Who will not be intersted to know who will be the Next Prime Minister? Who is going to form the Government in the center? Every one has their own predictions - exit polls , stock markets, the parties and ofcourse the common man - every body is free to have their opinion or rather wishful thinking. Statistics and probability of numbers don't consider bias of opinion and just throw some number as prediction. Here I will try to predict some numbers with respect to Lok sabha Elections of 2019. This is pure academic exercise based on the statistics available on public domain- 

Statistics is a vast subject with good potential used by many - investors, researchers, scientists etc. Even Astrology is a sort of Statistical study. But what I know in Statistics is Mean (average) and Standard Deviation. With these known things and excel data available online - here are the layman's deductions  - 

The following is the data collected from online. Data from 10th Lok sabha are taken , before that BJP was just starting or was non existent. Averages and  SD were calculated on EXCEL file. Other National Parties include BSP, CPI, BSP and NCP. Other than National Parties include State Parties like TRS, TDP, SHiva Sena, BJD, RJD, JD (U), SP , Unrecongnised and independents-

 

 

Average is Just the mean i.e sum of the Values divided by the number of values. Standard deviation is something which you have read in your school. It says how far the values are spread from the Mean Average.

 

Now Coming to the inferences part -

Please closely see the above table and note the Averages.

Averages of all the Headings, BJP, COngress, Other National and state parties were taken for 7 Lok sabha(A), 6 Lok sabha elections (B)  like that upto last Two Lok Sabha Elections(F). Whereas A1, B1 upto F1 represents Standard Deviation of the data for 7 , 6 upto 2 Lok sabha Elections. As can be seen from above, even though the BJP's Average was always hovering around 180 in all the six cases, the Standard Deviation is around 60 to 70. So it means that if we take any of the Averages (A to F) as our prediction, the value may not be near value and we have to express it is 180 ±  70. Same is the case with Congress whereas it Mean hovers around 120-140 , Standard deviation is also  high in this case. Same is case with others. One important observtion is except 16th Loksabha where BJP had 282 seats, it was always three front scenario - the third front being the parties which are state parties. 

To truncate the standard deviation , i tried one trick - I have took the Average of the Averages (A to F) . No idea how scientifc it is. It basically will have more weightage to the 16th loksabha elections and the weightage reduces towards 10th Lok sabha results. This  way the recent elections will have more weightage and may be outcome will be more near to this. 

But the most noteworthy feature is that the standard deviation of these Averages - A to F is quiet small. So the range of prediction is also closer. Based on this Average of Averages , following is the prediction - 

Before going to the final tally, some assumptions-

• It is difficult to BJP to repeat the 2014.
• Congress likely to gain. 
• Formation of SP-BSP alliance is likely to hit BJP badly in Uttarpradesh . Last year BJP had 71 out of 80. This time it may loose half. The vote share of BSP and SP last time combined is same as that of BJP (2014) in UP.
• However BJP may do better in West bengal and North Eastern states, Tamil Nadu and karnataka compared to previous. 
• Generally the prediction is given as Mean  ±  3 X SD (σ). σ is the standard deviation
• However BJP also done some good work and people do have some positive outlook on the party. Modi still has some Aura. Based on this, I am assuming BJP will go over Mean. As SD is 10 in BJP case, I feel the added factor will be +4σ or +5σ . That is +40 or +50. 
• Congress is likely to do better than their previous figure , which was the worst in the history. So I feel going with -4σ.  
• For Others I again go with -4σ 

 

Now coming to the final tally 

Parameter	BJP	Congress	Others
Average (X from the table)	180	132	220
Standard Deviation (x1 from table)	10	8	10
±  N X σ -> My assumption rather bias	+5σ	-4σ	-4σ
Likely figure	230	100	180

 Adding all it comes to 510 . Balance 32 can alo be added to BJP - implying 262.

Party	Likely Seats
Shiva Sena (Maharastra)	16
Akali Dal (Punjab)	4
JD (U) and LJP (Bihar)	8
Others	3
Total	31

 This brings our prediction of NDA to 262+31 = 293 Total.

TamilNadu is quiet unpredicatble when comes to the outcome. Hence I have not considered AIADMK as a major factor and the impact is averaged out in the BJP average.

So NDA is set to get the simple majority. Now coming to neutral parties which may either support NDA from inside / outside or may not support anyone. Looking at those parties (from my knowledge) and their predicted seat tally - 

Party	Likely Maximum Seats
BJD (Orissa)	20
YSR congress (Andhra Pradesh)	13
TRS (Telanagana)	11
Some independents - surprise parties	6
Total	50

In case of crisis, From this 50 , BJP can get support. So in all likelyhood, more chances of BJP returning to power.

 

Disclaimer - It is just a simple analysis for study purpose and nothing else.