Well let's start with the fact that you're once again continually disregarding MJ's wishes as stated by:
And I'm going once more politely to ask you not to promote such use on this forum.
In other words stop, this is another one of those things that's just getting old, and new to this one it's something that mars the forum.
As to your abuse of statistics there, you seem to be using the term "statistically significant" and "variant" in ways that simply aren't how statisticians define the words.
Statistically significant is worthless without stating the threshold of significance. 98% threshold is common enough, which means there is a 98% chance that the hypothetical statement is true. In this case the null hypothesis is that smoking increases chance of cancer by 25%.
Now in order to determine this we have a bit of work to do. I guess we could find the variance, but it's really just a stepping stone to the std deviation, so lets just skip right ahead to that.
std.dev=sqrt((sum(X-u))/(n-1))
Where X is the resultant values.
u is me not being able to type the 12th level in the greek alphabet on this forum and references the mean.
n is the population sample, and 1 is subtracted from it because we are not using the entire population.
n>40 and as such a std bell shaped curve may be assumed.
The sum(x-u) comes out to 2.5, the total difference between the sample and the expected.
std.dev=sqrt(2.5/999999)
=sqrt(2.500025e-6)
=.001581
So back to our null hypothesis.
Our hypothesis is that smoking increases the chance of cancer, so the easiest way to prove this is disproving that smoking and chance of cancer are not correlated.
For that the be true the std.dev of .001581 has to compensate for 2.5 people
2.5/.001581=316.3 std deviations
So... I don't know std deviations out that far so let me check.
My table doesn't go anywhere near that high. 5 std deviations will bring you to 99.9999% certainty. That's already past the number of significant figures we had in this equation, and as such the answer is 100%, most likely to be reported as 99.9999%, confidence in the statement that smoking increases the chance of cancer.