Four maps, no forecast

Three days before the election, the Washington Post published four Senate maps. In one, Democrats win 53 seats. In another, Republicans win 54. A third has Democrats at 55 plus an independent; the fourth is a 50-50 tie broken by Vice President Vance.

The piece is careful and well reported. It explains what would have to happen for each outcome, and what argues against it. What it never does is say which outcome is most likely. The authors state plainly that nothing they examined tells us whether the polls will miss, or in which direction.

So whatever happens on 3 November, the Post will have a map that "saw it coming." That is not a forecast. It is insurance.

I spent my career writing assessments for people who had to act on them, including for the President's Daily Brief. In that world, a paper that laid out four futures and declined to pick one would have come back with a single question in the margin: So what do you think?

What an estimate requires

The intelligence community wrestled with this problem long before election forecasting existed. In the 1960s, Sherman Kent, the father of American estimative analysis, found that policymakers read a phrase like "serious possibility" to mean anything from 20 to 80 percent. His answer was to tie words to numbers, so a judgment could be understood and later checked.

The lesson stuck. A finished assessment has three parts:

1. A judgment. What we think will happen.

2. A likelihood. How confident we are, in words tied to probabilities.

3. Indicators. What we would expect to see if we're right, and what would change our minds.

Scenario analysis, the alternative-futures technique, is a legitimate tool. But analysts use it to stress-test a judgment, not to replace one. Lay out the futures, then say which you expect and why. Skip that last step and you haven't written an estimate. You've written a list of things that could happen, and you've made yourself impossible to hold accountable.

Kent's problem is alive and well in Washington. The Cook Political Report, the capital's most trusted forecaster, still rates races in words: Solid, Likely, Lean and Toss Up. To its credit, Cook has graded itself, which none of the other sources here has done. According to the Washington Post, an internal review of 10,374 races from 1984 to 2024 found that its "Likely" ratings were right 97 percent of the time and its "Lean" ratings 94 percent. Few readers would guess that "Lean" means 94 percent. That's exactly the gap Kent warned about.

Then there's "Toss Up," the one rating that can never be wrong. It's no wonder that, as a Cook editor told the Post, campaigns "all want to be a 'Toss Up' at the end of the day."

That is the line between honest uncertainty and covering yourself. Uncertainty says "I think X, at about 70 percent." Covering yourself says "it could be X, or Y, or Z."

The hedging runs all the way down

The Post isn't the only one protecting itself. Look at the whole chain that turns voters' answers into the numbers you see on the news.

The pollsters. The Post found that in 13 of the 27 races it examined, polls agree with each other more than chance would predict. Its likely explanation: many pollsters now weight their samples to match how people say they voted in 2024. That's a defensible method, but it also means pollsters converge. No one wants to be the outlier.

Nate Silver goes further. Marist's new Maine poll showed the Democrat up 11 points for governor among registered voters, and tied among likely voters, in the same survey. Silver's leading theory is that Marist didn't believe its own Senate number, adjusted its likely-voter model toward the consensus, and broke its governor result in the process. He suggests Marist fears becoming the next Ann Selzer, the Iowa pollster President Trump sued after her final 2024 poll missed. If he's right, that's a pollster managing its own risk rather than measuring the public.

I have a confession to make here. Last week, in "Mainers Vote Against Trump but Bet on Collins," I relied on Marist's Senate numbers while questioning its turnout model, which quietly removed about five points of Democrats from the electorate even though its own enthusiasm measures favored them. The governor's result in the same poll suggests that doubt was warranted. I also found that Mainers themselves are hedging. They treat a tied governor's race as a Pingree win, yet most expect Collins to win hers, even in the Democratic-leaning 1st District, where Jackson runs strongest. Burned by Collins' 2020 upset, voters are doing what pollsters are doing: refusing to believe the Democrat's lead until they see it.

The aggregators. Then came Wedgewood Polls, a firm that admitted in October it had fabricated most of its data. Its fake surveys had been picked up by the New York Times poll tracker, FiftyPlusOne, Decision Desk HQ and Silver Bulletin. The first response from nearly everyone was the same: it didn't move our model. Silver explained that the data reached his model through the Times database, whose vetting he trusts. To its credit, FiftyPlusOne caught the fraud and admitted Wedgewood had fooled it too. But "our inputs come from someone else's vetting" is a familiar sentence. In my old business, we called it citing the source to share the blame.

The press. At the end of the chain, reporters inherit polls that have been herded, aggregated and partly outsourced. They then publish scenario pieces that cover every outcome. Each link can tell itself it acted reasonably. Together, they leave readers with no one who has actually said what they expect.

The exception: putting a number on it

In fairness, one link in the chain does take the risk. Silver's FLIPR model makes Democrats roughly 3-to-1 favorites to win the Senate. In his own voice, he goes further: Marist's Maine poll is indefensible, Hannah Pingree should win the governor's race comfortably, and he'd bet against FLIPR's own 70 percent odds for Troy Jackson. Those are judgments, with numbers attached, that can be checked on Wednesday morning.

Yes, Silver also reminds readers that a 3-in-4 favorite loses one time in four. Some will call that his own insurance policy. I don't. That is exactly what a likelihood is supposed to communicate. Kent would have recognized it. A forecaster who says 75 percent and loses hasn't necessarily been wrong. A forecaster who never says a number can't be.

The real test isn't whether anyone admits uncertainty. Everyone should. The test is whether they commit to a judgment despite it.

The scorecard

Before I grade anyone else, I owe readers my own record. On 17 October, in "The 3-Point Problem in Nate Silver's 16 October Senate Midterms Forecast," I showed that a polling miss the size of 2020 or 2024 would leave the Senate tied 50-50, with Vance breaking the tie. The Post's second scenario now makes the same point. But that piece was a warning, not a forecast. I showed how Democrats could lose without saying how likely I thought it was.

So here is my own estimate, on the record before the votes are counted. I judge that Democrats are likely to win the Senate, at about 60 percent, and that the most likely outcome is a narrow 51-seat majority. That's less confident than Silver's model, because Democrats' path still runs through rural states where polls have overstated them before, and through Maine, where Susan Collins beat her polls by a wide margin in 2020. If I'm wrong, the scorecard below will show it.

Here is what each source has actually committed to, on the record before the votes are counted. I'll fill in the last column after 3 November. Where a forecast updates daily, I'll grade the final version published before Election Day.

Notice which row I can't grade. Whatever happens, the Post was "right." That is the problem in a single line.

I'm not asking anyone to be certain. I spent years learning that certainty is usually a warning sign. I'm asking for what every analyst owes the people who read their work: a judgment, a likelihood, and the willingness to be found wrong.

Correction, 31 October: An earlier version described the 1st District as Jackson's home base. He is from Allagash, in the 2nd District. Thanks to a reader for the catch.

Sources

• Lenny Bronner and Eric Lau, "What if the polls are wrong about a Democratic Senate? Mapping 4 scenarios," Washington Post, 31 October 2026.

• What if the polls are wrong about a Democratic Senate? Mapping 4 scenarios. - The Washington Post https://www.washingtonpost.com/politics/2026/10/31/4-ways-polls-could-be-wrong-or-right-about-who-will-win-senate/?itid=deep_link&utm_source=webview&utm_medium=referral_top%20stories&utm_campaign=pb_2608%3Acontrol%3B

• Ian Ward, "What is the Cook Political Report, and why is Washington so obsessed with it?" Washington Post, 28 October 2026; Jennifer Agiesta, Ariel Edwards-Levy and Edward Wu, "Democrats in striking distance: New CNN polls in four states highlight party's path to Senate majority," CNN, 29 October 2026

• Nate Silver, "Fake polls are a real problem", Silver Bulletin, 22 October 2026

• Nate Silver, "The cognitive dissonance election," Silver Bulletin, 30 October 2026

• J. Mark Whaley, "The 3-Point Problem in Nate Silver's 16 October Senate Midterms Forecast", 17 October 2026

• J. Mark Whaley, "Mainers Vote Against Trump but Bet on Collins", 27 October 2026

• Axios: Election polling rocked by a new fake data scandal

• Mediaite: National polling company admits it's totally fake

• FiftyPlusOne: FiftyPlusOne just caught a pollster faking data

Bottom Line

Troy Jackson leads Susan Collins 49% to 47% among Maine likely voters in Marist's October 22–25 poll. That lead is statistically meaningless on its own, but the poll underneath it tells a more interesting story. Mainers dislike both candidates, they are voting largely against President Trump, and most of them still expect Collins to win. Whether Jackson's narrow lead holds depends heavily on an assumption Marist made about who will turn out, and the poll's own data give reasons to question that assumption.

A 2-Point Lead in a 13-Point Range

Marist surveyed 1,449 registered voters, but the headline number comes from likely voters, and that estimate is much less precise. Marist reports a margin of error of ±6.6 points for likely voters, against ±4.1 for registered voters. The reason is in the fine print: rather than screening out unlikely voters, Marist weights every respondent by their estimated chance of voting. That method is sound, but it costs precision. Although Marist interviewed 1,449 Mainers, its adjustments make the likely-voter result about as precise as a purely random sample of roughly 220 people.

The margin of error also applies to each candidate's share, not to the gap between them. The uncertainty on the lead is roughly double. A 2-point Jackson lead is consistent with anything from a comfortable Collins win to a comfortable Jackson win.

That does not make the poll useless. Marist is one of the most reliable pollsters in the business, and its result is in line with the other recent Maine polls: Beacon/Shaw and Trafalgar both show the race tied. Every credible measure says this race is a dead heat. The crosstabs are where this poll's value lies. They show why the race looks the way it does.

Unpopular but Leading

Jackson is leading without being liked. Only 39% of likely voters view him favorably, while 49% view him unfavorably and 11% are unsure. Collins is underwater too, at 44% favorable and 53% unfavorable. Yet Jackson wins 49% of the vote, 10 points more than his favorable rating, while Collins wins 47%, about 3 points above hers.

The gap is widest among independents, who make up 35% of Marist's likely electorate. Only 34% of them view Jackson favorably, but he wins them 50% to 41%.

The explanation is the president. Trump's job approval in Maine stands at 42%, with 58% disapproving, and intensity runs heavily against him: 53% strongly disapprove, while only 25% strongly approve. Asked about their vote for Congress, 41% call it mostly a vote against Trump, compared with 16% who call it a vote for him. Among independents the ratio is 46% to 9%.

Jackson's support, in other words, is anti-Trump first and pro-Jackson second. Collins, by contrast, outruns her party's leader. Only 32% of independents approve of Trump, but 41% vote for Collins. That 9-point margin is her crossover appeal, the same appeal that carried her in 2020.

Voting for Jackson, Betting on Collins

Asked who will actually win, regardless of whom they support, 58% of likely voters name Collins and only 40% name Jackson. Even 19% of Democrats expect Collins to win, as do 58% of independents.

The most telling number comes from Jackson's own stronghold. Maine's 1st District backs him 62% to 37%, yet only half of its voters expect him to win. In the 2nd District, 67% expect Collins to prevail.

Voters are not reading the race as a generic contest. In the governor's race, which is tied at 44%, 63% expect Democrat Hannah Pingree to win. Mainers treat an even race as a Democratic win, except when Collins is on the ballot. That is the residue of her history. Polls showed her trailing in 2020, and she won by more than 8 points.

Voters' expectations are not a forecast, but they may still matter. A Jackson voter who doubts he can win is a voter who might stay home.

The remaining persuasion is small. Both candidates' supporters are locked in: 68% of Jackson voters and 69% of Collins voters back them strongly, and fewer than 1% say they might switch. Only 3% are undecided. Democrats lead the generic congressional ballot 52% to 47%, which suggests most of the undecided are Democratic-leaning voters who have not warmed to Jackson. Collins draws exactly her party's 47%. Jackson's ceiling is about 52% if he can bring home his own party's voters.

The Turnout Question

The single biggest judgment in this poll is Marist's estimate of who will vote. Among registered voters, the sample is 36% Democratic, 31% Republican, and 30% independent. After Marist applies its likely-voter weights, Democrats fall to 31%, Republicans stay at 31%, and independents rise to 35%. Five points of Democrats effectively disappear.

That runs against the usual midterm pattern, in which the party out of the White House is the more motivated. It also sits uneasily beside the poll's own enthusiasm measures:

• Voting habit: 82% of Democrats say they always vote, compared with 80% of Republicans and 61% of independents.

• Sense of duty: 86% of Democrats call voting a duty rather than a choice, compared with 65% of Republicans.

• Campaign contact: 82% of Democrats have been contacted by a campaign, party, or civic group, compared with 69% of Republicans.

The partisan mix matters because Maine voters are almost perfectly loyal: Jackson and Collins each win 99% of their own party. The race therefore turns on how many Democrats, Republicans, and independents show up. If Democrats turn out at anything like the rate they claim, Jackson's position could be several points better than this poll shows.

Marist may well be right. People routinely overstate how likely they are to vote, and Marist's model probably draws on voters' actual turnout records, a sturdier guide than what they tell a pollster. Still, the gap between its turnout assumption and its own enthusiasm data is a question worth asking. That uncertainty matters more than the 2-point headline.

^{In most states, the oldest voters are the most Republican. Not in Maine, according to Marist.}

• ^{Voters 65 and older back Jackson 53% to 46%.}

• ^{Voters 45 to 64 back Collins 58% to 41%.}

• ^{Voters 30 to 44 back Jackson 58% to 40%.}

^{Generation X, not the Baby Boomers, is Collins' base, and it is also Trump's best generation in Maine, with 49% approval. The governor's race shows the same pattern. Subgroup samples are small, so treat the exact numbers with caution, but the shape is striking.}

^{What to Watch}

1. Democratic turnout in the 1st District. If Democrats vote at the rates they claim, Marist's model understates Jackson.

2. The 3% undecided. If they behave like the Democratic voters they resemble, Jackson approaches 52%. If they share their neighbors' doubts about him, the race stays a coin flip.

3. Collins' margin in the 2nd District. She leads there 55% to 38%. Anything less leaves her without a path.

Bottom Line

Maine voters are sending a message against Trump, and Jackson is the vehicle for it, not the reason. Collins remains personally unpopular, yet she holds her party's vote and keeps enough independents to stay within reach. Most Mainers think she will win again. Whether they are right depends less on persuasion than on turnout, and on whether Marist's assumptions about who votes match the electorate that actually shows up on November 3.

The old saying held that as Maine goes, so goes the nation. It has not been reliable for nearly a century, but this year it may come true in a narrower sense. Maine is one of the seats Democrats most likely need to win the Senate majority. If Mainers are betting wrong, the Maine Senate race could decide control of the chamber.

Source: Marist Poll of 1,449 Maine registered voters, conducted October 22–25, 2026. Likely-voter results carry a margin of error of ±6.6 percentage points.

The chart below measures first-day turnout against registered voters in the 19 counties for which figures have been released so far.

Statewide, 2.9 percent of registered voters cast a ballot on opening day. The chart also shows which counties went for Trump or Harris in 2024. In short:

• All six Harris counties beat the statewide rate at above 2.9 percent. Harris counties together yesterday came in at 3.2 percent vice 2.7 percent for Trump counties.

• The most heavily Republican counties were at the bottom. Saluda, Spartanburg, Lexington, Horry and Newberry all have the lowest rates. As a group, the counties Trump won by 30 points or more averaged 2.4 percent. Greenville was second only to Charleston in raw votes, but per registered voter it’s below average at 2.6 percent.

• However, some red counties are running hot. Clarendon, Aiken and Calhoun are near the top, all at 3.6 percent or higher.

• Harris counties also led on the first day in 2022, so this pattern isn’t new. But the gap between blue and red counties has widened this year.

One caution: since 2020, Democrats have been more likely than Republicans to vote early, so higher early turnout in blue counties may reflect habit rather than enthusiasm. The real test is whether red counties catch up later in the early-voting period or on Election Day.

It will be easier to see what’s going on in South Carolina early voting once statistics for all counties have been released.

Significantly, some of the increase is likely because voters have grown used to early voting, which first became available in a South Carolina general election in 2022.

But these predictions rests on a handful of races where Democrats lead by about three points. By shifting these polls three points toward the GOP candidates — about the degree of polling misses in both 2020 and 2024 — we’d see the Democrats end up with exactly 50 seats, meaning Vice President JD Vance would cast tie-breaking votes and Republicans would stay in control of the Senate.

What Silver's model says

On its face, the Silver’s forecast is good news for Democrats. His model rates the puts the average outcome at 52 Democratic seats, a net gain of five.

The near-certain gains and holds come first. Ossoff in Georgia (97 percent), Cooper in North Carolina (96 percent), and Pappas in New Hampshire (93 percent) are all heavy favorites. Michigan's open seat leans to El-Sayed at 81 percent.

Then comes the cluster that decides everything. Brown in Ohio (74 percent), Jackson in Maine (72 percent), and Talarico in Texas (71 percent) are each favored, but only modestly. Alaska (63 percent), Iowa (58 percent), and Kansas (51 percent) are closer to coin flips. Republicans are favored in Nebraska, where Pete Ricketts has a 70 percent chance against independent Dan Osborn, and Graham narrowly in South Carolina.

How Democrats got here

The polling trend lines tell a consistent story: Democratic candidates have held steady while Republicans have lost ground since late summer. The gains consistently come less from Democratic surges than from Republican erosion.

In Alaska, Mary Peltola has stayed near 48 percent while Dan Sullivan slid from about 47 to 44 since late September. In Ohio, Sherrod Brown has hovered around 48 percent for months while Jon Husted fell from 47 to a low near 43.5. In Kansas, Roger Marshall led by four points in August; Democrat Hamilton overtook him around October 1. Even Texas fits this pattern, with Ken Paxton unable to climb past about 46 percent.

South Carolina shows how unusual this cycle has become. Lindsey Graham died suddenly in July, and his sister Darline Graham, who had never held office, was appointed to the seat and then won the Republican nomination. She now leads Democrat Annie Andrews by under two points in the polling average, and Silver gives Andrews a one-in-four chance in a state Democrats last won in 1998.

The cold water: a 3-point problem

Most of the Democratic leads fall inside the range of a normal polling miss, and the misses of recent cycles have run in one direction — toward the GOP.

The pattern is hard to ignore. In Ohio in 2022, polls showed Tim Ryan close to JD Vance; Vance won by six. Texas polls had Ted Cruz up three or four in 2024; he won by about eight and a half. And in Maine in 2020, Susan Collins trailed in nearly every poll, then won by almost nine.

Now apply a uniform three-point shift toward Republicans to today's polling averages. North Carolina, Ohio, and Alaska stay Democratic, though Ohio's lead shrinks to about one point and Alaska's to under one. Michigan survives by less than half a point. Maine, Texas, Iowa, and Kansas all flip. Democrats finish with 50 seats, and Republicans keep control on Vance's tiebreaking vote.

One note on the numbers: if every current polling leader won, Democrats would reach 54 seats, higher than Silver's average of 52. Silver's figure is lower because it accounts for the real chance that some leaders lose anyway. Even the optimistic version, in which every leader wins, sits one polling miss away from a 50–50 Senate.

The math of control

Democrats need a net gain of four seats, and the fourth one sits at roughly a three-point lead. Rank the Republican-held seats by Democratic win chance and the order is North Carolina, Ohio, Maine, then Texas: the 51st seat. Alaska, Iowa, and Kansas are the cushion.

Silver’s tipping-point analysis makes the same point from another angle. No single race decides control. Maine and Texas each have a 12 percent chance of being the decisive seat, Ohio 11 percent, and Iowa and Michigan 10 percent each. Add Kansas and Alaska, and seven races share about 71 percent of the chance of deciding the Senate.

Silver’s Voter Power Index adds a human dimension. A single vote carries the most weight in Alaska (19.4) and Maine (12.2), where small electorates meet close races. Texas ranks near the top as a tipping point yet scores only 1.0, because millions of voters share that influence. In my home state of Nebraska, the index is 3.9: one vote in Lincoln counts roughly four times as much as one in Houston.

In fairness to Silver

None of this means Silver has missed anything. His 74.6 percent already accounts for polling error; that is what the ranges in his race-by-race charts represent, and why a three-point Alaska lead translates into only a 63 percent chance. The uniform three-point shift used here is also a simplification, since real polling misses vary from state to state.

The point is what his number means in practice. A 25 percent chance is not a long shot, and Silver knows this better than anyone. His final 2016 forecast gave Donald Trump about a 29 percent chance, and Trump won. Silver did a lot of explaining afterwards to make his point that an outcome shouldn’t have surprised anyone as a a more than one in four chance is going to happen every four or so times statistically

The Democrats' advantage, then, is real but conditional. It depends on the polls being roughly right in Ohio, Texas, and Maine, the very states where they have recently been wrong in the same direction.

What to watch

Three markers over the next two and a half weeks will show whether the forecast holds.

1. Texas polling. The two most recent heavily weighted polls both showed Paxton up a point. If that tightening continues, Texas leaves the Democratic column and the Dems lose their 51st seat.

2. South Carolina’s early vote. In-person early voting begins Monday, October 19. South Carolina doesn’t register voters by party, so the totals will show turnout levels, not who is ahead; the useful comparison is turnout in Republican-leaning versus Democratic-leaning counties.

3. Election night’s first signal. South Carolina’s polls close at 7 p.m. Eastern. If Andrews is running close there, the Democratic leads in Ohio, Texas, and Iowa are probably real. If Graham wins comfortably, settle in for a long night.

Silver’s model says Democrats are favored. The history of the last two cycles says the model’s favorites are standing on thinner ice than 74.6 percent suggests. Caveat qui praedictiones examinat.

Source: Silver Bulletin FLIPR 2026 Senate forecast, updated October 16, 2026.

I began my career as a CIA political analyst on the South Africa account during the apartheid era, where I was the “white politics” analyst. I often relied on polling data to gauge the mood of the white electorate, and in one key election I predicted the outcome in individual districts with 97.3% accuracy. Like all CIA analysts, I qualified my judgment in that President’s Daily Brief piece, but only with standard estimative language: “likely to,” “probably will,” “in our view.” I also included a map showing which party was likely to win each district.

Nate Silver’s Oct. 9 Silver Bulletin piece on Substack, “Could Minnesota spoil the blue wave?”, qualifies its conclusions far more heavily. Policymakers don’t have time for a dozen “on the other hands.” They need the bottom line up front. As I read Silver’s piece, I found myself translating his careful, heavily hedged analysis into the kind of judgment I used to write. Here it is.

Bottom line: Democrats are likely to win the Senate. Silver’s model gives them about a 75 percent chance. Republicans keep control about one time in four.

Key judgments:

• Democrats are almost certain to flip at least one Republican seat. North Carolina is close to a lock.

• A “double flip” year, with each party taking seats from the other, is possible but unlikely, at about 31 percent.

• If Democrats lose one of their own seats, the race becomes a coin toss. If they lose two, they’re almost certainly done.

• The seats to watch for a Republican upset are Michigan and Minnesota, but Silver expects Democrats to hold both.

• Most interestingly, Silver sees the Democrats’ most serious weaknesses not in swing states but in blue states like Minnesota, where voters believe Democrats have governed poorly.

The table below summarizes every competitive race. Bear in mind that this is only a snapshot. Forecasts move, and a 75 percent favorite still loses one time in four.

Data is from Nate Silver’s Silver Bulletin, as of Oct. 9, 2026. I recommend his full analysis on Substack for anyone who wants the details: https://substack.com/@natesilver/note/c-356942094

The Times treats comments the way a maître d' treats a dining room: a limited number of seats, a dress code, and a host deciding who gets in. The Post, at least until recently, has treated them more like a town square: open by default, policed for the worst behavior, but otherwise left to sort itself out. Both models are defensible. Both come with costs. And in the last two years, one of them has been tested in a way its architects never intended.

The Times: curation as house style

The Times has never tried to hide that its comment sections are curated, not democratic. For years, comments were open on only a small slice of the paper's daily output — historically around one in ten articles — because the Times insists on human moderation rather than automated filtering as the primary gate. A community desk of more than a dozen staffers reviews thousands of comments a day, reading for tone as much as for rule-breaking. The paper has experimented with machine-learning tools to help triage the volume, but it has resisted handing the actual judgment calls to an algorithm the way many outlets have.

On top of that moderation layer sits a curation layer: "NYT Picks," a badge awarded by staff to comments judged the most thoughtful, alongside a separate "Reader Picks" tab driven by upvotes. The effect is a comment section that reads less like an open debate and more like a letters page — a format the Times essentially invented and has been reluctant to abandon, digitally, in the underlying instinct.

That instinct produces a real structural side effect: because moderation happens in shifts concentrated on the East Coast workday, commenters in the Mountain and Pacific time zones routinely find threads closed, or their comments queued for review long after the conversation has moved on. It's not a policy anyone designed on purpose. It's a byproduct of treating comments as a small, hand-tended garden rather than an open field — and it reveals something true about the Times' priorities. Consistency of tone matters more than equal access. The comment section is meant to protect the authority of the article above it, not to compete with it.

The exclusionary quality of that system isn't only a matter of geography, either — it shows up in individual attempts to participate, too. This writer has submitted about half a dozen comments on Times articles over the years — real engagement with the reporting, not venting — and only one was ever approved; another was cut off entirely when the comment window closed before it could be submitted. The one approved comment only sharpened the contrast: a brief, offhand remark dashed off in seconds, about adult children quitting Facebook, sailed through with no friction, while the other, more substantive submissions sat rejected for reasons the Times never explains. Neither experience is evidence of bad faith; it's the predictable shape of a system built around a narrow, curated conversation. But it's a cost worth naming: a "quality-first" model filters out some of the very readers it claims to be selecting for, simply because the gate is narrow and the clock runs out.

The Post: openness as identity

The Washington Post has historically taken the opposite bet. By its own account, the paper offers comment sections on nearly all of its articles, moderated by a standing team rather than a small curatorial desk, with lighter editorial intervention and more tolerance for messy, contentious threads. Where the Times treats comments as content to be selected, the Post has tended to treat them as a feedback channel — noisier, less filtered, but closer to a genuine cross-section of what readers think.

This isn't incidental. It fits a broader identity the Post has cultivated as the paper of record for a city built on argument — a place where the comment section functions less like a curated salon and more like a Sunday-shows greenroom that never quite ends. Openness has been a competitive differentiator, a way of signaling that the Post wants engagement and debate rather than merely approval.

When the philosophy gets tested

That openness collided with something else entirely in 2025, and the collision is the most interesting thing to happen to either paper's comment section in years.

In October 2024, owner Jeff Bezos blocked the paper's editorial board from endorsing a presidential candidate, breaking decades of precedent — a decision that preceded roughly 300,000 subscription cancellations by Election Day. Two months later, in January 2025, Pulitzer Prize-winning editorial cartoonist Ann Telnaes resigned after editorial page editor David Shipley declined to run her sketch of Bezos and other tech and media executives kneeling before an incoming President Trump. Shipley said the call was about avoiding repetition with other scheduled content, not the cartoon's target; Telnaes wrote that in her time at the paper she had never before had a cartoon killed over who she'd aimed her pen at. Then, in February 2025, Bezos went further still: he directed the Opinion section to write only in defense of "personal liberties and free markets," explicitly stating that dissenting viewpoints on those subjects would be left to other publications. Shipley resigned from that role as well, within hours of the announcement. Longtime columnists and editors followed him out the door over the following weeks, several publicly, and subscriber cancellations continued climbing into the hundreds of thousands. That fall, opinion columnist Karen Attiah was fired after commenting publicly on the killing of conservative activist Charlie Kirk. A labor arbitrator ordered her reinstated nearly a year later, ruling that the Post "did not have good and sufficient cause" to terminate her — a sign that the pressure on dissenting opinion voices was being felt, and contested, well past the initial Opinion overhaul. Whether or not one agrees with Bezos's decisions, the sequence created a rare moment when the kind of dissent that Opinion and the cartoon desk could no longer freely accommodate found virtually its only remaining outlet in the Post's open-comment sections — accessible through the paper's website and its Android and iOS apps — inside the digital edition itself.

What makes this relevant to a piece about comment sections specifically is where readers went to register their fury: into the Post's own still-open comment threads. Reporting at the time put the volume of reader response across the publisher's note, related news stories, and opinion pieces at well over 100,000 comments — a real-time demonstration of exactly the "public square" dynamic the Post's open-comment philosophy was built to enable. The infrastructure did what it was designed to do. Readers showed up, argued, and were not gatekept out.

The Telnaes episode gave that dynamic a longer half-life than a single news cycle. The Post kept publishing editorial cartoons after her departure, cycling through several cartoonists rather than replacing her with one successor — and the comment threads under those cartoons have kept surfacing the same sentiment, regardless of which cartoonist or subject is on that day's page. One representative example: a September 2026 thread under a cartoon on an unrelated, non-political subject carried a comment reading simply "I really miss Ann Telnaes…", quickly drew 33 upvotes and multiple reply threads — ordinary engagement numbers for that comment section, but notable for how detached the sentiment was from the cartoon it was ostensibly responding to.

Screenshot by author, 4 October 2026

The sentiment isn't confined to that one thread, either. Roughly two weeks earlier, in a comment thread under a political cartoon about the Trump administration, a different reader wrote "Bring back Ann Telnaes!!" and drew 29 upvotes; two weeks apart, on two different cartoons by two different hands, on two different subjects — one political, one not — separate readers reached for the same two words. The absence, not any particular successor, is what keeps drawing comment. That's a small but pointed data point: readers are using the Post's own open-comment architecture to protest a decision the architecture itself had nothing to do with, which is exactly what an open forum is for.

There's a symbolic layer to this, too. The Post's masthead motto, "Democracy Dies in Darkness," predates Bezos's tightening of the Opinion pages by eight years, and despite widespread reporting in early 2025 that the paper was retiring it, company spokespeople insisted at the time that it remained the paper's public-facing tagline — it still anchors the Post's official social media presence today. But its visibility in the place that matters most for this argument — the digital editions where readers actually encounter stories and submit comments — has quietly receded. Opening the Post's Android app today surfaces no version of the motto anywhere on the front screen; the masthead reads simply "The Washington Post," full stop.

J. Mark Whaley served as a senior analyst on African politics in the U.S. intelligence community. He holds a master’s degree from the Fletcher School, where he specialized in diplomatic history and strategic studies.

Android Washington Post app w/o motto. Screenshot by author, 4 October 2026.

Screenshot of Android Washington Post app showing lack of motto taken by author 4 October 2026

Whatever the company's official position on the slogan, the digital doorway through which most readers now enter the paper has stopped saying it out loud. What changed alongside that fade was a separate, internal-only mission statement, "Riveting Storytelling for All of America," explicitly framed as a staff rallying cry rather than a public replacement. That combination — a motto quietly disappearing from the app's front door while an anodyne internal slogan takes its place backstage — left an irony sitting in plain view, and commenters did not let it pass. On the paper's own coverage of the endorsement decision, one heavily upvoted comment mockingly turned the masthead motto back on its owner, treating "democracy dies in darkness" as a description of Bezos's conduct rather than the warning he'd once championed. For a moment, the open comment section became the one place inside the institution where the gap between the Post's stated values and its owner's actions could be pointed out publicly, under the paper's own logo, without an editor's sign-off.

But the episode also exposes the tension sitting underneath that openness. The same owner who left the comment sections structurally untouched was simultaneously narrowing, by fiat, what viewpoints the paper's own opinion writers were permitted to publish. A comment section can be a genuine public square only as long as the institution around it tolerates dissent elsewhere too. Bezos's intervention didn't touch the comment infrastructure directly, but it raised an obvious question for anyone watching closely: how much longer does an ownership structure willing to police its own columnists leave the reader threads alone? The Post's comment policy, in other words, may currently be less a stable philosophical commitment than a holdover from an earlier era of the paper — one that hasn't yet been reconciled with the more centralized, owner-driven direction the rest of the institution is visibly taking.

Two philosophies, two kinds of risk

Set side by side, the two approaches carry mirror-image risks. The Times' curated model protects tone at the cost of access — it is, by design, exclusionary, and its time-zone blind spot is a small but telling artifact of an editorial culture built around a New York workday. The Post's open model protects access at the cost of tone — it is, by design, more chaotic, and it depends on an institutional tolerance for disagreement that is no longer obviously guaranteed from the top.

Neither newspaper frames its comment policy as a philosophical statement. Both would probably describe it, if asked, as a practical response to moderation capacity and audience strategy. But practical choices accumulate into worldview. A newspaper that opens its comments to almost everyone, almost always, is telling its readers something about how much it trusts them and how little it fears them. A newspaper that opens its comments rarely, and curates heavily when it does, is telling its readers something else: that the institution's voice comes first, and theirs is a privilege extended selectively.

The real test of either philosophy isn't the ordinary day. It's the moment a paper's readers are angriest at the paper itself. The Times has rarely faced that test in its own comment sections, because so few of its comment threads are open long enough, or broadly enough, for a mass reader revolt to gather there in the first place. The Post faced exactly that test in February 2025 — and for now, at least, the square held. Whether it stays open is a different question, and one that has less to do with comment-moderation policy than with how much dissent the paper's owner is willing to tolerate everywhere else.

In the Times' salon, the maître d' decides when the conversation ends. In the Post's square, the crowd does — at least for now.

I've spent much of the past year documenting a subtler version of this same instinct — one that never required an announcement at all.

When an ABC News correspondent asked Saudi Crown Prince Mohammed bin Salman about the 2018 murder of Jamal Khashoggi last November, the exchange lasted only seconds. Trump publicly attacked ABC shortly after it aired. Twenty-four hours later, the FCC — under Trump-appointed chairman Brendan Carr — opened a review of the legal ties between national networks and their local affiliates, ties that determine whether millions of Americans receive network news at all. I submitted an op-ed to the New York Times at the time arguing the sequence fit a recognizable pattern: subtle, deniable, and potentially dangerous. The Times passed.

The pattern repeated twice more. In January, the same FCC reversed a rule letting talk shows interview single candidates without triggering "equal time" obligations. In February, CBS lawyers told Stephen Colbert he couldn't air a planned interview with Senate candidate James Talarico; he aired his objection instead and posted the interview to YouTube. Seven months later, the identical sequence played out at ABC, with the identical candidate — Kimmel taped Talarico, then told his audience the segment wouldn't air, citing FCC pressure. It went to YouTube too, where it's now surpassed 12 million views. ABC is the network that broadcast the Khashoggi question and drew the affiliate review within a day. It's also the network that pulled the Talarico interview ten months later. Two different regulatory levers, one network on the receiving end of both, the same chairman overseeing both.

None of that required a public announcement. That was the point — a regulator doesn't need to coordinate anything to produce a chilling effect. It only needs to make the cost of certain journalism, or certain bookings, unpredictable enough that networks route around the risk themselves.

Today's ban is different in kind. There's no affiliate review to interpret, no equal-time statute to parse, no ambiguity to leave outlets guessing. It's a president simply declaring three of the largest newsrooms in the country unwelcome at the seat of government, over coverage he doesn't like. The indirect version relied on institutions internalizing pressure quietly. This version doesn't bother asking them to.

Eleven months ago, I tried to warn about one such signal, in an op-ed a major newspaper decided wasn't worth running. It has since repeated at two networks, against one candidate, through two different regulatory mechanisms — and now, out in the open, at three more. I would like to be wrong about where this leads. The track record so far does not encourage that hope.

Donald Trump fell squarely into the second category. Shortly after the question aired, he publicly attacked ABC. Then something curious happened. Only 24 hours later, the Federal Communications Commission — under Brendan Carr, the Trump-appointed chairman who has led the agency since early 2025 — announced it was opening a review of the legal and contractual ties between national networks and their local affiliates, relationships that determine whether millions of Americans can receive network news at all.

I submitted an op-ed to the New York Times at the time, arguing the sequence fit a recognizable pattern: subtle, deniable, and potentially dangerous. The Times passed on it. I'm returning to the argument now because the pattern has since repeated — twice.

In January 2026, the same Carr-led FCC reversed a long-standing rule allowing late-night and daytime talk shows to interview single candidates without triggering "equal time" obligations to every other name on the ballot. Kimmel later noted the irony: he'd interviewed candidate Donald Trump in 2016 under that same exemption without incident.

The rule change found its first target in February 2026, when CBS lawyers told Stephen Colbert his network could not broadcast a planned interview with James Talarico, the Texas state representative then running for the U.S. Senate against Republican Ken Paxton. Colbert called the legal guidance unprecedented in his two decades on the show, aired his objection anyway, and posted the interview to YouTube. Seven months later, the identical sequence played out again — this time at ABC, with the identical candidate. Kimmel taped an interview with Talarico, then told his television audience the segment would not air, citing FCC pressure on the network and its affiliates.

The interview went to YouTube instead, where it has now surpassed 12 million views — nearly four times Kimmel's best-ever monthly average of 3.15 million nightly viewers. The White House and Carr disputed Kimmel's account as "play-acting"; Carr maintained the equal-time statute is "about more speech, not less." Whether or not any explicit threat was issued, ABC's lawyers reached the same conclusion CBS's had seven months earlier, against the same candidate, under the same reinterpreted rule.

ABC is not a bystander here. It is the network that broadcast the Khashoggi question and drew the FCC's affiliate-contract review within a day. Ten months later, it is also the network that pulled a candidate interview from air rather than risk the FCC's reinterpreted equal-time rule. Two different regulatory levers, one network on the receiving end of both — and the same chairman overseeing both.

None of this requires believing in a coordinated conspiracy to be alarming. Two separate mechanisms have now produced the same outcome twice, at two networks: journalism that would once have aired on broadcast television migrating instead to YouTube, where the FCC has no jurisdiction at all. A regulator does not need to coordinate anything to produce a chilling effect. It only needs to make the cost of certain journalism, or certain bookings, unpredictable enough that networks route around the risk themselves.

Political pressure rarely announces itself. It operates indirectly, through signals that institutions learn to interpret. Over time, they internalize the boundaries. The danger is not simply censorship; it is self-censorship — a quiet recalibration of what questions seem worth asking, and which guests seem worth booking.

Jamal Khashoggi's murder was a stark reminder of the cost of challenging power where the press is not free. The United States is not Saudi Arabia. But a democracy's commitment to free inquiry is not guaranteed by tradition alone. It requires noticing when levers of state power are deployed — or appear to be deployed — in response to legitimate journalism, especially when each instance is small and deniable enough to dismiss on its own.

Democracies weaken not through a single dramatic rupture but through a slow accumulation of signals that narrow the space for scrutiny. Eleven months ago, I tried to warn about one such signal, in an op-ed a major newspaper decided wasn't worth running. It has since repeated at two networks, against one candidate, through two different regulatory mechanisms. I would like to be wrong about where this leads. The track record so far does not encourage that hope.