klapauzius
Well-known member
OrientExpress said:Volusiano said:You need to compare apple to apple here.OrientExpress said:17 cars out of a population of 25000 is .00068. That is more than 3 standard deviations from the mean, or in layman's terms, an edge case.
It shouldn't be 17/25000, nor should it be 17/500 because I'm sure there are more capacity losses not reported than the 17 reported.
The only meaningful sampling ratio at all should be
# of reported loss / (# of reported loss + # of reported NO loss) in Phoenix.
So far I've only heard of maybe a few from Phoenix who reported that they haven't lost a bar yet. Some of them who said they hadn't lost a bar have come back later and said they finally lost a bar.
The real ratio that's meaningful to me is more like 17/20 at this point. If you're going to sample, you have to pick samples from the same pool.
I have it right. The comparison are the know outliers vs. the total universe. Cherry picking the total universe from just Phoenix or even the southwest is not valid because one cannot say with certainly that similar condition do not exist outside of your suggested sample universe. Also we cannot say for certain what the total number nor location of all of the outliers are, so by default the entire universe must be considered in the sample.
No you are wrong, because "hot climate" and "cold climate" constitute definitely different distributions and you cannot apply statistics from one distribution to another.
The battery fail statistic in Phoenix is definitely not going to be the same as e.g. in Seattle.
And we dont even need statistics to tell us the difference, battery physics and the known climate stats for different geographical regions make it quite clear why we cant compare across the whole universe.
Now, I wonder what the solution to the conundrum would be:
Suppose Nissan would play extremely nice and replace all those batteries with new ones. Wouldn't they start to lose capacity a year later again? The cant possibly replace them every year?
Installing a thermal management system posthoc would probably be impossible too and would not help much either.
And they cannot pull a new heat resistant battery chemistry out of the hat...
In the end it might turn out that current battery technology is simply not suited for hot climates
Maybe they should offer a refund for the whole car?