One pollster has not put Netanyahu’s bloc below 60 since September; everyone else says the 50s
Sixteen days before the 27 October vote, the gap between Israeli polling houses on the same fieldwork dates is larger than the gap between the two leading parties.
The Middle East Forum’s poll-of-polls tracker, which logs pollster, publisher, fieldwork date and sample size for 321 polls taken in 2026, shows Likud and Yashar averaging the same figure: 23 on the tracker’s own average, 21.4 on the simple mean. The distribution underneath is not the same. Across five polls, Yashar’s range is 20 to 23; Likud’s is 19 to 26. The spread comes almost entirely from one source. On fieldwork date 07 October, Direct Polls for i24 News (n=518) put Likud on 26, Yashar on 20 and the Netanyahu bloc on 60. The same day, MP+TM+SN+A for Channel 13 (n=1,013) had Likud and Yashar tied on 20 and the bloc on 55. Two days earlier Midgam for Channel 12 (n=501) put the bloc on 49.
That pattern holds backwards. Direct Polls for i24 recorded blocs of 60 on 01 October (n=504) and 60 on 24 September (n=537), and has not published a bloc figure below 60 on the tracker since 17 September; Channel 13, Maariv, Kan and Channel 12 have ranged between 49 and 55 over the same weeks. The Jerusalem Post has flagged the divergence twice, noting that in one batch of five polls Yashar led four and Likud led only the i24 survey, and that bloc averages stood at 53.6 for Netanyahu’s side against 53.4 for the Zionist opposition. Haaretz’s 07 October write-up of the Channel 13 poll — Netanyahu’s bloc on 55, neither side secure without Arab parties — matches the tracker row exactly, which is a useful check that the aggregator is not drifting from the Hebrew originals.
Assessment: The bloc number is not a measurement, it is a modelling choice: who gets counted, and whether Arab parties are treated as available. Haaretz’s 01 October framing is the load-bearing one — the opposition’s own refusal to count those seats is what turns a Yashar plurality into a Netanyahu path, and no amount of fieldwork changes that. So read seat totals, not blocs, and read sample sizes: the outlier house is running roughly half the n of Channel 13. The Gil interview will be tested against these numbers over the next fortnight, and the temptation on both sides will be to pick the series that already agrees with them.