Thanks for summarizing this important study. It's frustrating that it sits behind a paywall, since much of the research involved, and even the salaries of some of the participants in the Novim panel that conducted this "meta-study", relied on public funding. So my comment reflects what I've read about the results, including your post, rather than careful scrutiny of the published paper.
With regard to the logic underpinning the assertion that methane leakage is worse than the EPA's inventory indicates, the first bullet point under 1. seems like a better argument for concluding the inventory might overcount emissions. Given the vintage of many of the 200 studies surveyed by the panel, and the age of the data in those studies, their observation that "the sampled devices are often older and not representative of modern techniques used in the energy sector" is important. However, based on the very large set of onsite measurements conducted by the recent UT study, which looked at current natural gas operations and found leakage rates below the EPA's estimates, there's every reason to suspect that leakage rates were higher in the past than today.
Another pillar of the study's conclusions depends on the observed gap bewteen atmospheric measurements of methane concentrations and the results of "bottom-up" estimates. There are many possible explanations for that gap, and I'd like to see more detail about the assumption that went into the modeling used to extrapolate ground-level leakage from airborne concentrations.
In any case, rather than casting doubt on the efficacy of switching transportation energy from oil to gas, which will take years to have a measurable impact on emissions in either direction, the main value of the study would seem to be in focusing attention on the fixable aspects of methane leakage, and in particular on the contribution of faulty equipment and the susceptibility of such leaks to be remedied.