Having worked on CMC movement for 5+ years now, I think the average games people make will benefit a whole lot more from the types of things you can make in a client auth model, and the ease of making them, will be outweight by the `slower` cheat detection/protection model of heuristics

Now, obviously bad connections happen and clients might `look` like they are cheating if they miss some gameplay state, in a client auth movement scenario, but that's why we can track things over time.

if 9 clients get their movement speed reduced by 40% when X happends, and the 10th player is still at max speed, ban that guy.

Sever auth movement is the holy grail and prevents cheating, but it also prevents a whole lot of gameplay mechanics and feel things

yeah, rope physics door puzzle game type thing.
Try to port to 5.5 and make a small little demo

come December I've got holidays and plans with @seannoonan.bsky.social to start working on a thing again.
so there is a chance I post things :D

oh no

One downside here is without C++, you are stuck with a UObject that has no world. And therefore you cant use functions that require a world, like trace functions.
You can create a C++ base class from UObject, which has UCLASS(meta=(ShowWorldContextPin)), and subclass that in BP.

It works best on objects that get reinitialized or are CDO by themselves.
So things like actor/scene components, widgets work well.
Pure UObjects as well In the vid above I basically make a double jump inside this UObject

This is how it looks when it's working, though. It's SUPER COOL!

The math on this was kinda wild, @mrtapa.bsky.social helped a TON. His solution ended up far more elegant. My initial efforts were... uhhhhhhh, less good. But it works now! It saves me SO much time!