I have suggestions. from my experience, there is practically only 1 combination that can consistently survive above depth 140. I think it would be a good idea to have more combinations that are actually very good (now it seems to me that the only good late game strategy is aspiring aspect + familiar strategy , or bee strategy. This could be done even without adding new followers, just fixing some combinations that seem like they should work, but they don't (e.g. familars don't get statistics from another one, you can't upgrade a unit's combat trait with an artificer many times, you can't give a dragoon's combat trait to a unit, the additional attack of an additional trait is not multiplied by a dancer (if applied to another dancer too), etc.) another thing is that some followers or additional traits are much too weak, even if not on their own, they simply don't fit into almost any combination, and it's definitely worth changing them to better fit different strategies. these new combos could then be difficult to do in a normal playthrough (to prevent from abusing them ), and would be geared towards endless mode. finally, an alternative to crow would be useful (high random or AOE damage), and in the initial few depths, a slightly larger store size would be helpfull (I often had a situation where I simply restarted when I didn't get faith's production at the beginning, without which it's hard to good start)
Yeah... Aspiring + Familiar is definitely the anomaly since it's the only cubic scaling in the game. I'm currently in the process of looking at all the least used followers and see if there can be a similar cubic scaling mechanic, but it's tricky since the while system is reeeaaally not designed to have scale infinitely (as in endless mode). Multiplying dancer's attack could work since charging aspect is the only way to make dancer attack twice and it's already pretty rare, I'll do some testing on that to see if it'll be too op. However, artificer and dragoon will be hard to implement, since the "+1 to all values" effect is not a variable, but a different version of the follower; and what dragoon is doing is change the first aspect into its intrinsic trait instead of actually getting 2 aspects.
I agree that early game need some small adjustments. The current plan is to adjust the spawn rate of initiate and supplicant when the player is on low faith and make the 4-choice recruit events show up 1-2 depth early. Hopefully that'll make the early game feels more fair.