MASTER NOTES: Crawlin'

As usual on a Saturday, last weekend I finished the chores mandated by Mrs. Masternotes in the morning so I could be ready to watch a little baseball in the afternoon. I have the Extra Innings package, like many of you, and on this Saturday, it happened that the first live game was on MLB.com.

During that game, I took an interest in the items appearing on the “crawl” that modern TV production designers have implemented for fear we might want to actually watch the programming for more than four seconds running. In particular, I was taken how these statistical nuggets had literally no value to help viewers understand player performance. In the age of Statcast and SABR and Baseball-Savant and even a few announcers cautiously mentioning launch angles, many of the news items seemed to be actively and purposely innumerate. About the best that could be hoped for would be “Huh! How ’bout that?”

Here are some examples, with some analysis and then some broadcast-quality hot takes!!!

For example, I saw one blurb pointing out that Carlos Correa had gone 1/14 over his preceding four games. This is an example of imposing a narrative on a situation, even when there’s no narrative actually present, or at least no narrative that actually makes any kind of point. It’s as though Herman Melville interrupted the whale hunting in Moby Dick to note that Fedallah was 29 for his last 116 harpoon tosses over the last six days.

In 2017-19, Correa had five different runs of four games in which he went 1-for-14. They were about as useful for “predicting” about the next day’s game as examining the entrails of goats!!! Not much. Overall, he was 10-for-19, but that was made up of a 2/4, a 3/3, another 3/3, a 1/4 and a 1/5. On Saturday, he went 1-for-4 again.

Next came a note that MIN is 13-3 in games in which they hit two or more HR. That’s a .813 win percentage, which would be a 132-win team in the regular season. So it sounds pretty impressive.

But it seemed intuitively obvious that teams hitting 2+ HR (2+) in games will score more runs than when they don’t, which portends well for its chances of winning those games. Indeed, the stats bear that out. In 2018, 1,542 games included a team hitting 2+ HR. In 270 of those games, both teams hit 2+, so we’ll leave those games out. The average runs scored per game (RS/G) in teams’ 2+ games was 6.4, while the average in non-2+ games was 3.7, a 2.7-RS/G increase.

Not surprisingly, the added runs meant added wins. In the 1,272 games where only one team hit 2+, those teams’ winning percentage increased by a fairly wide range, from +.332 points (DET) to +.057 (CIN). The average gain was +.217.

But MIN’s gain in 2+ games was no big news. It was almost exactly the same this season as last. In 2018, the Twins gained +146 points in winning percentage in 2+ games (.481 overall, .629 in their 2+ games). This season, through May 10, MIN has gained ... 146 points (.667 overall, .813 in 2+ games). So what this item said was, “MIN has won 15% more games this year when they hit 2+ HR than when they don’t, the same as last year.” And the keen baseball fan like, you, caller, will say, “Yeah? So?!!!”

What is interesting is that the Twins' 16 2+ games through its first 36 games through May 10 puts the team on pace for 72 such games in the season. If we assume that they win those games at the same .655-ish percentage, that’s 47 wins. If they play .500 in the other games, that’s 45 more wins   

Next came some pitcher items. The first was a capper: The crawl said that CHC LHSP Cole Hamels was 8-3, 3.57 in 16 career starts vs MIL. This item is just silly, for two reasons. First, 16 starts in a particular context doesn’t say much about any particular start in that same context. It’s like making a point about Hamels’s career record in games pitched on birthdays of the Danish Royal Family.

More problematic in this instance is that the context isn’t even the same, because we’re talking about a single franchise but a bunch of different teams. It’s easy to talk about  “the Brewers,” but which Brewers? Hamels' first start against MIL was on May 18 ... in 2006! That was 13 years ago! The MIL batting order that day, ignoring pitcher Dana Eveland, started with Rickie Weeks Jr., followed by Jeff Cirillo, Geoff Jenkins, Carlos Lee, Prince Fielder, Bill Hall, Brady Clark, and Chad Moeller.

In Saturday’s start, Hamels faced a MIL order that led off with Lorenzo Cain, followed by Christian Yelich, Jesus Aguilar, Yasmani Grandal, Ryan Braun, Mike Moustakas, Hernan Perez, Ben Gamel, and Orlando Arcia. You don’t have to attend First Pitch Arizona or the Sloan Conference to notice that these are two entirely different teams!!! What Hamels did against the 2006 Brewers has no bearing whatsoever on what he will do against the 2019 edition.

For the record, in that 2006 start, Hamels went 6.1 IP, giving up 4 ER on 5 hits and 4 walks—a 5.68 ERA (2.11 higher than career average). On Saturday, Hamels tossed a small gem: 7 IP, 1 ER, 3 H, 3 bb—a 1.29 ERA (2.28 lower than his career average). So we might be able to say that looking at Hamels’s track record against another franchise (not a team) is helpfully accurate to 3.5 ER—plus or minus 2.10. That's about as useful for predicting as ailuromancy!!! *

Finally, the crawl told us that TOR RHSP Marcus Stroman had given up 11 ER in his last 2 GS. This is similar to the other items in that it is almost meaningless.

Before we get to the details, let’s consider this question: “What are we supposed to know or understand from this information?” Does Stroman's mean he's struggling and should be avoided? Or does is it mean he's “due” for a rebound start and should be played? Typically, this analysis comes post facto, with the 20:20 hindsight beloved by so many of our baseball broadcasters.

Chris Liss and I talked about this a couple of weeks ago on the April 26 edition of BaseballHQ Radio, and Chris’s point was that these kinds of streaks do exist, in short and long runs, but that they are absolutely useless as predictive tools. My own 2013 research for BaseballHQ.com disproved the favored broadcaster trope that “home runs come in bunches.”

I haven’t dug into pitcher streaks or droughts, but it was easy to check Stroman’s actual record to see what happens when he has two bad starts in a row. And by “bad,” I’ll use the MLB crawl standard—but while they reported 11 ER in his last two starts (May 1 and May 6), Stroman actually only surrendered nine ER, with two more runs unearned. We’ll go with 9 ER over any two consecutive starts.

In his career, Stroman has had 116 starts before his May 11 start. That implies 115 pairs of games in which to establish a bad streak to MLB.com’s definition. In his 115 start-pairs, Stroman gave up 9+ ER 14 times, with four of the pairs coming in a row to start the 2018 season (4, 5, 4, 6, 6 ER).

How predictive were these bad start-pairs? Judge for yourself. In the games at the end of each 9+ ER start-pair, Stroman gave up 2 ER, 1 ER, 4 ER, 2 ER, 4 ER, 1 ER, 0 ER, 8 ER, 6 ER, 2 ER, 1 ER, 0 ER and 3 ER. Go ahead—bet your kid’s college fund!!!

The crawl had some other dubious notes, but while I’m as suspicious, or more, I didn’t have the space or time to parse out the details. But in true broadcast form, that won't stop me from shouting out some more “hot takes”:

  • Mitch Garver: 3 HR, 6 RBI, .571 through his previous four games. Hot take: Anything can happen in four games!! On Saturday, Garver went 1-for-4 (.250) with a triple, an RBI and a Run Scored.
  • Somebody (I couldn’t read my own handwriting in my notes): No strikeouts over his last 21 PA. Hot take: I know there are lots of strikeouts, but not whiffing in five or so straight games is about as remarkable as mustard on a hot dog!!! According to Baseball-Reference.com’ s wonderful Play Index machine, from the start of the 2018 season through May 10 of this season, there were 262 streaks of 21 or more PA with no strikeouts. One player had eight such streaks, including the top three. (Guess who?**)
  • Jackie Bradley Jr.: 4-for-12 (.333), two doubles, one HR, 4 RBI vs Felix Hernandez. Hot take: Twelve AB is as useful for predicting as throwing the runes!!! Bradley went 0-for-1 with a strikeout against King Felix, who was knocked out of the box in the third after giving up a walk, four consecutive singles, and another walk. Bradley did hit a bases-loaded 2-RBI single off Felix replacement Roenis Elias.
  • Trevor Bauer: “(A)t least 2 H in three of last seven games.” Hot take: Huh!!! How 'bout that!?!

There’s not much we can do with this innumerate nonsense except be aware of it, and maybe berate the perpetrators if you hear them on sports talk radio and you can get through. In the meantime, don’t be fooled by small-sample narrative!!!

* Ailuromancy, aka "felidomancy," is using the movements and jumps of cats to predict the future.

** The master of making contact was Andrelton Simmons, whose top-three whiffless streaks were a 20-gamer and two 14-gamers. Other players with streaks of four whiffless games or more, comprising 21+ PA: Michael Brantley (7); Alex Bregman (6); Joe Panik, Miguel Rojas, Nick Markakis, Yuli Gurriel (5 each); Anthony Rizzo, Daniel Murphy, Eduardo Nunez, Jean Segura, Wilmer Flores (4 each).


Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.