Mobile devices Toolkit
Story: BlackBerry has twice failure rate of iPhone
Whose math is fuzzy?
There is a high-level point to be made here: because all BlackBerry failures are lumped together, it could be that there is one bad "apple" in the bunch---a BlackBerry model that has very high failure rates and therefore accounts for a disproportionate number of the total BlackBerry failures. However, your math is *completely* wrong. The failure rate is calculated as follows:
failure rate = total failures across all models / total number of phones
So let's say for the sake of argument you were right, and the failure rate per model was 1.3%. Furthermore, let's assume we saw 1000 phones for each of the 16 models, meaning for each model, 1.3% x 1000 phones = 13 phones failed. What would the total failure rate be?
failure rate = total failures / total number of phones
= (16*13)/(16*1000)
= 13/1000
= 1.3%
If the average failure rate for each model is X%, then the average failure rate across all models is also X% and vice versa.
Even though the formula is meaningless, I'll also note that 12% failure rate / 16 phones = 0.75%.
Full Talkback thread









