The efficiency indices I was talking about are entirely things that could be calculated from TTW, EM, and win rate. I'm most of the way convinced that for the purposes of any efficiency index, it's pretty much a best practice to assume TTW=TTL, because TTL is a very weird number.
Also, with regards to the more all-inclusive one I'm working on, I was pretty much planning on using it in terms of turns played rather than time overall, because time per turn varies wildly between various players as well as various connection speeds, and regardless of how complex a deck is, it seems like the time any given turn would take a given player would eventually converge on the same number regardless of deck as they got more used to it.
As far as electrum/score won, I was planning on doing it all on averages, rather than recording every single game, much like FGei assumes 40 electrum per win. Same with card spins, as FGei assumes X% chance of winning a card any given game. Score is nearly the same as electrum, but for the unified EI I was talking about, I was more considering factoring in the value of unupped cards won from rushing bronze or whatever, as well as the sell values for spun upped rares.
Here's a very rough sketch of what I've been thinking about. This is all from memory right now and is in no way complete yet. The constants I mention are things that can probably be reasonably guessed by the community, and if not, I can just start keeping track when I'm testing things and come up with some overall rough averages.
-Start with turns per win, calculate how many games you'll win (and lose) on average over 1000 turns.
-Multiply games won by 2/3rds of the stated reward of the given target (e.g. FGei states ~60 electrum and uses 40 electrum per non-EM win). Multiply this number by (1+(2*EMrate)) to factor in extra money for EM's at the same rate used by FGei.
-Subtract cost for games lost.
-Determine a constant for cards won from regular spins. IIRC, FGs use approximately 0.5 cards per win, so if we used that as a starting place, each win would add 0.25 times whatever constant we came up with for the average value of a spun unupped card. As a shot in the dark, I'd say something around 45, which I believe is a bit above the non-pillar card value average (and given the prevalence of more expensive cards and winning a bonus five electrum from the spun card, I think that'd be fine). I'll verify that number if I ever actually write this up fully.
-Probably otherwise ignore the +5 electrum from two-card spins, because it's pretty minimal.
-Use the tables in the OP to determine average number of special spins over the stated period, and figure out some constant for how often you successfully spin a special spin (I'd probably say 0.66 or 0.75 as a starting point), and add the values of the unupped or upped rare to the total.
That'd give you expected money over 1000 turns, which is a much easier-to-record stat than time. Anyways, that's pretty much just an off-the-top-of-my-head description of what I've been thinking about, so I'm probably missing a point or two in there, but as it stands, it's all calculated entirely off constants, TTW, win rate, and EM rate, so it'd be able to be applied to basically any deck in any (PvE environment) at any point.
That's a lot more than I meant to write just then. If anyone sees any holes in this or has thoughts on values for constants, please let me know. If necessary I'll give this its own whole post sometime.
edit: For the purposes of comparing to FGei, money over 333 turns is money over one hour if one turn takes 10.8 seconds. Money over 250 turns is money over one hour if one turn takes 14.4 seconds.