Catcher Is Baseball’s Most Endangered Position

The rebuilding Baltimore Orioles hope Adley Rutschman becomes the new face of the franchise. Dubbed the best MLB draft prospect since Bryce Harper by Baseball America, Rutschman was taken with the first overall pick last June, the first full-time catcher selected No. 1 since Joe Mauer in 2001. But he could also become a new archetype for his position: a catcher whose primary value is his hitting.

Rutschman told FiveThirtyEight that the story behind his bat is one of nurture as much as nature. As a young player, he taught himself to switch hit with the aid of his father, a baseball coach at George Fox University, a Division III school in Oregon. They would head to the college field with a 5-gallon bucket of balls, or he would hit into a portable net in their home’s driveway. After his freshman season at Oregon State, when he hit just .234, Rutschman rebuilt his swing to add more lift and power. As a sophomore with the Beavers, Rutschman slashed .408/.505/.628, helping the school to a College World Series title and setting a series record for hits. He was even better as a junior, batting .411 with a 1.326 OPS. MLB.com and FanGraphs project him for above-average hitting and power in the majors.

In striving to improve, Rutschman tries to embrace change. “Being OK with being uncomfortable is a huge thing I like to go by,” he told FiveThirtyEight.

Catchers and MLB teams might soon have to get very comfortable with change. Catcher is the one position on the field for which teams have still routinely traded offense for defense: It’s the rare position that has gained in defensive importance this century. But no other position is threatened with having so much defensive value stripped away. An automated strike zone, which MLB has begun to experiment with, would eliminate the value of receiving, and advances in training are already shrinking the gulf in defensive skill at the position.

Yankees catching instructor Tanner Swanson told FiveThirtyEight that an automated zone would completely change the position: “It would be almost a second DH.”

All of this means that finding bats to man the backstop might become very important. It’s a potential sea change in baseball, but Rutschman and the next wave of catchers might be well positioned for it: MLB.com’s top 10 catching prospects include five catchers projected to be better-than-average hitters.

Catchers have been among the worst hitters in baseball — and sometimes the very worst offensive performers in a given season. They’ve often competed with shortstops for the distinction of worst hitters in the game, according to weighted runs created plus (wRC+), which adjusts for the ballpark and the run-scoring environment in any given year. But since the 1970s, shortstops have consistently improved in offensive performance, crescendoing in recent years to what is now described as a golden age at the position. Catchers, meanwhile, continued to lag.

Teams can afford more offense at shortstop, once a defense-first position, because of the confluence of the rise in strikeouts, home runs and defensive shifts: Shortstop is one of many positions seeing fewer and fewer defensive opportunities. If defense is less important, it makes sense that clubs would seek more run-production ability from that position.

But unlike other defensive positions on the field, the position of catcher has added defensive value in recent years. Why? Over the last decade, first hobbyists then teams have quantified the hidden and powerful skill of pitch framing, a catcher’s ability to receive borderline pitches in a way that gets them called as strikes. Teams began to understand the vast difference in framing skill at the position and the number of runs that could be saved by good framers — and runs cost by poor receivers. They started to target players who possessed framing skills and even teach those who didn’t. The value in framing is thought to account for the majority of defensive value behind the plate.

But the skill faces two threats, one existential, the other more immediate.

Framing relies on the fallibility of the human eye, the home plate umpire’s perception of the strike zone. But MLB now has the technology to automate the strike zone. It has experimented with such a zone in the Arizona Fall League and had planned, before the pandemic struck, to bring it to some minor league parks in 2020. Adding an automated zone has also been discussed this year as a way to allow umpires to maintain appropriate distance from the catcher and batter if there is a season. But if MLB chose to automate the zone, catchers could lose much of their value.

“Without question, it [would] completely shift the scope of skills that are required to now play the position. It really shifts the population of who you put back there,” Swanson said. “[Catcher] then becomes more of an offensive position.”

Catchers still perform other important defensive duties, of course: They throw out runners, make pitch selections and block bad pitches. But catchers’ arms have become less relevant as the number of stolen bases drops. “The blocking piece becomes more important,” Swanson said, “[but] I don’t see the throwing piece changing unless you start playing less athletic catchers, and maybe to combat that teams start trying to run more. So it’s the cat-and-mouse game.”

One of the greatest drawbacks of the automated zone, Swanson said, is that a catcher would not even have to catch the ball unless there were runners on base or two strikes. He suggested that stealing first base might need to be an option to ensure that catchers actually try to catch pitches, particularly awkwardly located ones.

But even if MLB decides against an automated zone, there’s another challenge to the framing value of the position: Catchers are getting better at it, so the gap between the best and worst framing catchers is shrinking. Over the last decade, the standard deviation of framing runs — a measure that relies on pitch-tracking technology to determine the probability of a called strike and the associated run value for each pitch, for qualifying players<a class="espn-footnote-link" data-footnote-id="1" href="https://fivethirtyeight.com/features/catcher-is-baseballs-most-endangered-position/#fn-1" data-footnote-content="

Catchers with at least 1,000 pitches caught in a season.

“>1 — has fallen.

The skill gap in pitch framing is shrinking

Difference by standard deviation in framing runs saved by MLB catchers by year since 2008

Year Standard deviation in framing runs saved
2008 12.21
2009 10.23
2010 9.36
2011 11.69
2012 10.04
2013 9.28
2014 9.34
2015 8.27
2016 8.38
2017 9.62
2018 6.84
2019 7.42

Source: Baseball Prospectus

If the floor continues to rise, teams will be forced to look elsewhere to create competitive advantage at the position.

Orioles assistant general manager Sig Mejdal said the team had considered that catchers’ defensive value could be threatened before they selected Rutschman with the No. 1 overall pick. But Rutschman has a strong arm, and the throwing component will always be a part of the position, Mejdal noted. Rutschman has received high grades for his all-around ability behind the plate and hitting at it. Rustschman graded as an average framer in a small minor league sample last summer, but some evaluators are bullish on his framing upside.

“I think how we evaluated catchers a decade ago or two decades ago, when we really didn’t appreciate pitch framing, I think we would return to those times” if the value of framing diminishes, Mejdal said. “Of course, there are less stolen bases in the game now, but I still think there is going to be significant appreciation of their defensive skills.”

That appreciation could perhaps include skills that have largely been unquantifiable to date. Mejdal said the new Orioles regime has begun to investigate the value of pitch calling. “It’s irresponsible not to,” he said.

It’s possible that pitch calling will become the new pitch framing. But as innovation changes the game, and without more value placed on specific skills behind the plate, catchers might need to take a whole new approach to their position. Catchers might soon need to turn into sluggers, just like Rutschman.

We Might Never Know Who 2021’s Best College Football Recruits Are

Opinions vary about how much college football recruiting matters, but it likely falls somewhere between irrelevant and all-encompassing. Each of the past 18 national champions has signed at least one top-10 recruiting class in the four years leading up to its title, but a highly ranked class guarantees nothing, which is how Appalachian State can rank 100th in a four-year weighted recruiting average but 29th in ESPN’s Football Power Index, and Arkansas can rank 25th in recruiting but 91st in FPI.<a class="espn-footnote-link" data-footnote-id="1" href="https://fivethirtyeight.com/features/we-might-never-know-who-2021s-best-college-football-recruits-are/#fn-1" data-footnote-content="

A skeptic might note a direct correlation between how much emphasis a coach claims to place on recruiting and how strong his last class was. “What rankings don’t do though is crack their chest open and look at their heart,” Texas head coach Tom Herman said after signing the 25th-ranked class in 2017. When he signed the third-ranked class in 2018, he said, “This needs to be the new normal.”

“>1 The early part of summer is when programs build their futures by hosting recruits on visits, evaluating them at camps and seeing them at their high schools.

But the coronavirus pandemic has shut down those activities, along with everything else in college athletics, and the effects could be sweeping. Nike has canceled high-profile Elite 11 regional quarterback camps. Under Armour’s recruiting camps are postponed, too. The NCAA has mandated a recruiting dead period through June 30, and stay-at-home orders wiped out virtually all on-campus recruiting events.

COVID-19 has placed the rest of the 2021 recruiting cycle in flux, but any recruiting evaluation this far ahead of the early signing period in December is inexact. For the class of 2020, 49 players were not in any of the three major scouting service’s rankings<a class="espn-footnote-link" data-footnote-id="2" href="https://fivethirtyeight.com/features/we-might-never-know-who-2021s-best-college-football-recruits-are/#fn-2" data-footnote-content="

The ESPN 300, the Rivals.com 250 and 247Sports.com’s “Top 247.”

“>2 during the week of July 22, 2019, but did appear in one of the final rankings. Fifteen prospects jumped by more than 100 spots in an average of the three rankings.

Some players made huge leaps their senior seasons

The 15 football players ranked in the class of 2020′s top 300 on ESPN, Rivals and 247Sports with jumps of at least 100 spots from last year

Aggregate rankings
Player Committed school In July 2019 Now Change
Jahmyr Gibbs Georgia Tech 339 83 +256
Edgerrin Cooper Texas A&M 339 157 +182
Van Fillinger Utah 339 162 +177
Marvin Mims Oklahoma 327 154 +173
CJ Stroud Ohio State 215 43 +172
Jalin Hyatt Tennessee 339 170 +169
Jimmy Calloway Tennessee 339 170 +169
Henry Parrish Mississippi 339 178 +161
Donell Harris* Texas A&M 199 51 +148
Josh Downs North Carolina 228 95 +133
Davin Vann NC State 339 232 +107
Loic Fouonji Texas Tech 339 234 +105
Eric Shaw South Carolina 339 235 +104
Alfred Collins Texas 160 58 +102
Ozzy Trapilo Boston College 313 211 +102

Players not among the 338 ranked in the top 300 of at least one of the three major leaderboards were assigned a ranking of 339.

*Harris reclassified from the 2021 class to 2020 in June 2019.


When the dust settled on the 2020 class, the biggest riser was Jahmyr Gibbs, a running back from Dalton, Georgia. In the 247Sports.com composite rankings on Feb. 15, 2019, Gibbs was the 1,010th-ranked player in the country. At this time last year (May 31), he was 587th, and by the start of the season, he was still only 438th. But he opened his senior year by rushing for 420 yards and eight touchdowns on 23 carries, and he finished last season as a Sports Illustrated All-American, with 2,554 rushing yards and 40 touchdowns — the kind of explosion that makes everyone take notice. By Oct. 15, he was up to 300th in the 247Sports.com composite rankings, and now he is 178th on ESPN.com, 70th on Rivals.com and 44th in the 247Sports.com rankings.

A breakout season like Gibbs’s is not uncommon, but this fall, a season like that may not be possible. States across the country are all facing questions about whether the 2020 high school football season can go on. Some coaches in California have even discussed moving part of the season to next spring, which could be after signing day. With college camps and on-campus visits also canceled for much of the year, the recruiting rankings next spring could look much like they do today — which, as we know, would not account for potential player improvement. Had he not played 11 games last fall, Jahmyr Gibbs would remain unknown to most of college football outside the state of Georgia.

“These guys that are five stars and have been committed since eighth grade, it’s not going to affect those kids,” Dalton coach Matt Land told me this week. “But there’s so many kids that are not those kids, that are one-star, two-star, three-star kids, that earn these scholarships because they have 2,000 yards on the year or they have 11 interceptions on the year. And they’re not going to have those numbers, potentially, because they’re not going to have those games.”

Gibbs’s rise didn’t change his college decision. He committed to Georgia Tech on May 25, 2019, and signed there in February despite late interest from Alabama, Ohio State, LSU and Florida. But many fortunes do change as the rankings fluctuate over the course of a season. A year ago, quarterback CJ Stroud’s most recent offers had come from Baylor, New Mexico, Kansas and Washington State; since then, the four-star California recruit has vaulted from No. 860 to No. 41 in the 247Sports.com composite rankings and signed with Ohio State.

A slower recruiting cycle could mean that midtier schools hold onto their overlooked prospects, which is telling, because right now the 2021 team rankings show a number of anomalies: No. 2 Tennessee, No. 4 North Carolina, No. 8 Minnesota, No. 9 Iowa, No. 20 Georgia and No. 46 Alabama. Those rankings may be misleading, as they value not just average player rating but also the number of players committed to a school at that time. Georgia and Alabama recruits so far are averaging ratings of over 94 on a 100-point scale; if those schools fill out a class with the same average player rating, they’ll end up in the top five — and push out other schools that look good right now.

But North Carolina, for example, may be in the top 10 for good. Mack Brown’s program already has 14 commitments, including 10 four-star prospects. Some of them could surely change their minds, but 13 of the 14 commits are in-state players, and if the coronavirus continues to shut down or limit travel into the fall, recruits may be unable to visit campuses as usual and hesitant to pick a school far from home. “You love your family, you’re wanting everybody healthy,” Brown told the Associated Press. “So I do think that this is encouraging people to stay closer to home.” California defensive end Korey Foreman, the nation’s No. 1 overall prospect in 247Sports.com’s composite rankings, initially committed to Clemson but has reopened his recruitment in part because of the long distance.

College football has come to be defined by a measure of near-certainty: The elite programs<a class="espn-footnote-link" data-footnote-id="3" href="https://fivethirtyeight.com/features/we-might-never-know-who-2021s-best-college-football-recruits-are/#fn-3" data-footnote-content="

Only 11 schools have made the playoff, and four — Alabama, Clemson, Oklahoma, Ohio State — have done so more than once.

“>3 routinely sign elite recruiting classes,<a class="espn-footnote-link" data-footnote-id="4" href="https://fivethirtyeight.com/features/we-might-never-know-who-2021s-best-college-football-recruits-are/#fn-4" data-footnote-content="

Only Alabama, Georgia and Ohio State can claim a top-two recruiting class in any of the past four years.

“>4 and those recruits have routinely developed into elite players. But the coronavirus brings an infusion of uncertainty, the kind of force that could rearrange a hierarchy. Alabama coach Nick Saban, the pillar of reliability in college football, seems to have a grasp of what we’re in for the rest of the year: “We’ll do the best we can with what we’ve got,” he said, “and I’m sure it’s going to keep on changing.”

Politics Podcast: Coronavirus Is Becoming Even More Partisan

After spending much of the past three and a half years out of the spotlight, former President Barack Obama has become more vocally critical of President Trump’s administration. In turn, Trump has called for Obama and his administration to be investigated. In this installment of the FiveThirtyEight Politics podcast, the crew asks who is benefiting from a high-profile conflict between Obama and Trump in an election year. They also look at evidence suggesting that Americans’ views on the response to the coronavirus pandemic are growing increasingly partisan.

You can listen to the episode by clicking the “play” button in the audio player above or by downloading it in iTunes, the ESPN App or your favorite podcast platform. If you are new to podcasts, learn how to listen.

The FiveThirtyEight Politics podcast publishes Mondays and Thursdays. Help new listeners discover the show by leaving us a rating and review on iTunes. Have a comment, question or suggestion for “good polling vs. bad polling”? Get in touch by email, on Twitter or in the comments.

The Industries Hit Hardest By The Unemployment Crisis

On Thursday, the U.S. Department of Labor reported that 2.98 million more Americans had filed for unemployment insurance in the previous week, bringing the total for claims since March 8 up to 36.8 million. Last week’s jobs report for the month of April estimated the national unemployment rate at 14.7 percent, and Thursday’s report implies that perhaps another 2.5 percent of the labor force has filed for benefits since the beginning of May.<a class="espn-footnote-link" data-footnote-id="1" href="https://fivethirtyeight.com/features/the-industries-hit-hardest-by-the-unemployment-crisis/#fn-1" data-footnote-content="

Not all these people will show up in the official unemployment number by the end of May. Some may find jobs in between, and some may stop looking. Unemployment rate and initial claims are computed through very different processes, but this number gives at least a sense of the ongoing scale of the crisis.


It’s a grim situation, though not every area has been hit equally hard. So which parts of the economy have seen the most damage? Using data from the most recent jobs report, I dug into changes in total nonfarm employment by sector from February — the last full month before the coronavirus-triggered economic crisis truly began — to April, starting at the ”supersector” level, or the most broadly grouped parts of the economy:

Which major industries are hit hardest by unemployment?

Economic “supersectors” by percent change in seasonally adjusted nonfarm employees from February to April 2020

Employees (millions)
Supersector Feb. Mar. Apr. 2 mo. % chg.
Leisure and hospitality 16.9 16.4 8.7 -48.3%
Other services 5.9 5.9 4.6 -22.0
Construction 7.6 7.6 6.6 -13.2
Trade, transportation and utilities 27.8 27.8 24.7 -11.2
Education and health services 24.6 24.5 21.9 -10.8
Manufacturing 12.9 12.8 11.5 -10.6
Professional and business services 21.6 21.5 19.3 -10.4
Information 2.9 2.9 2.6 -8.9
Mining and logging 0.7 0.7 0.7 -8.0
Government 22.7 22.7 21.7 -4.4
Financial activities 8.8 8.8 8.6 -3.0

Source: bls.gov

Over the past few months, every single major area of the U.S. job market has shrunk at least some. But it’s clear that the leisure and hospitality industry was hurt far more than any other. In February, nearly 17 million Americans worked in that supersector, which comprises the arts, entertainment and recreation sector — including spectator sports and gambling — and the accommodations and food services sector. By April, that number had fallen by nearly 50 percent, to just 8.7 million.

The leisure and hospitality fields suffered heavy job losses

Percent change in seasonally adjusted leisure and hospitality employees from February to April 2020, by supersector, sector and subsector

Employees (millions)
Industry Feb. Mar. Apr. 2 mo. % chg.
All leisure and hospitality 16.9 16.4 8.7 -48.3%
Arts, entertainment and recreation 2.5 2.4 1.1 -54.5
Performing arts and spectator sports 0.5 0.5 0.3 -45.4
Museums, historical sites and similar institutions 0.2 0.2 0.1 -26.1
Amusements, gambling and recreation 1.8 1.8 0.7 -59.9
Accommodation and food services 14.4 13.9 7.6 -47.3
Accommodation 2.1 2.0 1.2 -42.3
Food services and drinking places 12.3 11.9 6.4 -48.1

Source: bls.gov

Nearly all the jobs in this supersector involve some form of in-person services rendered for gatherings of people — which, of course, became impossible to maintain when social distancing requirements were put in place by local authorities. As such, leisure and hospitality businesses were always going to struggle under the pandemic’s terrible constraints. Every subsector of this industry has lost at least 25 percent of its employees since February, and most have lost 40 percent or more. Among all the sectors and subsectors tracked by the Bureau of Labor Statistics for the month of April,<a class="espn-footnote-link" data-footnote-id="2" href="https://fivethirtyeight.com/features/the-industries-hit-hardest-by-the-unemployment-crisis/#fn-2" data-footnote-content="

Some very specific sub-subsectors don’t have April data yet.

“>2 only one — scenic and sightseeing transportation, which falls under the umbrella of trade, transportation and utilities — saw more job losses than amusements, gambling and recreation, which shed 60 percent of its employees in a two-month period.

Although it wasn’t hit as hard as leisure and hospitality, the “other services” supersector also saw a 22 percent reduction in positions from February to April. That area has three subsectors:

Services are struggling (especially hairdressing and laundry)

Percent change in seasonally adjusted “other service” supersector employees from February to April 2020, by supersector and subsector

Employees (millions)
Industry Feb. Mar. Apr. 2 mo. % chg.
All other services 5.9 5.9 4.6 -22.0%
Personal and laundry services 1.5 1.5 0.7 -53.5
Repair and maintenance 1.4 1.4 1.2 -16.5
Membership associations and organizations 3.0 3.0 2.8 -8.5

Source: bls.gov

None is doing particularly well, but by far the biggest unemployment disaster has played out in personal and laundry services, which includes hairdressing, dry cleaning, laundromats and a vast array of other service jobs such as pet care, dating services, photofinishing, parking-lot services and so forth. The struggles of hairstylists and cosmetologists during this crisis are well-documented, as these professions made up more than half the jobs in this supersector in 2019. And although laundry workers have been deemed essential in some states, many have seen business volume drop significantly — not only because customers have been avoiding unnecessary trips outside but also because homebound teleworkers aren’t dressing as nicely as they otherwise would. As a result, the number of employees working in that slice of the economy has fallen by 54 percent since February.

For most other supersectors, jobs have dropped by somewhere between 8 and 13 percent since before the coronavirus crisis started, meaning almost every corner of the economy has been touched by the pandemic. The only areas experiencing more modest (though still significant) losses were in government — with workers at the federal level faring better than their state or local counterparts — and financial activities:

Financial and government jobs are weathering the storm

Percent change in seasonally adjusted financial activities and government supersector employees from February to April 2020, by supersector, sector and subsector

Employees (millions)
Industry Feb. Mar. Apr. 2 mo. % chg.
All financial activities 8.8 8.8 8.6 -3.0%
Finance and insurance 6.5 6.5 6.5 -0.5
Monetary authorities – central bank 0.0 0.0 0.0 +2.1
Credit intermediation and related activities 2.7 2.7 2.7 -1.0
Securities, commodities, investments, funds, trusts 1.0 1.0 1.0 -0.4
Insurance carriers and related activities 2.8 2.8 2.8 -0.2
Real estate and rental and leasing 2.4 2.4 2.1 -9.7
Real estate 1.8 1.7 1.6 -5.9
Rental and leasing services 0.6 0.6 0.5 -21.5
Lessors of nonfinancial intangible assets 0.0 0.0 0.0 -2.5
All government 22.7 22.7 21.7 -4.4
Federal government 2.9 2.9 2.9 +0.7
Federal, excluding U.S. Postal Service 2.3 2.3 2.3 +0.9
U.S. Postal Service 0.6 0.6 0.6 -0.2
State government 5.2 5.2 5.0 -4.3
State government education 2.5 2.4 2.3 -8.8
State government, excluding education 2.7 2.7 2.7 -0.3
Local government 14.7 14.7 13.9 -5.5
Local government education 8.0 8.0 7.6 -5.9
Local government, excluding education 6.6 6.6 6.3 -4.9

Source: bls.gov

The central bank and federal government (excluding the U.S. Postal Service) had more jobs in April than in February, an absolute rarity in the current economic climate. But even among the supersectors weathering things best, certain areas have seen plenty of job losses: The real estate market is teetering on the edge of a rent crisis, as unemployed tenants struggle to make payments, and rental/leasing jobs are down more than 20 percent from a few months ago. Both state and local government education are also down more than the overall government average over the past few months.

But those areas are still doing better than some others. If we dig deeper into the subsectors of the economy, the differences between those with the heaviest and lightest job-loss numbers are striking:

Which subsectors lost jobs — and which are surviving?

Subsectors with the largest and smallest percent change in seasonally adjusted employees from February to April 2020, by size of subsector

Large subsectors
Largest Losses 2 mo. % chg. Smallest Losses 2 mo. % chg.
Amusements, gambling and recreation -59.9% Federal government, excluding U.S. Postal Service +0.9%
Food services and drinking places -48.1 Insurance carriers and related activities -0.2
Accommodation -42.3 State government, excluding education -0.3
Transportation equipment -24.4 Credit intermediation and related activities -1.0
Administrative and support services -18.0 Food and beverage stores -1.5
Medium subsectors
Largest Losses 2 mo. % chg. Smallest Losses 2 mo. % chg.
Clothing and clothing accessories stores -58.9% Couriers and messengers +0.1%
Personal and laundry services -53.5 U.S. Postal Service -0.2
Performing arts and spectator sports -45.4 Securities, commodity contracts, investments, and funds and trusts -0.4
Transit and ground passenger transportation -37.4 Telecommunications -0.8
Sporting goods, hobby, book and music stores -36.2 Computer and electronic products -0.9
Small subsectors
Largest Losses 2 mo. % chg. Smallest Losses 2 mo. % chg.
Scenic and sightseeing transportation -62.1% Monetary authorities – central bank +2.1%
Motion picture and sound recording industries -48.3 Other information services +1.0
Furniture and home furnishings stores -46.3 Pipeline transportation +0.0
Apparel -40.6 Rail transportation -1.7
Museums, historical sites and similar institutions -26.1 Data processing, hosting and related services -2.0

Category sizes are based on whether the area ranked among the top, middle or bottom third of subsectors in total employment in February 2020.

Source: BLS.gov

Not surprisingly, the largest losses at this level have also disproportionately hit the retail, service and hospitality subsectors. But no matter the size of the industry, only a few areas managed to resist a net decline in employees, compared with their pre-coronavirus norms.

With even more people filing unemployment claims every day that the coronavirus lockdowns continue, these numbers are likely to get even worse. By definition, recessions always exact a terrible toll on the job market. But this one is unusual in just how many people are out of work — and how wide-ranging the unemployment has already been.

Would it hurt the economy to let people die from coronavirus? l FiveThirtyEight

Why A Voting App Won’t Solve Our Problems This November

At 106, MacCene Grimmett is one of the oldest voters in the state of Utah. Though women didn’t have the right to vote when she was born in 1913, by the time she was of voting age, the 19th Amendment had passed. She has voted in every election since, she told her local Fox affiliate, including the Utah County municipal general election last November.

But that time, the centenarian cast her ballot in a novel way: She voted via an app.

America is 174 days away from a presidential election. It’s also in the middle of a pandemic that upended normal life, requiring mass shutdowns and social distancing. Those two things don’t exactly jive.

Having millions of Americans stand in crowded polling places for hours to cast a ballot on Election Day sounds like the makings of a public health disaster — especially if there is a second surge of COVID-19 infections in the fall, as some experts predict. So now, election officials are looking for ways to hold elections remotely. One option that has been proposed is voting via an app on a smartphone or electronic device, just like Grimmett did last fall (though so far, states seem to only be considering this option for certain groups of voters, such as voters with disabilities).

It seems like an obvious solution: With so much of our daily lives now virtual, why couldn’t our elections be moved online too?

Voting online or via an app has even been tested in small elections a handful of times, but election security experts and even the founder of one of the most prominent voting apps on the market, Voatz, say there’s a laundry list of reasons why this technology isn’t ready for prime time. (Not to mention the fact that 19 percent of Americans still don’t have a smartphone, and as many as 21.3 million Americans still lack access to broadband internet, according to the Federal Communications Commission.)<a class="espn-footnote-link" data-footnote-id="1" href="https://fivethirtyeight.com/features/why-a-voting-app-wont-solve-our-problems-this-november/#fn-1" data-footnote-content="

As of 2017, the most recent numbers the FCC makes available. The agency’s standard definition for broadband is 25 Mbps download speed and 3 Mbps upload.


“I don’t know what I can say to explain this better: This is an incredibly dangerous idea,” said Mike Specter, a computer science Ph.D. student at MIT who has researched voting technology.

Specter told me there are a number of security and privacy concerns with voting online, which includes voting via an app, and that no technology so far has been able to solve these issues.

For starters, there is currently no way to ensure that each individual voter’s device is secure. Malware covertly installed on a voter’s phone could potentially alter the voter’s ballot or prevent it from being properly transmitted, Specter said. And even if the device is clean, election security experts say there are too many steps required to ensure that the ballot a voter submits online is the one actually counted. With a paper ballot, a voter marks their vote by hand and can visually verify it’s correct. A hard copy is also retained, which can then be audited. But with a digital vote, there are many steps that can create a gap between the vote cast and the vote counted.

“If you think about it, we have several versions of what that vote is and there is no way to verify that all those versions are the same,” said Duncan Buell, a professor of computer science and engineering at the University of South Carolina. “We have one version, which is what the voter sees in the form. We have another version, which is what gets transmitted by the software. We have a third version, which is the version that gets received by the storage system and then we have another version, which is what gets printed out and tallied by the election officials.”

And if the vote is intercepted at any point in that chain, there is no way to verify that a change had been made. It’d be like passing your paper ballot down a chain of strangers and trusting that nobody adjusted it before the vote was counted.

For the Utah County municipal election in which Grimmett voted by app, military and overseas voters and voters with disabilities could vote remotely using Voatz. But a report earlier this year by Specter and his colleagues at MIT found multiple security vulnerabilities in the chain of information that a hacker could exploit, including learning how a user voted, changing the user’s ballot or even accessing the user’s private information.

Voatz claimed the researchers’ methodology was flawed, but every online voting platform has faced similar challenges, according to Maggie MacAlpine, co-founder of Nordic Innovation Labs, a security consultancy firm that specializes in safeguarding elections. MacAlpine said when election officials have run trials of other online voting software in the past, they invited white hat hackers (computer security experts who attempt to hack into a system the purpose of assessing vulnerabilities) to test the software live.

“They have always gotten in with laughable ease,” MacAlpine said. “Every single time.”

It’s a longstanding problem, too. In 2010, for example, Washington, D.C., was considering a new online voting platform and invited researchers from the University of Michigan to test it. But when the Michigan fight song began playing after every ballot was successfully cast, it was clear the system wasn’t as secure as officials had hoped. And as the MIT analysis of Voatz shows, things haven’t gotten much better in the last decade.

MacAlpine noted that even if there was a completely secure system, there’s currently no way to have an online vote that is both anonymous and auditable. An anonymous vote protects against voter coercion, suppression, or vote selling. An auditable vote protects against any errors or breaches, because officials can conduct a recount. But that combination, which is possible with a paper ballot, isn’t yet possible online.

Voatz, though not the only online voting vendor in the market, has attracted a lot of scrutiny because it has been used by multiple state and local elections to facilitate absentee voting. The company’s co-founder and CEO, Nimit Sawhney, takes issue with a lot of the criticism the company has received, saying there are multiple layers to security and accuracy that protect against the issues raised. But even Sawhney said that at this point, the company couldn’t handle this fall’s presidential election.

“Nationwide would be a huge stretch,” Sawhney said. “We are a tiny little startup. There are about 25 people on our team. For us to be able to claim that we can do elections for 200 million people on a smartphone? That would be naive.”

So what’s a country to do when a pandemic is forcing us apart, but an online election is still a science fiction dream? Each of the experts I spoke to said the same thing: vote by mail.

Planning needs to start now, to make sure ballots are printed off and mailed in time, and that voters know their options for casting a ballot. In-person voting will still most likely take place as well. But experts told me if we want those well-spaced lines for the ballot boxes to be less than a few miles long, we’ll have to vastly ramp up mail-in voting by November.

“We’re going to have a hard time doing it this year,” Buell said. “But we have almost no choice.”