I draw conclusions based on the best information I have available, and how the hell is that "dangerous?" Nobody is going to get maimed if the iPhone 4S hasn't actually sold 100 million units yet.

But try this one on for size, if the share of iPhone sales which are not the latest model is greater than 10%, that just means the 4S will continue to garner more sales now that the 5 has launched, and continue to sell when the 6 launches. Not troll logic, actual logic. And since iPhone shipments increased by 78% in the last year and 75% the year before that, the 4S would likely sell more units in its lifetime if you're right than if you're wrong.

Don't believe me? Here's the math:

Two scenarios, let's be conservative and assume iPhone growth will slow from 75% down to 50% in both of them. In one scenario, the latest phone sells 90% of all units, and in the other, 70%. In both scenarios, the remainder of sales will be split evenly between two older models, so the 4S will see either 5% or 15% of sales in the next two years before it's off the shelves.

So now who isn't very good at math?

I am not sure if you are attacking my maths, but did you seriously count the total sales in order to come up with the annual extra difference?

Not so much attacking your math, because as near as I can tell, you haven't actually done any. But I am defending my own math and logic skills which you have impugned. I suppose I am attacking your logic though, for failing to recognize that your concerns are meaningless.

And I'd love to answer your question, but it makes absolutely no sense to me, so I'll try explaining my methods to you again.

At issue: Whether the Wii is the best selling single consumer electronics SKU of all time or whether it has been or will soon be deposed by a single model of the iPhone.

We know from this site that the Wii has sold approximately 97 million units, so the question is how many units does a single model of iPhone sell. Your contention is that my data point on the ratio of iPhone sales between models is inadequate to draw any conclusions, and my counterpoint is that if the ratio of iPhone sales which are the most recent model is lower, that merely means that any given model will see even more sales over the course of its lifetime (a longer tail, as they say), because the iPhone market is rapidly growing. In other words, trading 20% (or 30%, or 40%... the bigger the trade, the better the payoff) of this year's sales for 10% of the next two years' sales is a net gain, because the size of the pie keeps growing.

To demonstrate this principle, I started with the actual annual iPhone shipments for Apple's fiscal 2012 and the historic growth rate of ~75% derived from looking at previous years' shipments. To be conservative, I projected annual iPhone shipments to grow only 50% YoY from here on. Then I ran two scenarios of the three year lifecycle of a single iPhone model (3GS was discontinued after three years, iPhone 4 is on its third year now, and I expect 4S will exhibit the same pattern). In one scenario, I assumed that my one data point was representive, and 90% of all iPhones sold in a year are the latest model, with the remaining 10% split evenly between the two older models. In the second scenario I cut the ratio down to 70%, with 30% split between the older models.

The bottom line shows the result. If the 90% ratio is correct, the iPhone 4S has already sold more units than the Wii. If the ratio is 70%, the iPhone 4S will sell more units than the Wii within a year. If the ratio is even lower than that, it may take at most two more years. But failing a catastrophic collapse in iPhone sales, the 4S will sell more units than the Wii. That's just how much bigger the cell phone market is than the video game market. I don't understand why this is so hard to accept.