In this post I'm analyzing in detail the cbet success % on various board textures, which is my theory objective for the week. Warning: long post. Comments and observations welcome!
OVERVIEW:

Cbetting seems to be about equally as successful regardless of preflop action, which is a little surprising. It's quite successful when isolating limpers too.
ANALYSIS BY HIGH CARD:
A-high flop:

As expected the cbet is more successful here than usual, even when isolating limpers (whose limp/call range is sometimes Ax heavy).
K-high flop:

Fold equity is about the same. Villains seem to fold quite often after calling my 3bet.
Q-high flop:

FE decreases sensibly, particularly in 3bet pots
J-high flop:

FE keeps dropping in 3bet pots
T-, 9-, 8-, 7-, 6-high flops:





Percentages don't seem to drop below 40%.
ANALYSIS BY BOARD TEXTURE
Rainbow flop:

Two-tone flop:

Monotone flop:

As expected I have a bunch of FE on rainbow flops, less on two-tone, which have possible FDs, and a bunch on monotone flops, which I did not expect (I have low sample size though). One explanation for monotone boards might be that those boards look scary and it's hard for villain to call without a card of the right color.
Flop with no connected cards:

Flop with connected cards:

As expected, more FE on unconnected flops
Paired flop:

Unpaired flop:

As expected, more FE on paired flops.
WORST CASE SCENARIOS
Flop 2-tone, connected and unpaired:

Flop 2-tone, connected, unpaired and Q-high:

Flop 2-tone, T-high:

CONCLUSIONS
I confirmed that A- and K-high and paired flops are the ones with the post FE. Monotone flops might be a good spot too, but I need more samples to confirm. On a lesser measure, unconnected flops also have decent FE.
I assumed that cbet success % would drop below 30 in the worst case scenarios, but it's weird to note how they never seem to drop below 40 pretty much no matter what. Which means that cbetting 100% vs unknown is probably +EV on any board (I did not expect this).
My flop cbet is 67, I might be able to increase it slightly using these observations since villains seem to fold so often.
Comments/observations?