a = Probability of Continuous Violence on scale of 5

b = Probability of Poll-day violence on a scale of 5

g = Probability of MCC violation by the dominant candidate


In recent times, West Bengal has been counted among the most violent states regarding the polling exercises. There is indeed some truth in it. Two reasons can be thought of here. First, WB is a long standing hub of violent politics for last few centuries, and that inherent violence intensified through the various attempts of armed revolution against the British Colonial as well as the Sovereign Indian State. Being a border state now, cross-border operations through the inevitably porous border is rather easy to operate here. And hence, violence has always been a political game-changer in this region.

Secondly, given the current state of affairs, the political parties operating in this region are heavily dependent on siphoned government funds for their own expenses, for unlike other regions in India, there are no major industries here that could be extorted for the funds. Hence, at the lowest levels of the three-tier political structure, where the leaders and semi-leaders directly handle the development funds, people fight tooth and nail for the control of that fat moneybag, be it Panchayat or municipal elections.

For some strange reasons, the elections of these various levels of the administrative structure cannot be held simultaneously. During any 5 years cycle, we have 4 different elections here; no wonder “development” is perpetually “on the hold”. While on the other hand, major manufacturers of the local illegal arms like 9 mm guns and ammunitions have shifted their base from upper Bihar to South Bengal. Hand bombs with shrapnel, socket bombs and other IED are produced like cottage industry products here. In a way, the entire land strip between Delhi and Calcutta, namely WB, Bihar, and UP has been sitting on a huge pile of gun powder and other explosives. The latest polling spree has been a witness to that.

However, it cannot be said that all these arms are always used in political violence only. Yet there are pockets in this vast area, where 49 out of 50 crimes happened during any given time-span have something or other to do with the local political ambience. It’s sheer violence and muscle power that determines the hegemony in general, consent is often manufactured through the barrel of a gun, irrespective of what Mao suggested.


In our study, we have collected the crime reports from both the police and the crime beat reporters, 80% of which are never reported to media, because mainstream media as well as our legal framework have their declared policies in place of suppressing most of the crime data, especially where arms and atrocities against women are involved. Thankfully, there are still other ways to collect data than the official sources and mainstream media channels. We have volumes of data to be processed still lying here in 3 of our computers.

A small section of the data is processed and the result will be published here.

Our conjecture or the primary thesis revolves around a few hypotheses. Let’s list them first.

  1. There is no political consensus in general. No one is sure what the “correct politics” is. All what we get to see now, is the commonality of subjugation.
  2. Subjugation is quantified as the probability of continuous violence of political nature: P(a)
  3. Tendency to violate the MCC (Model Code of Conduct), denoted by P(g) increases with P(a), so that P(g)-P(a) is the actual influence during the election days in favour of the given perpetrator. The dominant candidate usually likes this value positive and hence as P(a) increases, there is more space for P(g) to increase, for the dissenting voices are already subdued by the perpetually violent hegemonic forces.
  4. The natural voting pattern in any given constituency, or booth is towards neutralizing this P(g)-P(a) value.
  5. Specially targeted Poll day violence appears there in order to add an extra leverage (L) to this P(g)-P(a) value. Therefore L = P(g)-P(a)+P(b), where P(b) is the probability of poll-day violence.
  6. In our kind of electoral system, win of a particular candidate in a particular constituency can be hardly predicted. But it is much easier to predict, even in perception, the possibility of Candidate A “losing” at a Constituency X.
  7. This possibility of losing therefore belongs to a closed interval [0,1], where A cannot lose at X returns a Surprise Factor ŋ(A)=P(A loses at X)=0. This value is often determined by continuous political exercises throughout a given term. Similarly, ŋ(A)=1 would actually assign ŋ(B)=0 to another candidate B at the same X. Yet, if A wins instead of B, then it comes as a shocker. Such a win is often attributed to an unforeseen L.
  8. Naturally, if by other perceptive analyses, the Surprise Factor in favour of one candidate A is minimized, the opposition B will try to increase Leverage in his favour by adding more P(b) since the other two components of L are beyond his control, and the Poll-day violence will be increased. Therefore L(B) is inversely proportional to ŋ(A). However, the common mathematics of L(B).ŋ(A)=K (constant) is difficult to prove here.

With this brief mathematical model at hand, it’s now time to explain the table above. During our pre-poll opinion survey among the political workers at the booth-level, we collected extensive data about both the perception (ŋ) and general crime scenario (a) of certain areas, not pertaining to the map of the constituencies. During the campaign period, we were keeping log of the instances of MCC violation (g), while on the polling days, we were collecting data about specific instances of violence within a radius of 1 km. around the polling booths (b). Quite naturally though, we had more volunteers and connections in the WB area, and we have more reliable data about WB than say, Bihar. The polling results of UP are still explicable in terms of Leverage, but Bihar is forever an enigma.


However, in the above table we have presented the seats won by BJP in WB. As I wrote earlier, 7 seats were confirmed beyond any doubt in favour of BJP. For those 7 seats, Surprise Factor was indeed zero. Those seats are easily identifiable in the table, by observing the column 4. Absolutely no doubt was there about Darjeeling. Any layman would know that Darjeeling could never vote for any party that has been ruling the state at least one. In their common perception, both TMC and LF have ditched them big time. There the margin of win was also the largest. On the other hand, the absolute shocker was the Burdwan-Durgapur seat, which was won by the lowest margin by a person who was in Darjeeling earlier, and was such a poor performer there that he could not become the repeat candidate from his own constituency, that is Darjeeling. The other candidates having surety about “not losing” are listed towards the bottom of the table. Please note that all these constituencies experienced maximum leverages (L) too. We know that it is quite natural, because TMC tried really hard not to lose these seats, AND they knew beforehand that they were losing. Another almost certain candidate was BJP’s Shamik Bhattacharya at Dum Dum who didn’t feature in this list, because he didn’t win though the common perception in the area was there that he could not lose.

In the other three columns, we have reduced the real data to fit into a scale of 5. That is, if 2.9 is the number on the continuous violence column, it is likely that if 5 crime scenes are reported, most probably 2.9 of them are political. Or in other words, given any crime on the recent history sheet of that area, its probability of being political is 2.9/5 = 0.58. Also, probability of a political crime being less than 0.4 is almost certainly an indicator that the area is generally peaceful, which Darjeeling certainly is not, but Bishnupur is.

This Dum Dum episode, in comparison with Alipurduar or Ranaghat gives us clue to another factor that was not calculated earlier, that of “vote-bank shift” from Left to BJP. There always was an indication that something like this was going to happen, but nobody except the conspirators could imagine that this was going to be so big in absolute terms. Not all BJP candidates won because of this unprecedented, manually controlled swing. Only some did. Bankura, Bishnupur, Bangaon, Ranaghat, and Alipurduars are there for sure, where such swing gave an unforeseen leverage to BJP, though the actual leverage was not much. On the other hand, this swing didn’t happen in Dum Dum, and only partially happened in the Jadavpur constituency. However, a quick comparison between the Surprise Factor and the Leverage columns would validate our hypotheses beyond any doubt.


It is always possible to predict the results of some seats with a degree of accuracy. But perception studies fall behind the real political moves made by one or another party. Hence, it is better to predict the minimum number of seats going to a particular camp, and not the maximum.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.