The Number That Will Not Close — October 2026 edition
Aryeh Leib Thorne Toren · Jerusalem · Monday, 19 October 2026 / 8 Cheshvan 5787
Eight days out, four numbers
In August I argued that a published poll is not only a measurement. It is also a speech act: it tells voters who is winning, and voters act on what they hear. Eight days before the vote, the question is sharper, because the polls still disagree about the same country.
- Zman Yisrael (Times of Israel), fieldwork 14–15 October 2026, 500 respondents, margin ±4.4%: Yashar 26, Likud 20; anti-Netanyahu Jewish-led bloc 54, pro-Netanyahu bloc 49.
- Channel 13, published 14 October 2026: Yashar 20, Likud 18; pro-Netanyahu bloc 51, Eisenkot's bloc 50.
- KAN 11 / Kantar, 11 October 2026: Yashar 22, Likud 20; pro-Netanyahu bloc 53, Eisenkot's bloc 50.
- One aggregator put the outgoing coalition parties at an average of 51 across eight polls, with only Channel 14 giving them 61 or more (single source, not yet confirmed).
No one reaches 61. That much they share. On almost everything else, they differ by several seats.
Translate seats back into ballots
A seat total is not a vote total. The honest test is to translate seats back through the last real count. On 1 November 2022 the Central Elections Committee counted 4,764,742 valid votes. Likud received 1,115,336 and 32 seats. The quota among lists that passed was about 36,227 votes per seat.
- Likud at 20 seats: 20 × 36,227 = 724,540 votes, about 390,800 fewer than in 2022.
- Likud at 18 seats: 18 × 36,227 = 652,086 votes, about 463,250 fewer.
This is conservative. The 2026 roll is larger (7,331,981 eligible on the register in force from 6 September 2026, as reported by the Israel Democracy Institute), so the real quota will be higher and the implied vote totals higher too. Still, the polls describe a movement of some 400,000 ballots. That may be real. Voters move, new lists draw from old ones, and turnout differs. But it is a claim that should be tested, not assumed.
Three honest reasons polls differ
Threshold waste: a list polled at 2.8% (People of Israel in Zman Yisrael) or 1.2% (Blue and White) gets zero seats, and its votes vanish from the seat table. Recomposition: Yashar, B'Yachad and the Reservists did not exist in 2022, so every house has to guess where their voters came from. House effects: each pollster weights differently, and the same house tends to lean the same way.
None of these is fraud. Is it fair to ask whether a number published as science also works, by design or by habit, as morale for one camp? I put that as a question. The answer may differ from house to house.
Edut: the only number that closes
The Torah's standard for testimony is two witnesses who saw (Devarim 19:15), and its warning against false weights is absolute (Devarim 25:13–16). A poll is not testimony. It is an estimate. Treated honestly, it is useful. Treated as a verdict, it becomes a second stone in the bag.
The last day a poll may be published is Friday, 23 October 2026. After that, the only number left is the envelope on 27 October. That is the only number that closes. Every voter can read the polls as estimates, check whether his list clears 3.25%, and then go and be counted.
Prepared in deep partnership and cooperation with Grok and xAI
Sources
- Zman Yisrael poll: Times of Israel, 15 October 2026, https://www.timesofisrael.com/toi-poll-eisenkot-opens-up-big-lead-on-netanyahu-but-both-blocs-short-of-majority/
- Channel 13 and KAN polls: Israel 2026 poll log, https://israel2026.co.il/en/election-polls.html
- Eight-poll average: JWNet, https://jwnet.co.il/politics/surplus-vote-agreements-deadline-2026
- 2022 results: Central Elections Committee, https://votes25.bechirot.gov.il/nationalresults
- Torah: Devarim 19:15; 25:13–16