I'm looking at the heat of formation for branched and straight alkanes and I've noticed a few odd trends I'm having trouble explaining. For straight-chained alkanes, the addition of carbon atoms to the main chain seems to consistently decrease the value of $\Delta H_\mathrm f$. This seemed logical as the compounds become more stable. However, for branched alkanes, I'm having a bit of trouble explaining them. The more bent it is the more stable it seems to be, as $\Delta H_\mathrm f$ decreases. But I'm not quite sure why.
Does it have something to do with geometry? This is a bit confusing because the boiling points of branched alkanes are less than that of their straight-chained isomers, implying that they're more stable. The location of the methyl groups seems to also have a part to play as if they're in the center (most spherical shape) they'd be least stable, but still, have higher $\Delta H_\mathrm f$ than straight-chained ones.