It can't go to 8/10 just by fine-tuning the difficulty in order to make PotD easier, trust me.
I don't mean just difficulty, there are a lot of small things that detract from the experience. Basic modern RTS shit like being able to use a common ability with multiple characters at the same time doesn't exist, I have to use it individually for each. Ability bar being only 1/3 the length of the screen, with all my mastered spells getting needlessly condensed into a separate menu I now need to access before using spells I use virtually every battle, formations getting scrambled after conversations, etc.
I've been playing this for a long time and I'm pretty sure there is something wrong with the ruleset itself. Ironically, after all the talk about balance, it feels like it gives really lopsided chances to one of the sides from the get go. If one side has an advantage it's very unlikely that the other will turn it around.
Example from Deadfire - you cause the Blinded affliction on a group of enemies, say with Chill Fog. Chill Fog targets Fortitude.
So, by beating their Fortitude defense once, you've first reduced their Perception by 5 (which in turn means -5 Accuracy for them, and -10 Reflex), increased their recovery time and secondly, since Blinded also automatically causes Flanked, this reduces their Deflection by 10, reduces their Armor Rating by 1, and they lose another 10 Accuracy. Reducing the armor rating by 1 point may also be just what you need to get an additional bonus - this time to damage (because of the Penetration mechanic). This effect comes on top of the reduction of Deflection, which by nature of the graze/hit/crit rules, also statistically increases the damage that you do - because lower enemy Deflection means higher chance to score a Hit instead of Graze, and a Critical instead of a Hit. You are getting one damage bonus on top of another.
To sum up - you defeat the Fortitude defense *once* to get a total of -15 Accuracy, -10 Reflex, -10 Deflection, -1 Armor Rating, +50% Recovery Time, -50% Range for ranged attacks. This effect lasts a good amount of time if the blinding attack hits. Enough to decide a combat encounter. For the price of one successful Fortitude attack, you are awarded being able to make multiple other, significantly more damaging attacks. So much more damage that defeating the encounter becomes trivial. And when you add that all your abilities are per encounter, and you have no incentive to economize their use...
Synergies like these lead to a nice puzzle-like feel of combat - you have to consider your party's abilities and the enemies' weakest defense in order to determine from which defense to start in order to dismantle their other defenses. In my experience the downside to these synergies is that success in combat tends to snowball very quickly - if things go badly from the start you are almost certainly getting crushed, and you can't turn the battle around. Conversely, if you open up correctly, the encounter becomes "going through the motions" and you likely won't even lose significant amounts of health. You already know the outcome and combat becomes a boring sequence of clicking abilities and targeting enemies, with the result being in the bag. This also leads to the player using the same approach and the same sequence of attacks and abilities in different encounters and against different enemy types, because what counts isn't what the enemy is, but which is its lowest defense. You tend to cycle as many "strategies" when approaching an encounter, as there are defenses.
I know it's less "intellectual", but I like D&D's RNG approach better, because it keeps you more on your toes. Josh's system seems overengineered by comparison, with so many interconnections in order for everything to "matter". The resulting information overload makes outcomes harder to predict instead of easier - with RNG at least you know where the turning point was.
I doubt anyone, even Josh has been able to rigorously test how powerful these cumulative effects really are. From his forum posts, he seems to be taking cues from the audiences' perception when balancing - if players complain about something, he boosts or nerfs it respectively, otherwise it's fine.
