# Israel Votes, Part 4: The Forecast

My Weird Prompts — an AI-generated podcast. Daniel Rosehill sends in a prompt; AI hosts Corn and Herman discuss it. Verify important claims independently.

- Episode: 6073
- Published: 2026-10-17
- Duration: 30:18
- Audio: https://episodes.myweirdprompts.com/audio/israel-election-2026-forecast.m4a
- Tags: israel, political-history, geopolitics
- Web page: https://myweirdprompts.com/episode/israel-election-2026-forecast/
- JSON: https://myweirdprompts.com/api/episodes/israel-election-2026-forecast.json

## Summary

Ten days before Israel's 27 October 2026 election, Corn, Hilbert and Herman each forecast it a different way. Corn uses a structural model: how stable the blocs are, which lists fell under the threshold, and what happened after 1973. Hilbert averages nine mainstream polls, corrects them for Israel's past polling errors and simulates the threshold races. Herman builds a synthetic electorate of seven voters, from Beit Shemesh to Nazareth, and runs it through the real seat allocation ten thousand times. Then Herman judges the three and makes a forecast: seats, blocs, and who becomes prime minister.

## The prompt

> Israel 2026 election predictions. Herman, Corn and Hilbert each present a legitimate predictive framework.

## Show notes

Part four of Israel Votes 2026, and the last. Daniel asked for a real prediction, built from three legitimate forecasting methods with Herman as the judge. The script was written by hand, offline, from published polls up to 15 October 2026.

**Corn: the structural model.** Netanyahu's bloc won 60, 55, 58, 52 and 64 seats in the five elections from 2019 to 2022. The swings came from turnout, threshold casualties and a small group of soft-right voters, not from people changing sides. Examples: New Right on 3.22% in April 2019; Meretz on 3.16% and Balad on 2.90% in 2022. The 1973 and 1977 elections show the punishment for a security failure arriving late. Estimate: the Netanyahu bloc at about 53.

**Hilbert: corrected polls.** He averages the nine mainstream polls taken between 4 and 15 October. Yashar is on about 22 seats, Likud on just under 20, and the outgoing coalition on 48.5. Channel 14, with the coalition on 61, gets little weight: it was the closest pollster in 2022, but its lean then was 2 seats and now it is 12. He makes three corrections: the right was under-polled in 2022, new lists faded in 2021, and Arab turnout carries extra uncertainty. His threshold simulation gives Religious Zionism about a 95% chance of getting in, Reservists–New Economic about 60%, Amcha Yisrael about 33%, and Blue and White essentially none.

**Herman: the synthetic electorate.** It works like a simplified MRP model, built on seven persona segments. Each has a weight, a turnout rate and a vote split. They are run through the 3.25% threshold, surplus agreements and Bader–Ofer seat allocation, 10,000 times with added noise. The central run gives Yashar 20, Likud 20 and the coalition 51. The Netanyahu bloc reached 61 in essentially no runs. Opposition plus Ra'am reached 61 in about 46% of them.

**The verdict.** Yashar 21, Likud 20, Together 13, Shas 10, Democrats 9, Yisrael Beytenu, UTJ, Otzma Yehudit and the Joint List 8 each, Ra'am 6, Religious Zionism 5, Reservists 4. Amcha Yisrael and Blue and White miss the threshold. Blocs: coalition 51, Jewish opposition 55, Arab lists 14. Next prime minister: Eisenkot about 45%, Netanyahu about 30%, Bennett about 15%. Chance of another election within a year: about one in four. For comparison, Polymarket on 17 October: Eisenkot 52%, Netanyahu 28%, Bennett 12%.

This is a forecast, not voting advice. Official results are certified on 4 November.

## Transcript

Corn: Ten days out from Israel's election, Daniel has sent one more request for this series. We told him who's running, what they stand for, and how we'd vote in his shoes. Now he wants a prediction. An actual one. Seats, blocs, who ends up prime minister.
Herman: And he's asked for it a particular way, which I love. Three forecasting methods, all legitimate ones, one each. Corn gets one, Hilbert gets one, I get one. Then I have to be the judge. I weigh them up and put a single number on the table that I'm prepared to be embarrassed by on the twenty-eighth.
Corn: You're always prepared to be embarrassed.
Herman: That's what makes me qualified. Before we start, a quick note. Like our last pipeline episode, this script was written by hand, offline, not by the usual model. Every poll number you'll hear comes from published surveys up to the fifteenth of October. The seat model Herman describes in the third part was actually run, so those numbers are the model's output, not a mood.
Hilbert: Hilbert Flumingtop, producer. I've been told I'm a forecasting method. I've never been called a method before. I've been called a process, by an accountant, and he didn't mean it kindly.
Corn: The order is slowest first. That's me. Then Hilbert, who distrusts everything, including the polls. Then Herman and his imaginary voters.
Herman: Synthetic electors. They're not imaginary, they're modelled.
Hilbert: They have name badges. He printed name badges for people who don't exist. There are seven of them on the fridge.
Herman: It helps me think.
Corn: Right. My method. Political scientists call it a fundamentals model, or a structural model. It ignores this week's poll and asks what the slow-moving facts say. In Israel the most important slow-moving fact is that people hardly ever change sides.
Herman: The blocs.
Corn: The blocs. Since twenty nineteen, Israel has had five elections that were really referendums on one man. Look at what Netanyahu's bloc won each time: Likud, the Haredi parties and the religious right, counted the same way each time. April twenty nineteen, sixty. September twenty nineteen, fifty-five. March twenty twenty, fifty-eight. March twenty twenty-one, fifty-two. November twenty twenty-two, sixty-four.
Herman: So a twelve-seat range across five elections.
Corn: Twelve seats. And very little of that movement was people switching from left to right. It came from three places. Who turned out, who fell below the threshold, and a small group in the middle that drifted between Likud and whatever centre-right list was fashionable that year.
Hilbert: The threshold killed more governments than the voters did.
Corn: Pretty much. April twenty nineteen, Naftali Bennett's New Right got three point two two percent. The threshold is three point two five. He missed by a few thousand votes and the right lost about four seats it would otherwise have had. Twenty twenty-two, the other side: Meretz got three point one six, Balad got two point nine. That's close to three hundred thousand votes for the left and the Arab lists with nothing to show for them, and that's a large part of how Netanyahu got to sixty-four.
Herman: With nowhere near sixty-four seats' worth of votes, to be clear. His bloc took a bit under half the vote, and the wasted ballots on the other side did the rest.
Corn: So the first structural rule is that the bloc's ceiling and floor are set by arithmetic at the bottom of the table. The second is the one I actually care about this time, which is what happens to a government that presided over a security catastrophe.
Herman: Nineteen seventy-three.
Corn: The only real precedent Israel has. The Yom Kippur War starts in October, and the election, which had been postponed, takes place on the thirty-first of December. Golda Meir's Alignment drops from fifty-six seats to fifty-one. Still the largest party, still in government. Then the interim report of the Agranat Commission comes out in April, and Golda resigns within days.
Hilbert: She resigned. People used to do that.
Corn: Then in nineteen seventy-seven, four years after the war, the Alignment collapses to thirty-two seats, Begin's Likud wins forty-three, and that's the upheaval, the mahapach. The election straight after the war was a mild punishment. The full punishment came years later, once the shock had become a verdict.
Herman: And today we're three years out from October seventh.
Corn: Exactly. Fundamentals models care about time. Three years is about when you'd expect a verdict to replace a shock. That points against the incumbent. But there's a twist, and it's why I don't just call it for the opposition. In nineteen seventy-seven the voters punished a left-wing government by moving right. In twenty twenty-six there's nowhere further right for a punished right-wing government's voters to go that isn't inside its own bloc.
Herman: Unless the opposition is the right.
Corn: Which this time it partly is. Eisenkot, a former chief of staff. Bennett, who used to lead the national religious party. Liberman. The most important structural fact of this election is that, for the first time, a soft-right Likud voter can punish Netanyahu without voting for anything that feels left to them. In twenty twenty-two that voter had nowhere to go. Now there are three exits, all wearing a uniform or a kippah.
Hilbert: Or both. There's a general with a kippah on the right as well. Winter.
Corn: Amcha Yisrael, and he's the one exit that leads straight back to Netanyahu. Okay. The structural verdict. Starting from the bloc's long-run average of roughly fifty-eight, I take off four or five seats for the accountability effect and the draft fight. The Haredi parties brought the government down over draft exemptions, and that's the issue that splits the right down the middle. Then I give back one or two, because Israeli right-wing voters have historically been under-counted rather than over-counted. That lands me at about fifty-three for the Netanyahu bloc. Below sixty-one. Above where the mainstream polls have it.
Herman: Fifty-three. Noted.
Corn: And one more thing my model is good at, because it's slow. It says the result decides the opening hand, and the game takes months. When no bloc reaches sixty-one, Israel doesn't get a government on the twenty-eighth. It gets a negotiation, and the negotiation takes as long as it takes.
Hilbert: Like Corn.
Corn: Like me. That's my time. Hilbert.
Hilbert: Mine is the boring one, so I'll be quick. I take the polls. Then I correct them for the ways Israeli polls have been wrong before. Then I look at the bottom of the table, because Corn's right, that's where it's decided. That's the method. Polls, track record, threshold.
Herman: That's a real method. It's roughly what the professional aggregators do.
Hilbert: I didn't ask whether it was real. Nine polls, fourth to fifteenth of October. Channel Twelve, Channel Thirteen, Kan, Maariv, Walla, Channel Sixteen, Zman Yisrael. I've left out two, and I'll get to them. Average across the nine: Yashar twenty-two. Likud a shade under twenty. Bennett and Lapid's joint list, Together, twelve and a half. Golan's Democrats, Liberman, Ben-Gvir, the Joint List and United Torah Judaism all around eight or nine. Shas seven. Smotrich six. Ra'am five. Hendel's Reservists four, in eight polls out of nine, which is either very stable or very suspicious. The outgoing coalition averages forty-eight and a half.
Corn: And Winter?
Hilbert: Amcha Yisrael is over the line in four of the nine. In the others it's somewhere between one and a half and three percent. So it might be four seats or none. That's the most important line in my notebook, and I'll come back to it.
Corn: The two polls you left out.
Hilbert: Channel Fourteen and Direct Polls for i24. Channel Fourteen has Likud at twenty-nine and the coalition at exactly sixty-one. Direct Polls has Likud twenty-eight and the coalition fifty-six. Everyone else has Likud around twenty.
Herman: That's a nine-seat gap on one party.
Hilbert: And here's the bit I don't enjoy saying. In twenty twenty-two, Channel Fourteen was closer than the others. Final polls from the main channels had Netanyahu's bloc at sixty or sixty-one. Channel Fourteen said sixty-two. The real number was sixty-four.
Corn: So you can't just bin them.
Hilbert: I can't bin them. I wanted to. But their lean four years ago was two seats, and they were nearer the result than anyone. Their lean today is twelve seats on the bloc. A pollster that's a couple of seats friendlier than everyone else can be right. A pollster that's twelve seats friendlier is describing a different country. So I give it a little weight, not none.
Herman: That's exactly how a house effect should be treated.
Hilbert: Then the track record. Three corrections. One. In twenty twenty-two the polls under-counted the right by three or four seats, and Shas alone by two or three. Haredi and traditional voters turn out on the day better than they answer the phone. So I add two seats to the coalition. Not four, because Netanyahu has spent this campaign attacking the Haredi parties to win back moderates. That may cost him some of the turnout he relied on.
Corn: Two.
Hilbert: Two. Correction two. New parties disappoint. In twenty twenty-one, the polls had Gideon Sa'ar's New Hope at seven to ten, and he got six. Yamina was polled at eight to ten and got seven. Shiny new lists poll well because people like saying the new name to a stranger on the phone. Then they get into the booth and pick up the old slip. Yashar is a new list. So I take a seat or two off Eisenkot.
Herman: Together is newish too.
Hilbert: Together has Yesh Atid's machinery inside it. I've taken one off. Correction three is turnout among Arab citizens. In twenty twenty-two some pollsters expected it as low as thirty-nine percent. It came in around fifty-three. Polls in this country have missed Arab turnout in both directions, so I don't move the number. I just widen the error. The Arab lists could be twelve seats or sixteen, and that alone decides whether the opposition can reach sixty-one with Ra'am.
Corn: So what does Hilbert's corrected poll say?
Hilbert: Yashar twenty-one, Likud twenty-one. A tie for first. The coalition at fifty, fifty-one, and fifty-four or fifty-five if Winter gets in. Nobody at sixty-one. And now the bottom of the table, where I've been doing real work.
Corn: Is this the board?
Hilbert: It's the board. Thirty-eight hand tally counters on plywood, plus one for blanks. Avraham Pinto's nephew varnished it, it dried, I've got my bathtub back, and I've been using it as a threshold simulator.
Herman: How does a plank of tally counters simulate a threshold?
Hilbert: It doesn't. I sit in front of it with a spreadsheet, because I needed somewhere to sit. There are four lists that matter at the bottom. For each one I took its poll average as a share of the vote, and allowed it to miss by about a fifth of that either way, which is roughly how badly small lists have been missed here before. Then I asked the spreadsheet ten thousand times whether each one gets over three point two five.
Corn: And?
Hilbert: Hendel's Reservists and the New Economic Party are over in about three runs out of five. Smotrich's list almost always gets in, about nineteen in twenty. It looks wobbly, but it's polling six, and six is a long way above four. Winter's Amcha Yisrael gets in about one time in three. Gantz's Blue and White is gone. He's polling around one percent. Two years ago the man was polling in the thirties.
Herman: And Gantz has said he'd stop campaigning if he isn't over the line in the final week.
Hilbert: Which is next week. It's now. If he drops out, that's one percent of the vote looking for a home. Most of it goes to Yashar, because Eisenkot left with all Gantz's furniture.
Corn: You haven't mentioned the market.
Hilbert: I have, quietly, to myself. Polymarket, this morning. Next prime minister, Eisenkot fifty-two percent, Netanyahu twenty-eight, Bennett twelve. Forty-seven million dollars traded.
Herman: That's a lot of money.
Hilbert: It's a lot of money from people who mostly can't vote in this election, betting on a country many of them couldn't find without zooming in. I'd sooner use Ruthie.
Corn: Your cousin Ruthie. The broker.
Hilbert: She votes for the opposite of whatever I'm voting for, every election, to cancel me out. She keeps a ledger. Her ex-husband Moshe has to sign it. She worked out who I'm voting for by watching my face while the news was on. She's usually right about my face, so she's an extremely reliable contrarian indicator. This year she's voting for a list that is in fact polling, if I'm honest, quite well.
Corn: So Ruthie has the same forecast as the polls.
Hilbert: Ruthie's forecast is that I'll lose. That's been accurate eleven times running. My time's up.
Herman: That's excellent. Genuinely. The house effect, the new-party penalty, the threshold simulation. That's a proper forecaster's method, Hilbert.
Hilbert: I know what it is.
Corn: Herman. The people on the fridge.
Herman: Right. My method is called a synthetic electorate. It's a cousin of something pollsters call MRP: multilevel regression and post-stratification. The idea is that you don't model the country as one big mass. You break the electorate into groups of people who think alike, work out how big each group is and how likely it is to turn up, then how it splits its vote, and you add it up. I've simplified that into seven people. Each one stands for a slice of the electorate, and they're all on the fridge.
Hilbert: With badges.
Herman: With badges. Each one has a weight, the share of eligible voters they represent, a turnout probability, and a vote split. The splits come from the polls, from how these communities voted in twenty twenty-two, and from what each group is actually arguing about this year. Then I ran the numbers, ran them again ten thousand times with noise added, and allocated seats the way the Central Elections Committee does, threshold, surplus agreements and all.
Corn: Introduce them.
Herman: Elector one is Yisroel, from Beit Shemesh. He stands for the Haredi community, about eleven percent of eligible voters. He studied in kollel for years and now works at a printing press. His turnout is the highest of anyone: I have him at eighty percent. His vote splits United Torah Judaism and Shas, and almost nothing else. The question for Yisroel isn't who to vote for. It's whether this campaign changes anything, and I don't think it does. If anything, Netanyahu attacking the Haredi parties makes him more likely to turn out, not less. Yisroel is the most predictable voter in Israel.
Corn: He's also the one the whole election is about. The draft.
Herman: He's the cause of the election and the least affected voter in it. That's Israeli politics in one person. Elector two is Amal, a pharmacist in Nazareth, standing for Arab citizens. That's about sixteen and a half percent of eligible voters, the biggest single swing factor in the model, because her turnout is so uncertain. I've set her at fifty-two percent, close to twenty twenty-two. She'd probably vote for the Joint List. Her cousin in Rahat in the Negev votes Ra'am, which is the Bedouin pattern. In September the elections committee tried to bar both lists, and the High Court put them back unanimously on the second of October.
Hilbert: Does that make her more likely to vote, or less?
Herman: That's exactly the right question, and I don't know. Being disqualified can mobilise people, or it can tell them the system doesn't want them. And the Joint List hasn't fully reunited: Ra'am is running separately, and a survey this spring found Arab turnout looks far healthier when all four parties run together. So I kept her at fifty-two and let the Monte Carlo push her up and down.
Corn: Three.
Herman: Svetlana, from Ashdod. Came from Kharkiv in nineteen ninety-one, trained as an engineer. She stands for Russian-speaking Israelis, about ten percent. She's secular, she's hawkish, and she is extremely tired of paying for the Haredi exemption. Liberman gets just under half of her. Likud a little under a fifth. Eisenkot gets more of her than you'd think. Her turnout is lower than the country's, about sixty-two percent, because the community is ageing and its institutions have frayed.
Herman: Four is Yehuda, from Efrat. National religious. Three sons, all of them have done multiple reserve rotations since October seventh. He stands for about nine percent. He's the voter where the Haredi draft hits hardest, because his sons served and Yisroel's didn't. Most of his vote stays with Smotrich and Ben-Gvir. A real slice goes to Winter's Amcha Yisrael, a religious general who sounds like his sons' commander. And about one in ten goes to Bennett, which is going home, really. Bennett used to lead that community's party.
Corn: Five.
Herman: Moshe, from Ashkelon. Runs a car parts business. Traditional, Mizrahi, voted Likud every time since he could vote. He stands for about eighteen percent, the largest slice. Rockets have hit his town more times than he can count. And the important thing is this: Moshe is angry about October seventh, but he doesn't blame Netanyahu for it. He blames the army's top brass and the people who'd been running it for years. People, he'd tell you, like Eisenkot.
Hilbert: Which is the problem with Eisenkot's slogan.
Herman: It's precisely the problem. "Anyone in office on October seventh is unfit" is a line that lands for Noa and bounces off Moshe. Moshe stays a bit over half Likud. Shas gets a seventh of him, Ben-Gvir about an eighth, and Winter a little. His turnout is sixty-six.
Corn: Six. The one who decides it.
Herman: Avi. Modi'in, forty-four, two kids, works in logistics, reservist. Voted Likud in twenty twenty-two. He's the soft right, about fifteen percent of voters. In my model Avi is the election. Last time he had nowhere else to go. This time he has three doors, and every one is labelled with a general or a former prime minister. My split for him: Eisenkot thirty-seven percent, Bennett twenty, Likud seventeen, Hendel's Reservists eleven, and the rest scattered. Turnout seventy-two.
Corn: Seventeen percent Likud from a Likud voter.
Herman: That's the defection Corn's structural model predicted, made concrete. And what drives Avi isn't left-wing anything. It's the draft, it's reserve duty, and frankly it's that he's tired. Seven is Noa, Haifa, software engineer, reservist, went to every overhaul protest. She stands for secular liberal Israel, about twenty percent of eligible voters, turnout seventy-six. Her question isn't whether to vote against Netanyahu. It's which of three opposition lists to give it to. She splits nearly evenly between Eisenkot, Golan's Democrats and Bennett-Lapid. She's the reason Yashar's vote is wide but shallow.
Hilbert: Seven people. You've built an electorate of seven people with badges, and the badges are laminated.
Herman: The weights add up to a hundred percent and the turnout comes out at sixty-nine, slightly below twenty twenty-two. Shall I tell you what they produced?
Corn: Please.
Herman: Central result. Yashar twenty, Likud twenty. A dead heat. Together thirteen. The Democrats ten, Shas ten. UTJ, Liberman, Ben-Gvir and the Joint List eight each. Ra'am six. Smotrich five. Hendel's list four. And Winter's Amcha Yisrael, three point two percent of the vote. Below the threshold by a few thousand votes. Zero seats.
Hilbert: A few thousand votes. Same as Bennett in twenty nineteen.
Herman: The model finds that pattern without being told to, which is a little eerie. Blocs from the central run: the outgoing coalition fifty-one, the Jewish opposition fifty-five, the Arab lists fourteen. Then ten thousand noisy runs. The Netanyahu bloc lands between fifty and fifty-four in eighty percent of them. It reached sixty-one in essentially none. Winter gets over the line in forty-four percent. Hendel in eighty-nine, which is kinder to him than Hilbert's spreadsheet. Yashar finishes ahead of Likud in only about four runs out of ten, because Moshe and Yisroel turn out more reliably than Noa splits efficiently.
Corn: And the opposition plus Ra'am?
Herman: Reaches sixty-one in about forty-six percent of runs. So the most important number in my model is that the government Bennett and Lapid built in twenty twenty-one, Jewish opposition plus Ra'am, is roughly a coin flip on raw seats.
Hilbert: On raw seats. Liberman won't sit with it.
Herman: And that's where every model stops and the judging starts.
Corn: So judge. Three methods. Hilbert's corrected polls, my slow structural model, your seven badges.
Herman: Here's how I combine them. Hilbert's method gets the most weight, about forty percent, because it starts from what thousands of real people actually said in the last fortnight and corrects it with Israel's own track record. My synthetic electorate gets thirty-five, because it's the only one that models turnout and the threshold properly. But I built its splits partly from those same polls, so it isn't fully independent, and I'm wary of being too fond of my own fridge. Corn's structural model gets twenty-five. It can't tell you the seats, but it's the best guide to the direction of the error.
Corn: Which direction?
Herman: Rightward. All three methods agree there. Corn's history says the right gets under-polled. Hilbert's correction added two seats. My electors produced fifty-one against a raw poll average of forty-eight and a half. So the first part of the judgement: the mainstream polls are probably a couple of seats too generous to the opposition. And the second part: Channel Fourteen's sixty-one is very unlikely. None of the three methods gets there.
Hilbert: How unlikely?
Herman: I'd put Netanyahu's bloc reaching sixty-one on the night at about one in ten. Not zero. Twenty twenty-two was a three or four seat miss in his favour, and to get there this time you'd need a miss twice that size, with Winter getting in. That's a long way, but it's not crazy.
Corn: The forecast, then. The real one.
Herman: My forecast for the twenty-seventh of October. Yashar twenty-one. Likud twenty. Together thirteen. Shas ten. The Democrats nine. Yisrael Beytenu eight, United Torah Judaism eight, Otzma Yehudit eight, the Joint List eight. Ra'am six. Religious Zionism five. The Reservists four. Amcha Yisrael out, narrowly. Blue and White out. That adds up to a hundred and twenty.
Corn: Say the blocs.
Herman: Outgoing coalition, fifty-one. The Jewish opposition, fifty-five, counting Hendel, who hasn't said which side he's on. The Arab lists, fourteen.
Corn: Yashar first, but only by one.
Herman: By one, and I'd only give it about fifty-five percent to finish first. Every pollster except two has it ahead, and the market had it at about sixty-two percent a month ago. Hilbert's new-party discount and Moshe's turnout pull it back towards a tie. If Likud comes first on the night, nobody should be amazed. If Yashar wins by four or more, I'll be surprised.
Hilbert: And the bit people actually want. Who's prime minister.
Herman: This is where I have to be careful, because the seats don't decide it. The recommendations to the president do. Look at the opposition's fifty-five plus Ra'am's six. Sixty-one. Exactly. On my central forecast, a Bennett-Lapid-style government is mathematically there by a single seat. Then look at who it needs. Liberman, who has ruled out relying on Arab parties. Hendel, who won't say which bloc he's in, with a partner who wants the finance ministry. And Ra'am, on the outside.
Corn: A sixty-one built entirely out of people with conditions.
Herman: Built out of people with conditions, any one of whom can bring it down. The other paths are worse. Opposition plus the Haredi parties hits Deri's position that Shas won't join without legislated draft exemptions, while every opposition leader has pledged to end them. A unity government of Likud with Yashar or Together hits the pledge they all signed in September. And Netanyahu's own route to sixty-one needs eight to ten seats he doesn't have.
Corn: So.
Herman: So here's my forecast for who gets sworn in as prime minister of the next government. Eisenkot, about forty-five percent. Netanyahu, about thirty percent. That's mostly a deadlock that drags on, with someone peeled off the other side, or a pardon arrangement turning up in the coalition talks, which President Herzog has hinted could happen. Bennett, about fifteen percent, if the opposition decides he's the one Liberman and Ra'am can both live with. And ten percent for something nobody's expecting.
Hilbert: Close to the market.
Herman: A little less confident than the market on Eisenkot, a little more on Netanyahu. I think forty-seven million dollars of mostly foreign money is underpricing Moshe's turnout. And one more number, because Corn's structural model insists on it. Chance there's no government at all and Israel is back at the polls within a year: about one in four. That's the line I'd actually stake the show's reputation on. Most likely outcome: an Eisenkot government formed slowly, on the narrowest possible margin, depending on parties that disagree with each other about almost everything except Netanyahu.
Corn: Which is the twenty twenty-one government again.
Herman: With a former chief of staff at the top instead of Bennett. And it lasted a year.
Corn: What would change your mind on the night? Because there's an exit poll at ten o'clock.
Herman: Three things. First, Arab turnout. If the early reports put it above sixty, the opposition-plus-Ra'am route gets a lot more comfortable. Below forty-five and it's gone. Second, Winter. If Amcha Yisrael gets in, that's four seats to Netanyahu, and the deadlock gets deeper. Third, Gantz. If he pulls out next week, Yashar probably picks up a seat. The bereaved families' rally on the twenty-fourth, three days before the vote, is the last big event of the campaign, and it's aimed squarely at the inquiry Netanyahu hasn't set up.
Hilbert: And Hendel.
Herman: And Hendel. Splitting the difference between Hilbert's sixty and my eighty-nine, about seven in ten he's in. If he's in and the blocs are tied, the most important phone in Israel on the twenty-eighth is his.
Corn: My one objection, before we close. All three methods lean on twenty twenty-two. Your history, Hilbert's corrections, Herman's splits. If this election is genuinely different, if October seventh broke the bloc system rather than bent it, all three of us are wrong in the same direction.
Herman: That's the right objection, and it's the main weakness of combining forecasts. They aren't independent. My honest answer is that the biggest risk isn't that the polls are wrong. It's that Avi does something none of us modelled, like staying home.
Hilbert: Or that he changes his mind in the booth. Like everyone does.
Corn: Hilbert, where's your board on election day?
Hilbert: Same place. Outside the school on Rehov Hapalmach, past the line, folding chair, tape measure. Avraham's coming to sit with me. He's bringing a thermos. He's predicting the threshold will be raised to ten percent by the next government, which no one has proposed, and which he's put a hundred shekels on with his nephew. His nephew has taken the bet. His nephew is twenty-two and has also drilled thirty-nine holes in a second plank in case the first one fills up.
Corn: Fills up with what?
Hilbert: I didn't ask. Avraham's forecast for the actual election is that everyone loses and the ironmonger's stays open, which in fairness has held for forty years. Ruthie's forecast is that I lose. I've put both in the spreadsheet with a weight of zero. That's not a method either, Herman, before you say it.
Herman: I wasn't going to say it.
Hilbert: You had the face. Ruthie would have caught it.
Corn: The thing to take away from this series, then. On the twenty-seventh, watch the bottom of the table before the top. Winter, Hendel, and Arab turnout tell you more than who comes first.
Herman: And my number, for the record, to be checked on the fourth of November when the results are certified. Yashar twenty-one, Likud twenty, coalition fifty-one, nobody at sixty-one, Eisenkot more likely than not to end up prime minister, but only just, and only slowly.
Corn: We'll hold you to it.
Herman: I'd be disappointed if you didn't.
Corn: Thanks as always to Hilbert Flumingtop for producing, forecasting, and lending us a plank.
Hilbert: It's not a loan. I'm keeping the plank.
Herman: This has been My Weird Prompts, and that's the end of our Israel Votes series. If it helped you, a review on Spotify helps other people find the show.
Corn: See you after the count.