Final Prediction: Meretz and Yisrael Beitenu may sit together, at home

In the past few days, I have added an algorithm to the model to correct for pollster errors. The algorithm computes the correction separately for every pollster, that is, the difference between the prediction based on that pollster’s polls and the actual result in past elections. It also computes the variance for the different corrections and includes the correction only if it is consistent across different election campaigns. For example, if in one election, the prediction based on that pollster was 20% and in practice the party received 15% (a difference of 5%) and if in another election, the prediction was 10% and in practice the party received 7% (a difference of 3%), the correction for that pollster/party combination would be the average, 4%. But if in the second election, the party received 15% (a difference of -5%), the correction won’t be applied because the variance for the two corrections 5% and -5% is too high. With the pollster corrections, the result that we get is as follows: As we can see, both the left-wing Meretz and the right wing Yisrael Beitenu parties are predicted to not pass the minimum threshold of 3.25% that is required to receive seats in the Knesset. Meretz still has some chance as we predict it will pass the threshold in a chance of 3 out of 10. The reason the model predicts this result is the very poor results these two parties have shown at the polls as well as the trend in the election cycle that has been on the decline for both parties. Tal Galili has written a post about the trends on his blog, “Bad Poll” (Hebrew with English graphs). Furthermore, based on past elections, pollsters tend to give both parties a result that is higher by 1-2% than the outcome in the elections. For example, in 2009, the polls gave Meretz between 5-7 mandates but in practice it received only 3, in 2006 between 5-6 and it received 5, and in 2013 between 5-7 with an upward trend, but it received 6. The downward trend alongside the correction for the predictions based on previous poll results that place the parties higher than the actual performance at the ballot-box push both parties below the threshold. Similarly, the absence of both parties and the allocation of the mandates they would have gotten amongst the other parties increase the chances for Herzog and Livni to set up a coalition. Previously, their chances were practically nil but now about 20%. We can see this also from the mandate allocations. A coalition of United Torah Judaism-Shas-Koolanu-Zionist Union achieves 53 mandates. The Knesset has 120 seats, but if the 13 MKs of the Joint List abstain as they have declared they would do, that is just one vote short of the majority needed at the crucial Knesset vote to confirm the government. (A comment – “Four vote majority” in the graphs means that the government is able to survive/pass laws even if 3 MKs defect and vote against the coalition position, though…

The support of the Joint (Arab) List is critical for a Zionist Union win

Benjamin Netanyahu has traveled to the United States to address the Congress on Iran. The effort can reset the news agenda to turn back the recent trend in the polls where the Likud has been losing ground. It seems his replacement is likelier than ever, but what are really the chances for his replacement? Can such a thing even be measured? Other sites like Batel Be60 present the chances for setting up various coalition makeups but do not present the big picture – what are the chances for the Likud to set up a coalition compared to the chances for the Zionist Union to do so? In order to compute these chances, I counted coalitions – how many coalitions from the entire set of possible coalitions does Netanyahu have in various scenarios and how many do Herzog and Livni have in those same scenarios. The results are presented before you.…