jeffreyh: So perhaps the constent fiddling doesnt help, just one theory.
I was wondering the same thing. But even if user resets are counted towards forced resyncs\MTBR one would hope the acceptable thresholds they have set are a bit less than 8 days, which I would say around what my average is!
Reading the patent for DLM (which seems to be the most informative thing I can find on DLM!) it looks like DLM can be configured\implemented in a few ways for how it decides whether something is a wide are resync, a user initiated or a "forced resync" (caused by line conditions) :
1) They will collect data across multiple connections to determine when an "area-wide" event causes a resync.
2) They may just set a certain figure as an estimate for how many user initiated resyncs there may be then subtrace the figure for this and area-wide to then determine how many forced resyncs there were
3) They may set a threshold so that a resync where the connection is down for less than a certain time period is considered forced, but if it's down for longer then the user probably turned off the modem and it's considered a user initiated resync.
4) They divide the day into time bins then record bins as either having connection\no connection and somehow count contiguous bins to determine which are user initiated and which are forced.
I guess the main take away is that if you're going to initiate a resync yourself, it's best to leave the modem off or disconnected for an extended period to maximise the chance it's recognised\classified as a user resync not a forced resync.