out of the inning. (Of course, these things wouldn't have happened in the alternate reality, but for illustrative purposes ...) The Yankees' chances had dropped from about 27 percent when Hicks had the 3-0 count to 15 percent once he'd struck out, and to 5 percent when the inning ended.

The third crucial 3-0 count wasn't given by Verlander, the stakes weren't as high and the outcome was a bit more typical. After allowing a home run to Aaron Judge, Brad Peacock quickly fell behind 3-0 to Didi Gregorius with a two-run lead in the eighth. Sanchez was on deck, and Peacock was on the ropes. But he pumped in an easy strike, as pitchers do, then threw a well-located fastball down.

Sometimes after 3-0, all a pitcher can do is pump in fastballs, close his eyes and pray that the batter just misses it; that's what happened here, as Gregorius popped out for the second out of the inning. The extra 2 percentage points of win expectancy that the Yankees had briefly built up mid-AB had been ephemeral.

So it was for the Yankees: The moments that they seemed on the verge of something would vanish just as quickly. By the end, it wasn't very close at all.