Future of auto-battler genre, and Riot's opportunity

Main comments from reddit thread:

1. "It already seems as though the top players have nearly reached a skill ceiling." In LoL which is very open-ended strategically and mechanically challenging there is a very high skill cap that takes years to develop. TFT not so much.
2. "the idea with TFT from Riots perspective ... is to keep changing the meta ... will make the game feel less stale for longer periods of time. Although I don't know if that will be sustainable ..." It will get frustrating from the player's point of view, since people who don't play much will never understand the meta, and people who play a lot will feel like 'here we go again'. Legendary LoL player apdo said LoL retains competitive players because it's a team game. Your effect on the outcome of a game is not transparent, so you don't know exactly how good you are, so you keep playing.
3. "The problem in my opinion is that the game feels mostly like a single player game. Interacting with other players is limited to checking enemy boards to see what's taken, pacing (indirect), pivoting or switching composition mid game, and positioning--that's it." The challenge is increasing player interaction without making it mechanical.

Looking at the future:

1. "[positioning] is gimmicky...[and devolves into who can micro faster]" SUGGESTION: "Highlighting the opponent you’re going to play [next] is a good idea. I also think they should freeze that opponents board (to [all players'] vision[, NOT to their actions]) with ~15 seconds left."
2. "I feel like giving players more control over the automated behavior of their team might be the future of the genre, because it would cause players to be playing against each other, rather than just playing against the games rng, regardless of the development of any meta." Here lies the opportunity for Riot. Without developing advanced artificial intelligence, TFT units will always be way more stupid than even a stupid human. This was just as true in Starcraft, but in Starcraft units (more or less) fight the way you want them to fight. Of course, Starcraft is very mechanically intensive. We are looking for a middle ground between ‘micromanaging battle commander’ and ‘release the NPCs!’, beyond understanding how NPCs will behave and then positioning them on a hex-field.

CORE SUGGESTION: Automation Settings for strategic depth and player interaction

A picture (or example) speaks a thousand words. Imagine a drop-down with options [1/2/5/10/20/30/35/40 seconds]. That’s how long all your units will sit, attacking any enemies in range, but not auto-repositioning. This would delay assassin jumps as well (but they remain invisible for first ?20? seconds of match, or maybe as time passes they randomly become visible, or until they jump).

An objection: would this be balanced? What about RFC carries? I recommend balancing unit and item capabilities around player capabilities. In other words, buff players, nerf NPCs.

There is probably room in the UI (and without making the game overly complex) for 4-6 automation settings (each defaulted to current way the game works). Other ideas: a) 'at 1/2/5/10/20/30/35 seconds retarget ally All/1hex-range/2+hex-range units onto enemy with highest damage dealt if they are targetable’ b) 'on enemy unit death, retarget to enemy with LowestHP/LargestRANGE/LargestDMG if within 1/2/3/4/5 hexes of repositioning’ c) toggle '1rng units automatically reposition to block the path of enemies from getting in range of ally unit with LowestHP/LargestRANGE/LargestDMG'