Thanks for the info. I have followed the poles and measured on google maps - its actually less than 250m, something around 215m. Obviously this doesn't include the additional wiring inside the cabs but it should be fairly accurate, and nowhere near the 400m+ mark expected for this level of attenuation.
Is your line carried by poles all the way to your home (ie your line runs pole -> pole -> pole), so you can follow the actual path? Or is the distribution cable underground, only coming out for each pole to do just the last drop to the homes?
If the latter, then the line could go down one side of the street, feeding poles on that side, before crossing and coming back up the street, feeding those poles. If that were the case, you'd probably see the changes in the estimations for different houses following the street layout.
Anything else, and the problem is more likely to be gauge or material. 0.4mm copper has resistance of around 133 ohms/km, compared to 85 ohms/km for 0.5mm copper; a line of 0.4mm copper would be getting behaviour approaching 1/2 or 2/3 of our normal expectations. Aluminium lines put a similar kybosh on reality over expectation.
Anyway, this is by-the-by while there is a fault around. Dropping broadband by picking up the handset is definitely indicative of a problem...