Which is why I have always said that the 90/95% numbers are just numbers that may or may not be useful depending on what one is wanting to do with them.tiny data sets like this are very misleading.
Works the same way with the ResMed leak numbers. Can be very misleading and sometimes causes unnecessary stress/panic when someone sees a slightly above red line in the sand and they think "OMG my therapy is totally in the toilet because of big leaks almost all night long". In reality maybe only a very short period of time was the therapy in the toilet (if it was even in the toilet in the first place).
90/95% numbers are too easily skewed to be particularly useful on a nightly basis.
Long term we might see more value and I mean months and not days or weeks long term.
The "or below" part of the definition leaves a big open door for variances.
An overall average (not the median average) might be a more useful tool to be using.