Last weekend’s no-hitter by Edinson Volquez brought up a frequent debate amongst baseball fans: is the no-hitter jinx real? This “unwritten rule” is that if you mention that a pitcher’s hurling a no-hitter, it won’t come to fruition. This is often practiced by teammates in the dugout, fans in the stadium, broadcasters covering the game from the stadium, and of course by fans on Twitter.
We decided to get to the bottom of this longtime legend. If you see a no-hitter in progress this year, are you free to talk (or Tweet!) about it, or are you going to be the jinx? Let’s take a look at the data to find out.
There’s not quite a feasible way to measure how much conversation there is out in the world about a (potential) no-hitter, but Twitter data serves as a great proxy in this case. And since - to some! - the folklore allows for creative turns of phrase (“no runners have reached base other than a walk”), just not the words “no hitter”, there’s a simple way to query the conversation by looking for the number of Tweets that use that particular phrasing (along with a few hashtagged versions & alternate spellings). We want to explore the relationship between the number of Tweets mentioning a no-hitter in progress, and whether the no-hitter was successfully completed.
To dig into the jinx theory, we focused on the 2015 season, which had 34 no-hit bids (defined for this research as no hits through six complete innings) and seven completed no-hitters, giving us more games to base the analysis on than last year’s season with just one no-hitter or the one so far this season. (Thank you to Baseball Reference @Baseball_Ref for compiling game data.)
Looking at volume level from the entire game time will include Tweets that either celebrate a successful no-hitter or react to it getting broken up. To remove those surges and to standardize the time period for each game, we opted to focus on the amount of Tweets in approximately the first seven innings of the game. We measured the number of no-hitter Tweets from start-of-game through 40 minutes before final pitch for a broken-up no-hitter (in 2015, the average MLB game was around 3 hours, which means that each inning lasts approx. 20 minutes) and 32 minutes for a completed no-hitter (average no-hitter game time was 2:31).
Here’s what the pace of conversation looks like for a no-hitter compared to one that got broken up, focusing on two of @Max_Scherzer’s outings (he also had an additional no-hitter *and* an additional no-hit bid that season):
Now it’s time to zero in on the potential jinx. First, we checked how many Tweets mention “no hitter” on an average in-season day in 2015 (excluding days that saw a no-hitter or bid): around 700. Games that had a no-hitter into the 7th, but did not end as a no-hitter, saw around 2,500 Tweets in the -30min window. And no-hitters? Those had over 5,000 Tweets in the -32min window.
So what does that tell us? There are more than twice as many Tweets saying the specific phrase “no-hitter” in (approximately) the first 6 innings for games that are indeed successful no-hitters than there are in games that fail to see it through. In other words: based on Twitter data, talking about a no-hitter does not jinx it. (But we won’t go so far as to suggest a new superstition that you have to Tweet about a no-hitter in order to make it happen…)
Next time you see a no-hitter in progress (I’ll be crossing my fingers that it’s coming from my @SFGiants), feel free to Tweet about it! In fact, you may just jinx it if you don’t.