modest creek
I need some help debugging my Client prediction setup, Ive added a little video to help detail one of the issues im facing, and im just wondering if anyone has any idea what is causing the discrepencys & if they have any smart ways of debugging a complicated system like this?
If you havent seen my game before its essentially supposed to be a semi competitive FPS on breakable boats, the PVP aspect of this has lead me to persue Server Authed movement with client prediction & reconciliation, Due to my players movement being heavily influenced by the movement of my boats this has also required me to consider my boats in the prediction and reconciliation. Im slowly debugging the discrepancy between Clients Scene & Server, and have something fairly stable but there are still some jitter during certain events, which suggests re simulation is not working as intended or that theirs an actual difference in the Clients View Vs the Servers View.
Im currently working on debugging the boat part of my whole setup, As you can see, when splitting a boat apart there is small but noticeable jitter, I believe its likely related to my re-simulation of these boat parts somehow.
if you look closely on the third destroyed segment the boat splits in half, at this point for a fraction of a second both parts do a strange jitter.
Breakdown of how Boat objects are destroyed:
BoatObject Destruction event
Resim -
Initially Turning the BoatObject back on if ServerUpdate is before destruction event, re-adding joints & applying Servers VelocitysEventually re-destroying the BoatObject again which should lead us to a fairly similar state.
(I do not believe the issue shown here is due to the non deterministic nature of NGO Physics, as the difference here is quite noticeable, and most discrepancy which derive from the non determinism result in a much much smaller difference )
void idol