New Delhi: Now that the results for the just concluded assembly elections are available, it is time to examine the post poll predictions for CNN-IBN-The Week. This analysis is written by Rajeeva Karandikar Director, Chennai Mathematical institute.Given below is the comparison between the predicted seats and the actual outcome for the four states.Assam Prediction Actual Cong 64â72 80 AGP 16â22 8 AUDF 11â17 38 BJP 7â11 others 12â20Kerala Prediction Actual LDF 69-77 68 Udf 63-71 72Tamil Nadu Prediction Actual DMK+ 102- 114 30 ADMK+ 120-132 204West Bengal Prediction Actual TNC+ 222-234 228 LF 60-72 61From one point of view, we got it correct in three states - Assam, Tamil Nadu and West Bengal as we predicted the winner correctly - Congress, the ADMK alliance and the Trinamool-Congress alliance respectively while we got Kerala wrong as we had said slight edge to the Left Front.However, looking at it more carefully, it is clear that we were completely off the mark as far as Tamil Nadu is concerned. We did not predict such a clean sweep in Tamil Nadu - no one did. Some surveys had in fact predicted majority for the DMK alliance. Indeed, it appears that the AIADMK alliance had at least 15 per cent vote lead over DMK alliance â something that our survey did not pick up. We had in fact estimated the gap in votes to be 2 per cent and 15 per cent is way outside the error margin. So either our survey was completely wrong or perhaps there was a fear factor at work, namely even after voting for AIADMK alliance, people whom we talked to were reluctant to say so. It may be noted that Jayalalithaa's victory 204/234 gives a strike rate of 87 per cent as compared to Mamata Banerjee's victory in Bengal 228/294 which gives a strike rate of 78 per cent. All the political analysts, psephologists, journalists had talked of a Mamata wave and yet no one had foreseen the Jayalalithaa Tsunami. So clearly there was something happening at the ground which all of us missed. In any case, the underestimation of the victory margin in the survey led to under estimation of victory margin.Of course, in West Bengal we were spot on, both TMC+ and LF got seats that are within our predicted bounds. Of course here most other surveys were also correct. Assam we were the only ones who predicted a Congress majority and indeed Congress got even more seats than what we predicted. About Kerala, we did get the winner wrong but the actual seats are just off by one seat from our upper/lower bounds for UDF and LDF respectively. Indeed, we had predicted a close finish and had indeed said that too close to call with slight advantage to LDF. So in my assessment too, we got three out of four right: Assam, Kerala and West Bengal and we missed reading the Tsunami that hit Tamil Nadu.