Just a quick thought regarding the idea that there is "practically zero time" to analyze the size and implication of each contribution - That is by design. Each individual meeting is not meant to contain proof that work was done. It's meant to prove one thing: that consensus was reached.
See https://james-mart.medium.com/proof-of-nothing-c6d00bfc83e7
All analysis must happen between meetings. And if you publish your analyses, you not only help correct subsequent valuations of the claims on a particular project, but the analysis itself is a contribution you can use to earn more community respect for yourself.
In conclusion, you could be correct that Fractally is unable to cope with what you call type 'L' events. In fact, those types of events are extremely difficult to correct for in any consensus mechanism. But my word of caution is that we can't analyze fractally by looking at individual meetings. We must assess it over time.