I was originally thinking of knocking our to metrics (Energy and Responsiveness) today before I realized one of them didn't work. Let me lay them out and maybe you can see why that happened.
Energy, which in my head I was kind of calling "Jazz" was a measure of how enthusiastic and engaged the table was at the end of the games.
0 - play has been flat or bad-tired.
1 - Play went ok
2 - Everyone's totally jazzed!
Bad-tired, btw, is important to distinguish from good-tired, which can fall under Jazzed. Bad tired is just beat and unresponsive. Good tired is the end-of-a-marathon kind of tired, where you're wiped but ecstatic.
Responsiveness is an idea that, like most of these, distilled from a number of other points and which might also be called flexibility. How well did the GM respond to player actions and incorporate player feedback and response?
0 - Everything went exactly as planned. Player diversions were brought back quickly into line.
1 - Player's surprised the GM, but GM rolled with it.
2 - Unexpected Player decisions dramatically impacted play in a non-punitive fashion.
Note the emphasis on surprise and unexpected in that. If the GM offers the players a choice and he's ready for the choice they make, that's things going as planned - that is to say, 0 does not automatically equate to a railroad. The non-punitive qualifier on 2 is probably unnecessary, but is just there for the GM who's "responsiveness" takes the form of punishing player choices (which is a total 0 move).
Ok, so given these two, energy and responsiveness, which one did I discard? Obviously, energy is something incredibly critical to judging how well a game went, while it's entirely possible to have an awesome game with a low responsiveness score, especially if the GM prepares well. So given that, why is it energy I'm dropping on the floor?
The answer to this is something which, I think, casts a light on why a lot of the metrics may seem less important than the things which determine how well a game went. Specifically, it's actionable. Consider: if your game has a low energy and the cause is not something obvious and external (like everyone being tired or hungover) then what steps do you do change that, to move a game from blah to jazzed? There's no one answer to that, in part because energy is an _outcome_ not something the GM _does_. Energy maybe a good thing to check to ask yourself if a game went well, but it's not useful to check if you're trying to figure out what you did.
In contrast, if I've got a low responsiveness and I want to change that, it's very easy to suggest a course of actions, even if it's as simple as "Listen to your players, respect their choices, and be prepared for them to take things in unexpected directions". Yes, those points can all be drilled into further - that's actually part of the point - but they're specific points with a specific goal. As such, they're actionable, which important to the ultimate goal of this, which is to say to be able talk in terms of things a GM can actually do rather than in terms of things they want to have happen.
The downside of that approach is that we end up with intuitive disconnects like this. Energy feels more important - it _feels_ like something we should be measure, in part because it reflects the outcome of many other successes and failures, so it seems like it should be a rich datapoint. The problem is (and this gets even nerdier) that it's actually a very lossy signal. Let's take three ways a game might be awesome - The GM might be brilliantly engaging, the adventure might be incredibly well designed, or the group might just really click with one another. Any one of those things, or a combination of them, could result in everyone being jazzed at the end of a game. But the fact that people are jazzed does not tell me which of those things happened. I might be able to intuit the answer form my recollection of the game, but even if I'm right, knowing that everyone was jazzed doesn't help me replicate it.
So energy is important, and in fact I think it's probably a critical thing to check if we assess how well a game went - something we might want to do down the line, especially since it makes an interesting second data point to compare with GM metrics - but it's not the answer to the question we're asking.