The GMTKN55 data set is a collection of standard benchmarks in molecular quantum chemistry that spans small- and large-molecule thermochemistry, reaction barriers, and non-covalent interactions. The error across this collection of benchmarks is reported as a weighted mean absolute deviation (WTMAD). We identify a flaw in the canonical WTMAD definitions, which weight some benchmarks orders of magnitude more heavily than others. For example, the top 3 benchmarks contribute as much to the total WTMAD-2 as the bottom 36, potentially giving a misleading picture of overall functional performance. We propose a new WTMAD-4 metric and use it to conduct an outlier analysis on hundreds of DFAs from the literature, including the new DM21 and Skala machine-learned functionals. The exchange-hole dipole moment (XDM) and many-body dispersion (MBD) corrections are also assessed on GMTKN55 for the first time. XDM shows excellent performance, especially when paired with minimally empirical functionals and Becke’s recently proposed Z-damping function. We next investigate whether the strong performance of the Z-damped XDM variant extends to the solid state by testing on molecular crystals and the LM26 benchmark for layered materials such as graphite, hexagonal boron nitride, lead(II) oxide, and transition-metal dichalcogenides. The importance of three-body dispersion interactions is explored via inclusion of the Axilrod--Teller--Muto (ATM) term, yielding the best performance achieved on LM26 using semi-local functionals to date. Finally, we introduce a new minimally empirical local-hybrid functional based on B86bPBE that has only three empirical parameters when paired with XDM(Z). Testing on GMTKN55 shows that a carefully constructed minimally empirical local hybrid can perform consistently and accurately, competing with the best hybrid functionals in the literature.