Saturday, January 23, 2010

Hoping with statistics

I hope Sarath Fonseka would win the presidential election that would be held in 26th, two days from today. These two days are crucial. Although campaigning should have ended by 12 pm yesterday government media would continue to campaign for current president for these two days as well.

Fonseka has a real chance of winning this race. Those who pretend to know the probable mandate of the public is only counting on the past provincial council election results. Provincial council results are not effective enough to be concerned with because most of opposition supporters didn't vote because they didn't want to be in loosing side and political landscape has changed dramatically following Fonsekas entry as the common candidate, specially in the last few weeks.

I would try to analyze previous presidential election results and try to come into a conclusion about the possible outcome this time around.

Following are statistics of 2005 in millions

Votes of Mahinda 4.8
Votes of Ranil 4.7

JVP votes went to Mahinda 0.4
JVP votes that would go to Mahinda this time around because of Wimal's split and discontent about their current political allegiances 0.1(25%)

After change of JVP votes

Fonseka 5.0
Mahinda 4.5

UNP votes that would be cast to Mahinda this time 0.4(which is 8%)
SLFP votes that would be cast to Fonseka 0.1(whic is only 2%)

standing after vote changes between the camps

Mahinda 4.8
Fonseka 4.7

North votes that were not casted previous time 0.5(This is a rough guess, not sure about number of votes and not sure they would go and vote in high numbers or how many would go to other candidates)

New Northern vote split between two
Fonseka 0.33 (66% of votes)
Mahinda 0.17

After northern votes
Fonseka 5.03 (50.3%)
Mahinda 4.97 (49.7%)

Which deliveres Fonseka a slight lead against Mahinda.

I was biased towards a fonseka win during the calculations, I changed my mind against some statistics when it looked like Mahinda would lead Fonseka. But I gave Mahinda more advantage when calculating, like in JVP vote split where I gave 25% split for Mahinda and in vote changes between UNP and SLFP i gave 8% unp vote changes against only 2% of SLFP vote change.

So people if you want change over more of the same keep that hope.

6 comments:

sittingnut said...

"lies , damned lies and statistics"

lol @ liberal lanka's attempt at number crunching.

2005 pe election mr 4.9m (4.88m) , rw 4.7m
2008/09 pc elections upfa 5.1m , unp 2.6m, jvp 0.24-

he dismisses provincial council results saying 'because most of opposition supporters didn't vote' but forget to mention that by going back to 2005 he increases unp vote by 80% ( did almost half of them stay way in pc election?)

by same action he reduces mr vote by 5% ( bc in pc election even with less ppl voting mr got more absolute number of votes than in 2005).

does he think bc upfa victory at pc election was so certain, that all mr voters voted in that pc election? hardly realistic and counter factual to all previous pc elections
it is likely that some of them too did not vote .

-
jvp was not able to get 0.24 m votes in pc but they also get increased to 0.4m ( 66%) or 0.3m (25%) ( without wimal ) for liberal lanka calculations

in the process he neglect to mention explicitly, that according to him mr got 4.4 votes sans jvp in 2005, meaning that by going back to 2005 mr lose 7m (14% instead of 5% )votes from what he got in 2008/2008 pc election with less voters and w/o jvp
-

he gives arbitrary numbers for exchange of unp votes for mr and vice versa. that is in addition to that 7m bsolute but not explained reduction in mr vote.
-

i agree that it is foolish to assume numbers for north and east . but in spite of saying this liberal lanka assumes numbers!
do not agree with 1/2 (34/66)split of north and east vote given the politics of non tna groups including tmvp in east .

-

he provides for jvp change of sides but does not do so for cwc and ucpf.

-
he assumes no increase in voter numbers ( other than northern ) from 2005 to 2010 most of them in rural south (bc most ppl live there) or from pc to pe election due to higher turnover, even tough that has happened in every election ( increasing absolute vote numbers of both parties)

-

instead of giving 'more advantage to mahinda ' as he claims he has actually concealed his reduction of mr votes
he has ignored unpalatable facts like cwc, and mr getting more votes in absolute numbers than in 2005 at pc election.
he has simply ignored other factors that could affect the result materially, probably bc of his ignorance

even after all that manipulation, favoring sf , what he gets is a close election.
w/o those manipulations. it would be a thumping victory for mr

Liberal Lanka said...

@SN

Funny how you have addressed an imaginary audience. Afraid to direct any questions directly to me as I do?

This is a forecast, and I have stated the reason for considering 2005 elections. Poor you, if you base your prediction on provincial elections. Wait for your candidate to win 70% of votes then.

I think you have not got the point and is jumping up and down for all the confused reasons, so typical of you. This is to elaborate that the race would be closely contested than government fear mongers like you claim. To urge anyone who wants change (unlike you who love the status quo) to vote for the combined opposition candidate. This is a follow up of my earlier post.

So all those about numbers are irrelevant the only wise point seems to be your objection of not counting CWC votes. Honestly I forgot about that cross over. CWC would have some influence but not much. Count my word Fonseka would clean sweep Nuwara Eliya.

//he assumes no increase in voter numbers ( other than northern ) from 2005 to 2010 most of them in rural south (bc most ppl live there)

Of course since I can't predict the pattern of newer votes other than assume it would be similar to previous overall vote percentages. My impression was first time voters(mostly boys not girls. girls seem to vote for their fathers candidate bc in our culture they rarely stand up for a course) prefer Fonseka over Mahinda.

//even after all that manipulation, favoring sf , what he gets is a close election.

Since you are normally slow to grasp the point i would help you here. That was the whole intention here. The race would be close. You need to seriously step out from your elite neighborhood and from your holiday homes.

Thanks for comment.

Anonymous said...

You can stuff u statistics up your A** on the 27th Morning... you dumb twit... People are not fools... they have much more brains and are much better at stats that you!

Liberal Lanka said...

@anon

Most people don't vote for the winning side so their or my statistics would not have any effect on their votes.

The election would be closely contested and even if Mahinda wins he would not be able to be the king he dreams to be, thanks for Fonseka. sorry if it breaks your heart.

sittingnut said...

lol@liberal lanka trying cover his manipulations of stats.

btw i don't address him bc he is immune to reason and facts

-
i did not base any predictions on pc elctions , nor have i said anybody is going to get 70% vote, but pointed out the effect on absolute vote number when going back to 2005.

here are liberal lanka's manipulations in brief. for details refer to my comment above >

by using 2005 instead of 08/09 pc numbers as base, he is increasing unp vote by 80% . did all of them stay away ?

mr got 5% more votes more in absolute numbers at pc elections than he got in 2005. (absolute numbers , not percentages. i repeat this distinction bc ppl like liberal lanka does not seem to understand the difference.)
he did it w/o jvp
so by going back to 2005 and by his reduction of arbitrary number of jvp votes, liberal lanka is reducing mr base vote by 7m (14%) in contrast to 80% increase for unp.

is it realistic to suppose that all of mr voters voted in pc elections, if victory was so certain? it is historically inaccurate. there are always more voters at pe than pc elections . so both parties absolute vote numbers increase. liberal lanka increase one by 80% reduce the other by 14% .

while making provision for jvp crossover ignores cwc and ucpf crossover.

he assumes numbers favorable to sf in north and east provinces ( while saying they are unknowable!)

after making a provision for increase in northern vote he ignores the increase in other parts of country . when reminded he makes comical comments on them ( you have read them above to believe them !). fact is increase in new votes will happen most, in places where ppl live most ( in rural south ) . is it right to assume they will reflect already manipulated national average of liberal lanka ?
--

w/o such sf favoring manipulations it is clear that mr will win big . but liberal lanka finds 'hope' in such manipulations

lol.

Liberal Lanka said...

@sittingNut

Have the guts. Don't run away.

You are depending on provincial council election while stating otherwise. You have the right to depend on it but I think it is not effective so I stick to 2005 presidential elections.

You are talking of votes government had increased in the provincial elections but is stating you are not depending on provincial elections. Many people voted for the government bc of the military campaign but most of those voters are discontent with the government. Also most first time voters of the opposition didn't vote in the provincial elections bc they didn't want to hinder military operations and didn't approve the conduct of the government at the same time. That is why I didn't take provincial council elections much seriously.

If anyone is going to take advantage of newer votes it is Fonseka but I have no way of predicting them. Please do so if you can predict since it would make the prediction complete but I am not able to do so.

Yes I have hope. You sound like it is a crime.lol.