Skewed Left: It’s Not a No-Hitter ’til _____

There were some years in there, before he got busy and I got Twitter, that my brother and I would have this arrangement. He would text me or I would text him “turn on the Phillies game” or “turn on the ESPN game.” The rest would go untexted, not because either of us was necessarily all that superstitious, but just because the rest was assumed. Somebody had a no-hitter going.

When to send the message, though? Too early, and you’re wasting somebody’s time. Not that baseball is a waste of time, but if we weren’t watching the game, we were probably doing something better. Too late, and you’re depriving him of some of the more exciting innings.

Some people actually have to decide this. Networks have to decide when to cut in or give special updates in a no-hitter. Writers have to decide when to start writing their specials; or worse, off-duty writers have to decide when to get in the car and come to the ballpark to help with the potential extra work.

Some have it decided for them. MLB.com beat writers are instructed to file running game stories on no-hitters if the no-hitter is through six innings. Associated Press news alerts come out at the same time and continue after every inning. But for most, it’s all up to us when to get excited (if, in this era of depressed offense and emphasis on out-avoidance rather than hits, you even get excited over no-hitters anymore).

How high do you want the chances to be before you’ll get yourself invested in a game you wouldn’t have ordinarily followed? Fifty-fifty shot he finishes it? Is 10 percent enough?

As it turns out, 10 percent is just about the threshold for institutions getting institutionally excited. In looking at every game played since 1950, if a pitcher (or a team, since we’re allowing combined efforts) is through six innings with a no-hitter, he’ll complete the no-hitter 9.98 percent of the time. This is a point that’s reached about every 75 games (every 150 team games). Get through seven and the chances more than double. With six outs to go, the no-hitter is completed 24 percent of the time, but more than three in four still end in disappointment and probably a standing ovation.

While things will change based on the match-ups, and offensive era has a little bit to do with it, you can create from these 65 years of history the likelihood of a completed no-hitter if it’s through any number of outs.

The chances of completing a no-hitter generally double every inning a pitcher gets through. So if you’re looking for 50-50, just throw the game on when he’s warming up for the ninth. You’ll catch a no-hitter just a shade over 50 percent of the time and a memorable finish no matter what.

There are a number of other questions this data can help answer. Is there a real danger point for no-hitters to be broken up? As the game gets later, one might think the first batter of an inning could see a spike in no-hitters being broken up, whether it’s because a pitcher is out of rhythm or maybe sat for a long time or was even out on the bases. Or maybe it’s the opposite. After a pitcher’s been out there a while in an inning, he’s more susceptible to giving up a first hit.

Both were reasonable theories, but we don’t see either one of them in practice. In each inning, there’s a different pattern for when no-hitters are broken up, and with the seventh inning being a hot spot, that’s likely a product of who would bat in the lineup in the seventh inning and the third time through the order penalty.

As expected, in addition to the downward trends within each trip through the order there is a general downward trend to the proportion of still-active no-hitters that are broken up at each number of outs in the game. The longer you go, the better the pitcher is that day. And finally, with 26 outs is the first time that that next out comes more than 80 percent of the time.

The most dangerous point in a no-hitter isn’t the first out, although that’s close. It’s trying to get the 10th out, which combines a second trip through the order with generally facing somebody near the top.

And finally for the gamblers and beat writers out there, here’s a little tip for you. If you enjoy the no-hitter pool, which I wrote about in this article on betting at the ballpark, this data gives a little way to get a bit of value from some bad luck.

As a refresher, when a pitcher has a no-hitter through three innings (which we now can say will happen every six games or every 12 team games), 10 bettors all put in $5 and pull a card from a deck that contains ace through nine and joker. Pull a 7, and you win the pool if the No. 7 hitter breaks up the no-hitter. Pull a Joker and you only win if the no-hitter is completed.

This data tells the exact historical value of the joker. Since a pitcher who is through three innings with a no-hitter will finish it 0.82 percent of the time, the card that you paid $5 for and gives you a chance to win $50 is worth 41 cents. If anyone will give you a buck for the thrill of taking that shot, just take the dollar, take your losses and run.

Thanks to Rob McQuown for research assistance.

All charts ignore games that don’t go a full nine innings and also eight-inning no-hitters. They also count any game where the no-hitter is broken up in extras as a no-hitter to avoid being reliant on the score of the game.