BRF GJP Trial Poll-1: How many seats will NDA win?

Locked
vijaykarthik
BRFite
Posts: 1169
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by vijaykarthik »

Mihaylo wrote:I am revising down my last projections just because I feel we are discounting the possible significant impact of EVM magic and rigging.

Before:
Week 5
Mihaylo: 0 25 50 25 0 0

After
Week 5
Mihaylo: 0 45 40 15 0 0

Ok. Let me play a bit of a devils advocate here. Call it DA, red teaming ... anything you like. So, we wanted GJP style functioning and thinking. Lets attempt it here. Helps to think in a Bayesian way.

This is a qn for everyone. not just mihalyo. So feel free to answer / structure and modify as reqd:

a. How much of the vote rigging do you think will impact the results. will it have a positive (upside) or negative (downside) effect? To whom (by looking at your results, I assume you are looking at vote rigging as being disadvantageous to Modi)
b. How easy does one think its going to be to rig. Will it be in pockets or will it be rampant? [What exactly is the process involved?]
c. How easy is it going to be, if and when the rigging is done, to find the event out -- if EC finds the rigging, surely they wont turn a blind eye? So in a nutsheel, how effective are the EC in finding reports of rigging. Is that scientific [it will be interesting to find that out! Say, if the EC found 80% of the cases of rigging, then we are much better off!]
d. How rampant will rigging be and how will that distribution look like -- will the rigger have a distinctive strategy and rig only a few booths / constituencies or like a blithering idiot attack constituencies of all hues and colors [alternatively put, if the rigger knew that a few are suspect constituencies, will the EC not be aware of it and make alternate measures. Alternate alternatively, if the govt [read UPA] knew that the constituencies were riggable and EC knows about it, will they have deliberately made lesser security arrangements in place? [If that were the case, will the EC not have screamed by now?]

From what I am given to understand and what I know of the process:

reg pt b of mine first. : Rigging is not so easy. I have talked about it to people in the know and have at least an understanding of the process involved. Most machines are clean when they arrive. [I saw a few reports which talked of malfunctioning machines at the factory itself. We will take them up later]. Besides, when the machine comes in, its the duty of the polling officer or his assigns to ensure that the machines work -- they press buttons, vote for different parties, check results etc etc to satisfy themselves, look at it, analyse it and then reset it before polling starts. [I am assuming they will perhaps do it before the volunteers of the parties too? I am not sure if this happens currently. Does anyone know?]

On the day of polling, the votes surely are counted. Volunteers from all parties are around and rigging is quite incredibly difficult? If it takes more than 1 person to rig something, its difficult. But on the day of polling, I am assuming to effectively rig, the entire booth needs to stink up to the high heavens. For successful rigging. Not very easy in an elections as this. [I may be wrong. But one gets the idea]. So, its a likely low probability event. [But I am still fascinated by this point and will like imputs / analyses of the rest if they have a better explanation / answer / rebuttal]

Since pt b makes it a low probability event [per me, atleast], lets attack pt a. Yes, rigging will ideally be disadv modi. but since rigging is quite difficult (I will say a 5-7% probability of overall rigging throughout the country). So, I will put it at a <5% event. [overall]

Pt c: EC has been looking at reports and proactively been advising repolling. --> implies to me that they are doing a fair job. I will say they are atleast taking action on 30% of the cases? [Is that more or less?]. Which given my %s tells me that of the <5%, close to 1.5% gets reported and found. Which brings up the tally of potential rigging that is unfound to be 3.5% [I will round it to 5% for now]


pt d: interesting but I don't have a lot of idea on this. I am assuming the rigger will be a strategist. So, he will rig only select constituencies. Which will those be? My assumptions: they will need to be safe seats of congress where they have worries about winning. perhaps constituencies where big shots of bjp stand, perceived closely fought seats (where margin of support is perceived to be even for 2 parties etc), home bastions / constituencies etc. But however, the EC and the volunteers will be aware of this and will be doubly careful? I would like to think so.

I don't know about the rest. But structuring this way helps me in my thought process atleast and helps to crystallize my thoughts. Whew. Long post!

Update: now that this has been mentioned and it is hopefully read and if someone is in agreement, the most important qn. Will it make you change your forecasts? Are you planning to do it? Lets get that done too if this convinces you. (if it doesn't convince, I am all for hearing the "other "side!)
Saral
BRFite
Posts: 1663
Joined: 16 Jan 2005 14:05

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Saral »

There was a lot more rigging in the days of paper ballots.. 1977 still happened and Congi lost. Maybe there is significant rigging in W Bengal or parts of UP/Bihar or Assam but those likely follow historical precedents in those areas. Example: http://www.thenewsminute.com/stories/Th ... 2mxPvldU5o If anything, EVM rigging (and I don't mean hi-tech hacking which is even less probable) does happen and, overall it has to be at most in the low single digits of vote percentages and that too concentrated in a few constituencies.

Exit polls do a decent enough job in seat projections and we should have a very clear idea on May 12 itself. Suppose there is significant rigging, say to cause a swing of 30 seats in one direction, there would be massive discrepancies between exit poll projections and the outcome on May 16. And the pattern of discrepancies would be very telling so we'll get a clear understanding of problem areas. I don't recall such things happening in the past (Sure, there have been repolls. but all those in select constituencies {such as in Bihar} where there has been prior issues). Given the amount of interest in these elections, data from exit polls would be even more robust than previously and that should have a carry-over effect in seat projection accuracy.

Personally, there was some concern when I saw deleted voters list issue crop up in MH and other places.. but, on balance, I concluded that it wasn't a big enough issue. Turnout has increased tremendously, on the order of 10% which is quite unusual. Now, it is unlikely that pre April polls factored this non-randomness into their sampling (differences in turnouts across segments). So for me poll rigging is not a relevant issue at all, in terms of overall impact (yes, it does happen in some booths, so what?). It is used by NaMo as a debating point, to get people excited etc. So that (use) is very different.

Turn-out, not EVM rigging, will be the story of this election. People who are more enthusiastic and pro-change are more likely to show up to vote. Amazingly, this turn out percentage is set to exceed 1984 where Congress got 400+ based on sympathy wave. This increased turn-out will generally favor NDA and, going purely by macro indicators, there is no reason why NDA cannot get near 300 or even exceed it. For same reasons, IMO, UPA will have hard time breaching 120.

Rigging is like cheating in board exams. Does it happen in some places? You bet it does, esp if the exam monitor in a particular class is in cahoots with the students. How pervasive is it? Not enough to make a systematic difference because most schools are reasonably well-run and most monitors follow the rules.
Altair
BRF Oldie
Posts: 2620
Joined: 30 Dec 2009 12:51
Location: Hovering over Pak Airspace in AWACS

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Altair »

Marten wrote:Everyone is talking about higher turnout. Booth volunteers can only capture numbers, but in a lot of places, the party agents are not very vigilant. I believe even the number of those that have voted is inflated. EVM manipulation is probably not only through changes in vote attribution, but through additions that might not be from real voters.
There is one other factor. Every EVM has an average of 1000 vote limit. Assuming 60% turnout we have a upper ceiling of 600 votes for the EVM.
Now whosoever is doing rigging is not foolish enough to vote all 600 for the same candidate. There will be 40%-50% vote for the primary candidate. The rest of votes are divided among rest of candidates in fray. At most 300 votes per booth will be the advantage for the candidate doing booth rigging. Now doing this at 10 or more will attract attention. Can it be done? yes. Will it impact? Depends on the constituency and the size of voter base and voters per booth or per EVM.
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

Altair wrote:
Marten wrote:Everyone is talking about higher turnout. Booth volunteers can only capture numbers, but in a lot of places, the party agents are not very vigilant. I believe even the number of those that have voted is inflated. EVM manipulation is probably not only through changes in vote attribution, but through additions that might not be from real voters.
There is one other factor. Every EVM has an average of 1000 vote limit. Assuming 60% turnout we have a upper ceiling of 600 votes for the EVM.
Now whosoever is doing rigging is not foolish enough to vote all 600 for the same candidate. There will be 40%-50% vote for the primary candidate. The rest of votes are divided among rest of candidates in fray. At most 300 votes per booth will be the advantage for the candidate doing booth rigging. Now doing this at 10 or more will attract attention. Can it be done? yes. Will it impact? Depends on the constituency and the size of voter base and voters per booth or per EVM.

Ans. EVMs can record a maximum of 3840 votes. As normally the total number of electors in a polling station will not exceed 1500, the capacity of EVMs is more than sufficient
Booth Capture :

1. EVM has recorded 90% votes in one candidate's favour
2. EVM has 90% and above turnout and 10% higher than previous turnout.
These indicators read with previous turnout figures, observer's report and Videography, Newspaper clippings, Candidates' complaints etc would result in repoll. I think EC has ordered video coverage and webcasting as well of all polling booths in Amethi(?).

As far as rigging is concerned , it is not very difficult and it does happen in some cases. Eps. in those booths where other contenders have weak presence.

The instances of booth rigging is less now a days but not all together eliminated.

It might affect results , if remain undetected, in less number of seats. Especially in Parliamentary seats , where number of voters are large with several assembly segments, it is difficult. So candidates resort to luring voters.

Personally , in my opinion parties do make compulsory noises and wherever indicators and observer's report show signs of booth capturing EC does order repoll. The overall outcome will not be affected during this election.

I hope that answers all the four questions of Vijayakarthik.
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

Saral wrote:
Turn-out, not EVM rigging, will be the story of this election. People who are more enthusiastic and pro-change are more likely to show up to vote. Amazingly, this turn out percentage is set to exceed 1984 where Congress got 400+ based on sympathy wave. This increased turn-out will generally favor NDA and, going purely by macro indicators, there is no reason why NDA cannot get near 300 or even exceed it. For same reasons, IMO, UPA will have hard time breaching 120.
+100000......

Congis will be reduced to rump party with less than three digit figure, will not cross 100.
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

EVM undergoes Mock Poll several times. The last one is one hour before the start of the poll. All approved polling agents have to sign on the Mock Poll Certificates and EVM sealing materials.They will have to press all buttons and NOTA ith tally of at least 50 and then count is checked against each candidate. Once they are satisfies, All data is cleared from EVM and it is sealed. In case of malfunctioning of EVM during poll process CU component of EVM is sealed and then new set of EVM is brought up from the reserve. Mock poll is again conducted before it can be used. At the end of Poll Presiding offier will give a certificate indicating total number of votes cast from the register where every voter is made to sign along with voter ID number. ER copy is also marked. All Poll agents note down these numbers. A authentic copy is given to each poll agent.

Once EVM-CU is opened for counting, total is displayed. That total is tallied with the earlier total given. However before opening of EVM seal, Counting agents have to be satisfied that seals are intact. Only after that counting process progresses. In case of dispute observers present in counting hall are consulted. Disputed CUs are kept aside. If winning margin is higher than the total votes polled in that EVM the result is allowed to be published. Else Repoll is the option for that booth.

In case there is attempt to rig the poll presiding officer has the option to press the close button of the EVM -CU and then no further voting can take place. Of course one can physically destroy the unit. But data can still be recovered, if needed. Now if everyone colludes including poll agents of rival parties and officials and observers, videography then we don't have hope for democracy. Fortunately , these things are few and far between.
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

Yes all UP seats in this phase and the next are under webcasting.

Posted later. Looks like not all booths are covered. Inadequate covering. Also it seems CPF are not present in full strength. Bachane ki Kavayad Jaari hai Bhaiyya ko.
Please check this URL for live webcasting.

http://electionsmonitoring.in/amethi.html

And for other areas
http://ceouttarpradesh.nic.in/webcast.aspx
vijaykarthik
BRFite
Posts: 1169
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by vijaykarthik »

This is the stuff that I get worried about. I am sure a lot of people will raise points against me for doubting. But, its a democracy and people are allowed to have different opinions and we can always agree to disagree.
http://www.ndtv.com/elections/article/e ... eststories

My concerns still remain. He needs a mentor and a challenger. And needs to reduce the tone. Lets hope for the best.
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

vijaykarthik wrote:This is the stuff that I get worried about. I am sure a lot of people will raise points against me for doubting. But, its a democracy and people are allowed to have different opinions and we can always agree to disagree.
http://www.ndtv.com/elections/article/e ... eststories

My concerns still remain. He needs a mentor and a challenger. And needs to reduce the tone. Lets hope for the best.
NYTimes and rrNDTVee are something which I would not even bother to trust. Barry is known to be mouthpiece of inimical elements and has habit of running down Indian interest.
Saral
BRFite
Posts: 1663
Joined: 16 Jan 2005 14:05

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Saral »

Week 5

Saral 0 0 0 0 0 40 60
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

Chaanakya 0 0 0 0 0 0 100
harbans
BRF Oldie
Posts: 4883
Joined: 29 Sep 2007 05:01
Location: Dehradun

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by harbans »

How many think NDA can touch 400 or more? Or 350-400?
panduranghari
BRF Oldie
Posts: 3781
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by panduranghari »

harbans wrote:How many think NDA can touch 400 or more? Or 350-400?
In 2019, that should be the goal.
satya
BRFite
Posts: 718
Joined: 19 Jan 2005 03:09

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by satya »

Does ABC pattern of financial markets apply to election scenario ? If yes then where is the B /corrective wave . So far we have seen only directional movement as per pre-poll surveys .
KLP Dubey
BRFite
Posts: 1310
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by KLP Dubey »

harbans wrote:How many think NDA can touch 400 or more? Or 350-400?
Definitely a few Nishadas (boatmen) of Varanasi...they have already stated their opinion clearly, why not include them in GJP ? ;)

Nishada committee member: 0 0 0 0 0 0 100 (and specifically 400+)
devesh
BRF Oldie
Posts: 5129
Joined: 17 Feb 2011 03:27

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by devesh »

Week 5 devesh same as week 4
Saral
BRFite
Posts: 1663
Joined: 16 Jan 2005 14:05

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Saral »

Given expansion of regional parties, 400+ in 1984 probably is like 330+ today. A 350+ would be a historic mandate. I am sure they are planning for 2019 and the assemblies in-between. For this year, not too much above 300 is my estimate. I think 280 to 300 is more reasonable (if optimistic by BR expectations) for this year.
vijaykarthik
BRFite
Posts: 1169
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by vijaykarthik »

chaanakya wrote:EVM undergoes Mock Poll several times. The last one is one hour before the start of the poll. All approved polling agents have to sign on the Mock Poll Certificates and EVM sealing materials.They will have to press all buttons and NOTA ith tally of at least 50 and then count is checked against each candidate. Once they are satisfies, All data is cleared from EVM and it is sealed. In case of malfunctioning of EVM during poll process CU component of EVM is sealed and then new set of EVM is brought up from the reserve. Mock poll is again conducted before it can be used. At the end of Poll Presiding offier will give a certificate indicating total number of votes cast from the register where every voter is made to sign along with voter ID number. ER copy is also marked. All Poll agents note down these numbers. A authentic copy is given to each poll agent.

Once EVM-CU is opened for counting, total is displayed. That total is tallied with the earlier total given. However before opening of EVM seal, Counting agents have to be satisfied that seals are intact. Only after that counting process progresses. In case of dispute observers present in counting hall are consulted. Disputed CUs are kept aside. If winning margin is higher than the total votes polled in that EVM the result is allowed to be published. Else Repoll is the option for that booth.

In case there is attempt to rig the poll presiding officer has the option to press the close button of the EVM -CU and then no further voting can take place. Of course one can physically destroy the unit. But data can still be recovered, if needed. Now if everyone colludes including poll agents of rival parties and officials and observers, videography then we don't have hope for democracy. Fortunately , these things are few and far between.
These are very interesting facts. thanks. Which makes me feel that my initial 5-7% of rigging possibility to be a figure on the higher side. I will perhaps put it as 3-5% chance of rigging with atleast a 30% chance of the rigging being found by EC, after looking at these new points. Which will perhaps put total rigging at 2-3.5% (achievable rigging that goes unnoticed). Which will turn out to be about 10-15 seats in LS put together (max number -- ceiling). So, the rigger better be aware of the select constituencies where he needs to target and target them effectively. (Does anyone know of a potential list of rigged constituencies. If anyone gets the data or has it handy, we can be more or less certain that those will be the suspect ones)

And the point on everyone getting involved is pretty interesting [as I had mentioned too. Rigging is difficult and unless everyone has a vested interest, it is quite remarkably difficult to rig with just 2 people who want to rig out of say a possible (at least) 10 people in the booth). And purely from game theory or a rational perspective, for all 10 to come in and allow rigging isn't beneficial for anyone. (The one possible flaw in my logic: when the big candidates really fear for their deposits and want to get more votes to get their deposits back. But large parties should ideally not have that problem?)]

Whichever way we look, as long as we look at it rationally, rigging seems like a low probability event and a difficult one to achieve. So, is it possible that they can rig it after the elections are done? Replace 500 sealed EVM's replaced with "like for like" another set of rigged 500 EVM's etc? Last time I checked (2010/11 state assembly elections, I think) the 3 layered security worked well and I have reason to assume that will be done this time around too. And the strong room was guarded reasonably well.

There are advantages in the wisdom of the crowd! Yeah!
Patni
BRFite
Posts: 886
Joined: 10 Jun 2008 10:32
Location: Researching sub-humans to our west!

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Patni »

Week 5
Patni 0 0 0 0 55 35 10
Anand K
BRFite
Posts: 1115
Joined: 19 Aug 2003 11:31
Location: Out.

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Anand K »

After the penultimate round;

Anand K - 50 25 25 0 0 0 0
rajithn
BRFite
Posts: 470
Joined: 27 Dec 2008 01:52

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by rajithn »

Week 5: 0 0 0 0 0 10 90
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

vijaykarthik wrote: Whichever way we look, as long as we look at it rationally, rigging seems like a low probability event and a difficult one to achieve. So, is it possible that they can rig it after the elections are done? Replace 500 sealed EVM's replaced with "like for like" another set of rigged 500 EVM's etc? Last time I checked (2010/11 state assembly elections, I think) the 3 layered security worked well and I have reason to assume that will be done this time around too. And the strong room was guarded reasonably well.

There are advantages in the wisdom of the crowd! Yeah!
Almost next to impossible. Each EVM is serially numbered. As you know EVM has two components. Ballot Unit (BU) which contains the buttons to be pressed. And Control Unit which records the votes. Both Units are paired for polling. They are sealed seperately. Sealing is done at the Polling station and contains the signature of Presiding officer and Poll agents of candidates for that specific poll station. Now these documents are given to Political parties.It contains the polling station Number, Name, EVM serial numbers used, and total votes cast as per ER Sheet.

Only CU is brought out for counting. BU remains sealed and inside the strong room. One table ( not more than 14 tables in one counting hall) will have only one CU. It will be examined by Counting agents of Political parties appointed for the purpose tablewise. They would have records for those EVMs and tally sheet and that who was Presiding officer and Poll agents who signed . Those officers are now off to their homes and not reused for counting process.

So if EVM is changes then serial number will change. If additional votes are put inside CU somehow ( remote control) tally sheet may not match. Signatures of Pol agents of political parties would not match.

And we assume that all officials present are supporting only one party and that they have somehow colluded to get that single candidate from that party win. This might have been possible in older times. Now with Sheshan spine still running through the ECI and various measures to ensure that Officials are insulated ( few exceptions apart) from local machinery and that if anything wrong is established, Returning Officer will be made mincemeat by ECI and all , such possibility seems difficult.

If EVM of a particular booth is found tempered with and Counting agent raise objections and Observer finds factual merit in that then matter would have to be referred to EC and their decision to be awaited. Counting for that booth will proceed only after clearance by EC or repoll if ordered, Meanwhile counting for other booths would be taken up.

Counting centre does have three tier security and no unauthorised persons are allowed in. Police Officials do not have permission to go inside counting hall. Local Govt officials not on election duty otherwise are not allowed at the first cordon. So the only link would be Returning Officer who remains inaccessible for everybody else except election officials. Again exception can't be ruled out. But at the most he would tell results a little bit in advance since others won't have electronic items inside CC. And Observer would be present. But then they may be keeping low profile. However , collusion at such large scale is difficult to arrange.

The only play they might have is Postal ballot where old tricks could still be played an in majority of cases PBs are not much of a deciding factor unless margin of win is less than 2000 or so.
Yagnasri
BRF Oldie
Posts: 10390
Joined: 29 May 2007 18:03

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Yagnasri »

Narayana Rao

Week 5 same as before.
Supratik
BRF Oldie
Posts: 6470
Joined: 09 Nov 2005 10:21
Location: USA

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Supratik »

last round no change

5 10 10 20 30 20 5
ramana
Forum Moderator
Posts: 59773
Joined: 01 Jan 1970 05:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by ramana »

Vijayakarthik, Thanks for the tutorial on how to make good judgments using available data to count and discount the information.

Folks to make it easy for saral please re-post your assessments. That would be a good practice for it makes his work less bothersome as he doesnt have to go back.

In future we need better format to run longer running polls like this. I was hoping we all get familiar with this tool and the technique to use it for better forecasts of more significant events.
Yagnasri
BRF Oldie
Posts: 10390
Joined: 29 May 2007 18:03

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Yagnasri »

I think we can now close this effort as by and large the people have made up their mind whom to vote.
rvishwak
BRFite
Posts: 384
Joined: 07 Jul 2010 14:03

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by rvishwak »

rvishwak

Week 5 :

0 5 10 70 15 0 0
Saral
BRFite
Posts: 1663
Joined: 16 Jan 2005 14:05

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Saral »

Maybe Week 5 can be the last set of estimates as May 12 is just around the corner and exit polls will be out? I can post week 5 later today or tomorrow.
KLP Dubey
BRFite
Posts: 1310
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by KLP Dubey »

Week 5 (revised): 0 0 0 0 0 30 70
SwamyG
BRF Oldie
Posts: 16267
Joined: 11 Apr 2007 09:22

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by SwamyG »

SwamyG wrote:NDA, week 5:

0 5 35 35 10 10 5
NDA, week 5 (revised):

0 0 40 35 10 10 5
vijaykarthik
BRFite
Posts: 1169
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by vijaykarthik »

Just a thought. How about closing this bit of the poll on the 12th instead? We should have the exit polls by then and should make it more interesting? Since we are looking for good judgment and we don't have a lot of factual data to base it on, a few days delay in closure for this round will perhaps help (particularly because the 12th should give us exit polls)?

That way, we can look keenly on the 12th data and the 15th data and see if we were able to formulate good probabilities and base our judgments with proper grounding.

At least I am currently reluctant to move from my probabilities that I set about 3 weeks ago (though in my mind I feel I should be slightly more optimistic than my current 262 seats that I have voted for here).
ramana
Forum Moderator
Posts: 59773
Joined: 01 Jan 1970 05:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by ramana »

Plan is to close on 12th after polling over. No more estimates to be considered. However thread will be open for saral to post summary and members to post their reasons.
merlin
BRF Oldie
Posts: 2153
Joined: 01 Jan 1970 05:30
Location: NullPointerException

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by merlin »

ramana wrote:Plan is to close on 12th after polling over. No more estimates to be considered. However thread will be open for saral to post summary and members to post their reasons.
That is good. However, saral needs to close week 5 now and members need to post week 6 estimates and then close on the 12th and then another summary. Polls went for 6 weeks this time so we need 6 weeks of estimates.
vijaykarthik
BRFite
Posts: 1169
Joined: 11 Aug 2016 06:14

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by vijaykarthik »

^^ how about a slight change? We consider wk 5 till the 12th. And wk 6 as the days from 12-15th.
It will be a great option (or so I feel) as we will have lots of data from 12th-15th. And will really help us to learn how people "judge" even with the exit polls hitting them in front of their faces. So, 12th evening [PST] we close wk 5 and wk 6 starts on 12th and closes on 15th. It can be slightly difficult for Saral though. :)
thoughts?
merlin
BRF Oldie
Posts: 2153
Joined: 01 Jan 1970 05:30
Location: NullPointerException

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by merlin »

Sounds good but the GJP is all about feel for the numbers without being affected by exit polls. So numbers before 6.00 PM IST 12/5/2014 only should be considered. Exit polls (when not used for psyops) would be more accurate and it would be good to check the GJP numbers against it without members being biased due to exit polls.
merlin
BRF Oldie
Posts: 2153
Joined: 01 Jan 1970 05:30
Location: NullPointerException

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by merlin »

Pre-emptive strike :-P

Week 6

merlin 0 15 70 15 0 0 0
Saral
BRFite
Posts: 1663
Joined: 16 Jan 2005 14:05

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Saral »

I yam traveling next week.. From now on, whatever you put, call it week 6 as merlin seems to have done. You can (and probably should) edit after exit polls.. Some are rumoring NDA 260 to 280 based on leaked exit polls.

Ok so it is week 6 from now till May 16 (before results start coming out).. There are several here without much faith in exit polls. EVM tampering remains a big concern :mrgreen:
chaanakya
BRF Oldie
Posts: 9513
Joined: 09 Jan 2010 13:30

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by chaanakya »

Well you are not giving any scale after 300. There is no alternative.

It is
Channakya 0 0 0 0 0 0 100
Denis
BRFite
Posts: 144
Joined: 05 Sep 2002 11:31
Location: India

Re: BRF GJP Trial Poll-1: How many seats will NDA win?

Post by Denis »

Denis

Week 6: 0 0 0 0 0 75 25
Locked