Assign symmetric and antisymmetric stetches in IR spectroscopy

Typically the antisymmetric stretch in IR spectroscopy is higher than the symmetric stretch for a given functional group. For example for $\ce{NO2}$ the antisymmetric stretch falls at $\sim 1530\ \mathrm{cm^{-1}}$ compared to $1350\ \mathrm{cm^{-1}}$ for the symmetric stretch. Likewise $\ce{NH2}$ has a similar pattern ($3400\ \mathrm{cm^{-1}}$ antisymmetric and $3300\ \mathrm{cm^{-1}}$ symmetric).

However, in some cases this pattern is inverted. e.g. $\ce{[HCO]+}$ has an symmetric stretch at $3896\ \mathrm{cm^{-1}}$ and an antisymmetric stretch at $2814\ \mathrm{cm^{-1}}$. How can this be explained?