The forecast for the 26th Knesset: ranges, not numbers
Not current These polls were all published by 23 October 2026. Israeli law bars publishing new polls until the polls close on election day; these are not current, and nothing can be learned from them about how the public votes, or what it thinks, today.
The largest list: Yashar (דרך) in about 50% of the scenarios, Likud (מחל) in about 50% of the scenarios. That is: no clear leader.
Each list: the median of 1000 scenarios, and in brackets the range in which 8 of every 10 scenarios fall. Each pollster's latest poll, 18 October 2026 to 22 October 2026.
A list that clears the threshold in fewer than 95% of the scenarios is marked. Percentages are rounded, and nothing is shown as "certain" or "zero": below 5% and above 95% this model does not know enough to say more. Without surplus-vote agreements (not yet officially published; they move a seat at most), and without כחול לבן, which withdrew from the race on 19 October. How it is calculated →
What is the chance your coalition reaches 61?
Tick lists. The chance is the share of scenarios in which together they hold 61 seats or more: arithmetic of seats, not a forecast of who will form a government. The site puts no list in any bloc.
How the forecast is calculated
- For each pollster, its latest poll from the 21 days before the newest poll is taken, and only polls published before the ban on publishing polls began.
- Each poll becomes vote shares: a list below the threshold gets the percentage the poll published; a list with seats gets its share of the seats plus half a seat, out of the votes left (seats are rounded, and a small list gets a little less than its proportional share).
- 1000 scenarios are drawn around the pollsters' average, spread as widely as they disagree, including lists' tendency to rise and fall together in the same polls.
- On top of each scenario goes an error shared by all the pollsters, of a measured size: each pollster's latest poll in the final ten days before each of the five elections of 2019–2022, against the Elections Committee's official results. A large list misses by more percentage points, a small list by more relative to its size.
- Each scenario goes through the threshold (3.25%) and the Bader-Ofer method: the same engine that reproduces exactly every election since 2019.
What the calculation does not know: a shift in opinion after the last poll, a polling miss larger than anything in 2019–2022, and the surplus-vote agreements. The numbers are an estimate, not a result.
The polls: estimates, not results
Everything in this part is derived from polls: sample estimates with sampling error and house effects, not a count of votes. It looks different on purpose from the part below it, which shows official results verified against the Central Elections Committee.
Based on 86 polls from 9 pollsters, since 26 August 2026.
The seat trend in the polls
Each dot is a single poll's result for one party. The thin dashed line is a trailing average of the last 3 polls for that party: an aid to the eye for seeing direction, not a forecast and not a model. Hover over a dot for its details, and over a party's name in the legend to highlight it.
Dot = an actual poll result Line = trailing average (an aid to the eye only)
The latest polls
Newest first. Party letters as in the legend above. Each row links to the pollster's original publication, where there is one.
The engine that turns votes into seats
Every scenario in the forecast goes through this engine: the threshold and the Bader-Ofer method, over all of its votes. It reproduces every election since 2019 exactly: here you can watch it on the official results.
The 25th Knesset · the seat allocation
2022 · threshold 3.25%
The votes left unrepresented
Votes for lists that did not clear the threshold, and no one was elected on them.
417,400 votes went to lists that fell below the threshold: about 8.8% of the valid votes, worth about 8 seats without the threshold.
The allocation method, the threshold and the surplus-vote agreements are all documented and linked to the official source. We show the numbers; we make no claim that has no calculation behind it.
The threshold's bite over time
How much of the vote was lost in each election
The threshold bit hard in 2019 (הימין החדש) and in 2022 (מרצ and בל"ד), and barely touched the elections in between. That is volatility that shows in the seat count but not necessarily in the electorate.
What connects to this: the forecast's next steps
The seat engine is the base. The live forecast is built on it. Meanwhile you can play with the threshold simulator, read How it works to understand the system, or look at the political map over time.