Two things are clear from Israel’s election returns Wednesday morning. Benjamin Netanyahu’s bloc leads. And nobody has 61.

With about 76 percent of ballots counted, the Netanyahu-led coalition of right-wing and ultra-Orthodox parties stood at approximately 54 seats, according to the Central Elections Committee’s running tally. Gadi Eisenkot’s Yashar-led bloc was at about 47 seats. Arab-led parties were tracking toward 19 seats. A coalition needs 61 seats in the 120-member Knesset to form a government. Netanyahu’s bloc is 7 seats short of that threshold. Eisenkot’s is 14 seats short. Official certified results aren’t due until November 4.

The exit polls that ran from Tuesday night through early Wednesday gave a different picture, and not a consistent one. Channel 12 had Netanyahu’s bloc at 52 and Eisenkot’s at 49. Channels 11 and 13 both showed them tied at 51 apiece. The partial count diverged from all of them. Eisenkot said Tuesday night that “only exit polls” were in and the country needed to wait for results. He hasn’t declared victory. Netanyahu hasn’t either. The Arab parties, whose 19 projected seats make them a potential kingmaker in any coalition math, haven’t committed to either bloc. Democrats leader Yair Golan called explicitly for a government that includes Arab-led party participation, a formulation that would exclude Netanyahu. Whether that combination can hold 61 is a math problem that coalition negotiators will be working for weeks.

Israel has been here before, and recently. Between April 2019 and March 2021, the country held four elections in less than two years because neither bloc could reach 61 and form a stable government. The first ended in a 35-35 tie between Likud and Blue and White and produced nothing. The third, in March 2020, was nearly tied. The fourth finally produced a functioning government after six months of caretaker status. The difference this time is that Netanyahu is asking voters to return him after a war his government fought following October 7, and the election is the first Israeli national vote since the Iran war began in February. What that context means for coalition negotiations, nobody has said publicly yet. (The Times of Israel is tracking the live count as ballots are tabulated.)

Brazil’s presidential runoff is certified. The legal challenge has begun.

Luiz Inácio Lula da Silva won his fourth term on Sunday, defeating Flávio Bolsonaro by approximately 4.4 million votes, with 100 percent of the vote counted. Brazil’s Superior Electoral Court confirmed the result. Lula’s fourth term begins January 5, 2027. Flávio Bolsonaro formally rejected the result on Monday and is pursuing a legal challenge. The stated basis is a Supreme Court justice’s decision to extend voting hours by two hours on election day, a decision Bolsonaro’s camp says was made without proper authority and affected turnout in regions favorable to Lula.

In 2022, Jair Bolsonaro, Flávio’s father, also rejected a runoff loss to Lula. His party filed a challenge with Brazil’s electoral court arguing that electronic voting machines used before 2020 should be invalidated. The court rejected that challenge the next day. The elder Bolsonaro eventually authorized his chief of staff to begin the transition process rather than conceding directly, and left for the United States before the inauguration. He is currently serving a 27-year sentence for his role in the January 2023 attempted coup, held under house arrest for health reasons since March. The pattern from 2022 doesn’t guarantee a resolution this time, but it’s the nearest precedent for what happens next. (CBS News reported Bolsonaro’s rejection statement; NPR confirmed the certified result.)

Four Salvadoran aviation technicians disappeared on October 18 after a gang ambush near Port-au-Prince. They weren’t soldiers. They were aviation support personnel embedded with the Gang Suppression Force, the UN-authorized mission operating in Haiti to suppress gang activity, and their role was supporting medical evacuations. The ambush hit their convoy. No group has claimed the attack. No bodies have been recovered. Their status remains unknown.

President Nayib Bukele announced a military rescue operation on October 21 and has deployed special forces to Haiti. Troop counts in press accounts vary: the Wall Street Journal reported approximately 200 soldiers, other outlets have cited figures between 100 and 450, and El Salvador’s government has not published an official number. Bukele has warned publicly that a failure to release the men alive will produce a more forceful response. The GSF, which has more than 2,000 troops on the ground from Kenya, El Salvador, Jamaica, Guatemala, Chad, and other contributing nations, is operating under a six-month mandate extension approved on September 29. It was authorized by UN Security Council Resolution 2793 in September 2025. The Salvadoran rescue operation is the most direct military confrontation with Haitian gangs any contributing nation has undertaken since the mission began. (Al Jazeera reported on the missing soldiers and Bukele’s deployment.)

Six days to November 3. The Senate forecast hasn’t settled.

Dave Manuel’s simulation model, updated through October 24, gives Democrats a 94.7 percent chance in the House and a 62.1 percent chance in the Senate. Polymarket puts Democrats at 52 percent for the Senate. Silver Bulletin’s FLIPR model ran as high as 73.9 percent for Democrats. That’s a roughly 22-point spread between the leading Senate models, which reflects genuine uncertainty in races that remain genuinely uncertain.

Nearly every modern midterm has cost the president’s party seats in the House. In 2022, Democrats lost 9 House seats. In 2018, Republicans lost 40. In 2010, Democrats lost 63. In 2006, Republicans lost 30, which ended their twelve-year House majority and brought Nancy Pelosi the speakership. The question for November 3 isn’t whether the historical pattern holds, since it almost always does, it’s whether it holds by 3 seats or 30.

The Iran diplomatic channel is still open. Iran’s Foreign Ministry says messages are passing through third-party intermediaries. No deal has been reached. The war is in its eighth month. The Pentagon submitted a supplemental spending request of $67.1 billion to Congress in June. Its most recent cost estimate to Congress, delivered in late September, put the total at $45.1 billion. The Congressional Budget Office has estimated the war is adding roughly $2 to $3 billion to that total every month. Brent crude held above $100 per barrel through Tuesday. What changes in the diplomatic calculus on November 4 is a question nobody in Washington has answered on the record. (Decision Desk HQ’s Senate forecast updates continuously through Election Day.)

The ceasefire window in Oleshky is open.

Ukraine’s General Staff and Russia’s military both announced a halt in hostilities starting at 3 a.m. on October 27, Kyiv time, in a designated corridor around the city. The White House, in its original announcement, said the evacuation operation would begin October 29. Ukrainian ombudsman Dmytro Lubinets confirmed after talks with Russian negotiators that the operation would proceed in stages, with about 1,700 to 2,000 residents, most of them elderly, plus approximately 50 children, able to choose evacuation toward Ukrainian-controlled or Russian-held territory. The city has had no reliable electricity for months. Aid delivery is part of the arrangement.

The broader war has no ceasefire. The fighting elsewhere hasn’t stopped. But in Oleshky, in Kherson Oblast, the two military commands have agreed on a specific corridor and a specific purpose. Engineers surveyed the route. Convoys are either moving or about to move. That’s not a peace agreement. It’s one city, one window, and a few days to get people out. Whether it holds is something the morning news will tell us. (Ukraine’s 24tv.ua confirmed the US evacuation announcement and the October 29 start date.)

Something good.

On Monday, OpenAI announced that its AI systems contributed to first-ever proofs in three areas of mathematics and science: harmonic analysis, complexity theory, and quantum field theory. Among the specific results was a proof of the local smoothing conjecture across every spatial dimension, a problem mathematicians have worked on since at least the early 1990s.

These problems aren’t engineering challenges. They don’t produce a product or solve an immediate practical need. They’re questions about the structure of things at a level of abstraction that takes decades of specialized training to approach. A proof matters in a specific way: once a conjecture is proved, it joins the permanent record of what is known. It doesn’t change. It isn’t revised. A problem that was open last week is closed this week. Mathematicians have debated whether results like these could come from outside the human research community. Monday’s announcement says some of them can. (Al Jazeera reported on the OpenAI announcement on October 27, 2026.)

Howard Fenn writes The Day, Monday through Saturday.