With novel work priors haven't been validated, so it fitting them isn't a sign of the methodology working or not. I think the conversation with this blog piece at the moment should be more focused on the methodology and it's validation. The results might be a mirage of the methodology.

I think it is a good experiment to run, but I don't trust the results. Namely ranking things on a numerical scale seems sort of like false quantification, and I am unsure how much tiny changes in prompt change LLM outputs. I think this is a starting point that needs more fleshing out.

How real is Drummond's shooting? He is hitting the few he takes at a good rate, but will that ever translate to guys guarding him in the corner?

Ah, afraid after that I mostly read comic books, which I don't think you are looking for.

Discworld is good. Very funny and well written, and Terry Pratchett isn't too far off of the guys your mentioned. Hitchhikers Guide to the Galaxy by Douglas Adams also very good, but a bit more dark humor.

While his mobility clearly is not where it needs to be, I do think he looks good strength wise on the court.

Thought for a sec Hamuda wouldn't have anything sweet in the Prowess list, but then I see the Eddymurk Crab and realize I was worried for naught.

Is Maxey the second best player in the East?

VJ is playing really well, but Maxey being one of the best off ball players in the NBA is such a cheat code.

I wonder if the Simic player knows I ran 3 Jackal, 2 Beastbinder and 3 Multiversal Passage because I didn't own anymore of those cards.

Yeah, having some GY hate would be nice. I just copied the list from a league dump. Online I think has had dropping numbers of Goryo's, so Endurance mostly fights Phlage. But that matters. No Shardless also makes it very hard to Evoke, also why no FoV.