There's an obvious answer of course – don't overtighten them, but bear with me.
Volcanic ash permitting I'm hoping that my R853 Pipedream Sirius will be with me this week. It has sliding dropouts, which do slide (rather than being simply replacable and held in place by the frame) so I assume that the bolts that hold them in place will be under a lot of load. I emailed Paragon Machine Works, who I belive supply them, to ask about the recommended torque range and was told 24-26 NM, fair enough. However the email went of to say at the upper limit of this range (and 2Nm is hardly much of a range), unless you use a new 5mm wrench there's the possibility of rounding out the bolt head on a steel bolt.
Now it strikes me as odd to supply hardware that might round out when used as recommended. I'll be running mine geared and was intending to threadlock the bolts in nice and tight and leave it at that, but if you were running singlespeed/Roholf dropouts and regularly sliding them to tension the chain all that loosening/tightening might be rather worrying.
Interestingly the email went on to mention that titanium bolts are available where rounding out isn't an issue and that they're a nice upgrade. Now I though that titanium was stronger than steel by weight, but not by volume and that therefore ti. bolts were weaker than their steel counterparts (my Raceface stem instructions warn against replacing the steel bolts with titanium ones). Since I've heard of steel bolts holding Orange P7 sliding dropouts in place breaking I'd have thought that titanium bolts torqued up to a level where you might round out the head of a steel one, may well stretch or break, but there you go.
So, am I likely to round out my bolts doing them up to the recommended torque (and I'd imagine that doing less tight runs the risk of the dropouts shifting) or is this an opportunist sales pitch?