Sorry Tim but that is still a possible variance of 1.5 mmols ish, all I was trying to say is there is a big allowable difference, thus who is to say which is correct?
Once again, no. The possible variance isn't 1.5 mmol/l. The possible range is 1.5 mmol/l, distributed on a normal curve. A test will not be 1.5 mmol/l away from the real value, it will be no more than 0.83 away from the real value, 95% of the time , which is a significantly different statement.
As per the image below, the range of values at 4 mmol/l,
95% of the time, lies between the two red lines on the curve. Or to put it another way, the maximum variance is 0.83mmol/l in either direction, but the value on the meter is much more likely to be around about the real value than at the extremes we always talk about.