SBSQ #24: Extra special nerdy edition
Calculus vs. statistics, quarterback clutchness — and my media consumption diet.
Welcome to the Rickey Henderson (#24) edition of Silver Bulletin Subscriber Questions. The questions in this edition err toward the nerdy side. So since we’re in a fastidious mood, let’s start with a bit of housekeeping:
Don’t neglect the item we ran this weekend on whether America is ready for a gay president — or at least, whether it’s ready for Pete Buttigieg — which was also inspired by a subscriber question.
SBSQs are supposed to post around the first of the month. While it’s almost the first of October, this is actually the September edition of SBSQ. While we could just take a mulligan to catch up, let’s try something a little different. SBSQ #25 will be a video edition featuring me and Eli McKown-Dawson; you can leave questions for either or both of us in the comments below. (You’re welcome to ask about anything, but questions about elections and/or Silver Bulletin are slightly preferred.) We’ll run #25 at some point in mid-October. The dirty little secret is that video content requires a bit less work to produce1, so this should hopefully let us get back on track for future months.
Eric Adams dropped out of the New York mayoral race on Sunday, although he’ll still appear on the ballot. I wrote about the race recently enough, including the contingency that Adams would drop out, that I don’t think this is worth a full post, but I’ve added a (fake) SBSQ prompt with a few thoughts.
ELWAY, our NFL team prediction model that will be a companion to our QBERT quarterback ratings, is still coming — but not this week. I guess “I’m too much of a perfectionist” is always a convenient excuse when your homework is late, but we’ve been busy adding features to ELWAY that initially we were going to pass on, particularly accounting for injuries and building a simulator to produce realistic game scores (e.g. 24-17 is much more likely to occur than 25-15). Basically, ELWAY has turned into a big project, almost as big as an election model. We are really nearing the end of the work. So it’s coming soon. But this week? No.
Onto this month’s questions:
Should high schools teach more statistics and less calculus?
Is there such a thing as a clutch quarterback?
Does Adams dropping out give Cuomo a real shot?
The jobs market was worse than we thought. How should that change our assessment of 2024?
Nate’s media consumption diet, a.k.a. Always Be Reading
Should high schools teach more statistics and less calculus?
Friend-of-the-newsletter Matt Glassman asks:
Where do you stand on teaching statistics, data analysis, and game theory in public schools? My view is that is prioritizing this over pre-calculus and calculus would on balance be better for most students, though there’s certainly a trade-off.
I’m not sure whether I’m the perfect person to answer this question or I’m hopelessly biased because I use probability and statistics nearly every day in my work, while I basically never2 directly use any sort of calculus (although techniques like linear regression are derived from calculus). But I’m also not alone. Claude estimates that 1-3 percent of Americans use calculus daily in their work, while 8-15 percent use statistics regularly. And that’s to say nothing about what happens outside of work. Some fuzzy notion of expected value is essential in everything from which route to take to drop your kid off at school to how to invest your savings.
So, yes, I’d want to see coursework in statistics, data analysis, and, more generally, statistically flavored logical thinking be the default path for high school students — while still offering calculus for students who want it.
However, I’ve been asked versions of this question for a long time — Matt and I aren’t the first people to wonder why high school curricula are still so calculus-intensive. And it’s not like I spend a lot of time in high school classrooms. So let’s make an effort to ground this in some data:
These are the number of students who completed AP exams in calculus, statistics, and computer science since 2002. There have long been two calculus tracks — AB and BC — and the College Board added a Precalculus exam in 2024. It also added a second computer science exam, Computer Science Principles, in 2017. There’s still just one lonely stats exam, meanwhile.
While the number of computer science test-takers has grown at a rapid rate, the gap between AP Calc and AP Statistics has only marginally declined — and it’s increased again if you include precalculus.
So it sure looks like there is a lot of stickiness in public school curricula. In a mid-20th-century world where professions involving math tilted toward engineering-type tasks — as its name implies, even Silicon Valley started out with semiconductors — calculus might have made more sense as a capstone than in today’s world dominated by computing and data science. Schools have been slow to adapt.
In addition to stubbornness, there may also be signaling value in calculus coursework. This study suggests that students who take AP Calculus are more likely than those who take AP Statistics to enroll in elite colleges — although they wind up with about the same earnings down the line. In an environment of persistent grade inflation, calculus may be taught as more of a “weed-out class” than statistics, and certain students and admissions offices might like that. Though I’d argue that’s part of the problem: most high school classes are taught with a goal of achieving proficiency to help students navigate adulthood. In contrast, calculus is either taught as a sort of mathematical IQ test or as quasi-professional training. But students don’t necessarily want or need that training and will have to relearn the material again in college if they do. Anyway, yes, more stats and less calc, please.
Is there such a thing as a clutch quarterback?
Charlie Mass asks:
Smart NFL commentators love to point out that teams usually don’t win more than 50% of one score games - it’s one thing Bill Barnwell in particular harps on when talking about teams regressing year to year. But Tom Brady was 251-84 in those games, Mahomes is cited as being even better by different sources, etc. So what gives? Is winning one score games a fluke, or do the best QBs over perform in the “clutch”?