The problem with doing this is that we have 3 months of data which includes, launch and a holiday.
Really, from what I can discern all you are doing is taking last gens figures and then extrapolating the figures based on marketshare to the numbers we have from the aforementioned 3 months of data.
I dont see a confidence interval, standard deviation or even factored in inflation, income levels or mention of saturation into the results
Its just simply too rudimentary of an equation, and to be honest quite pointless because youre not measuring a static figure. Youre attempting to extrapolate marketshare based on 3 moving parts and companies who will act and react with each other.
There is nothing wrong with your range necessarily if you had a concrete confidence interval to work with or compute, People saying the range is too wide, I feel, dont understand statistical analysis.I just dont think you can compute a decently accurate range with the variables you're using here.
Youre numbers are very sound and I honestly wouldnt be surprised if this is fairly accurate but I just cant imagine, like I said, 3 companies reacting to each other not adding in unknown variables and when we KNOW there will be unknown variables, unseen and unaccounted for, it makes this analysis, at this time, pointless.
After a year (this time next year), you could build a similar graph that I feel would reflect the changing market and give better figures.