Method
The ballot closed, and one list did not file
Candidate lists closed on 8 September with 38 filed. Every poll this model reads was taken before that, and 14 of the 15 lists it names are on the ballot. The table has 57 for the opposition bloc, 51 for the Netanyahu bloc and 12 for the Arab parties — six seats apart on a run where the simulations did not move at all.
At 22:00 on 8 September the doors of the Central Elections Committee closed. Over two days, representatives of 38 lists filed candidacies before its chair, Judge Noam Solberg, for the election of 27 October. Names, letters and disqualification petitions are still to be ruled on; final approval of the lists is set for 27 September.
For a forecast this is the most consequential date in the campaign that is not polling day. Until 8 September the set of things a voter could vote for was still moving — and this page has spent three weeks reporting it moving. A list re-formed on 19 August. Another was counted twice on 27 August. One left the ballot on 4 September and a second merged on the 6th; one of them came back on the 8th. Every one of those was a reading of a ballot that had not closed.
It has closed. So the reading can be checked, and this is what the check says.
Fourteen of fifteen
The forecast names 15 lists, plus a residual bucket that is not a party. Of those 15, 14 filed:
Yashar, Likud, Together, The Democrats, Yisrael Beiteinu, Shas, United Torah Judaism, Otzma Yehudit, Ra’am, Amcha Israel, Blue and White, the Joint List, Religious Zionism and The Reservists-New Economic Party.
Four of those are combinations this page reported as they happened, and the filings confirm each one. Religious Zionism filed with Zehut, as the sources have been reporting it since 3 September. The Reservists filed with Yaron Zelekha’s New Economic Party, the merger of 6 September. The Joint List filed as Hadash, Ta’al and Balad — and Ra’am filed separately, which settles a question the corpus could not: the reunification of 19 August did not extend to Ra’am, and the two Arab lists on the ballot are the two the model has been carrying. Ofer Winter filed Amcha Israel alone, after merger talks with Religious Zionism failed.
The fifteenth is Unity (fmr. Erdan-Edelstein), and it did not file. Gilad Erdan announced the list would not run; Aliza Bloch, who had joined it, is now third on Blue and White’s list, which filed just before the deadline and asked to keep the letters כן.
Unity is still in the forecast. This page gives it a 30-in-100 chance of
winning seats, and blocs.yaml puts it in the Netanyahu bloc.
Why a closed ballot is still open to the model
Nothing in this pipeline knows what filed. A list is in the forecast because a source still gives it a column, and the sources give Unity a column because the polls do — every poll in the 22 the model reads was fielded before 8 September, and several were fielded before Erdan’s announcement. The mappings record what a list is called, whose it is and which bloc it sits in. None of them records whether it is on the ballot, because until this week there was no ballot to be on.
This is not a lag that clears on its own. The polls in the window will age out over the next eight weeks and be replaced by polls that do not ask about Unity, and the number will decay rather than go to zero at the moment it became false. The honest description today is that the forecast’s list set is the polls’ list set, and as of 8 September those are two different things by exactly one list.
It is a small list — a median of 0 seats, and 30-in-100 is a chance of winning any, not of winning four. Removing it is not a small change, though: the chamber is fixed at 120, so a list taken out is paid back to every other list at once. That is a change to the model, not to this week’s data, and it is not in this round.
The one poll, and six seats that did not come from it
One poll arrived. Lazar for Maariv, fielded 7–8 September, sample 600: Yashar 25, Likud 20, Together 13, The Democrats 10, Yisrael Beiteinu 9, Otzma Yehudit 8, Shas 7, United Torah Judaism 7, the Joint List 7, Ra’am 5, Religious Zionism 5, The Reservists-New Economic Party 4. Amcha Israel polled 2.8%, below the 3.25% threshold, and won nothing. Blocs in that poll: 47 Netanyahu, 57 opposition, 12 Arab parties, and 4 in no bloc.
The Jerusalem Post’s report of it and the ingested copy agree party for party, on the sample size, on the fieldwork dates and on the sub-threshold percentages. So does the bloc arithmetic.
That one poll moved the printed table from 54 and 54 to 57 and 51. It did not move the simulations at all.
Both statements are true, and the gap between them is the whole of this round. Amcha Israel’s chance of reaching four seats went from 52.3% to 48.0% — a coin toss landing the other way. Its median went from 4 to 3, and the printed table seats a median of 4 in full and a median of 3 not at all, so its four seats were handed back and apportioned by largest remainder to Yashar, Likud, Together and The Democrats, one each. Three of those four are in the opposition bloc. Six seats of margin, from one list moving one seat.
Underneath, the simulated medians are 53 for the Netanyahu bloc, 54 for the opposition and 12 for the Arab parties — the same three numbers as two days ago. The Netanyahu bloc reaches 61 in 3.6% of simulations, from 4.3%; the opposition in 7.0%, from 6.4%. The opposition is the larger of the two in 57.7%, from 53.7%.
It also caught this site out. The sentence at the top of the page read “are close, at about 57 and 51 seats” — six apart, and called close. The word had been firing on the simulations rather than on the two figures printed beside it, and until today every run that reached it happened to have both. It now describes the gap a reader can see: within three seats the blocs are close, and beyond that the leader leads.
The rest of the order
The opposition bloc and the Arab parties reach 61 between them in 90.6% of simulations, from 90.1%. That is arithmetic about seats and nothing else: the 61 would come from seven lists that do not act as one, and no coalition is modelled here.
Yashar is the largest list in 54.3% of simulations, Likud in 37.6%, and they are level in 8.1% — 51.9%, 40.4% and 7.7% in the 8 September run this one is compared against.
At the threshold, Religious Zionism is printed at 5 again, its chance of four seats or more 68.6%, from 67.9%. The Reservists-New Economic Party, seated at 4 by Maariv, has a 22.0% chance of winning any seats, from 14.9% — the second poll to seat it under the merged name, after Kan 11’s on 6 September; the two of 7 September put it at 2.1% and 2.7% and seated it not at all.
Amcha Israel and Religious Zionism remain the two lists whose fall would take most off the Netanyahu bloc. Both are now certainly on the ballot. Only one of them is clear of the coin toss.