0
The above link is a related thread, but this is worth having its own thread, since this bug report should be in this forum instead.
I had a recent game (8-man) and there was an early cap on Assault in Forest Colony Snow. Those who participated on this match can feel free to comment about it...
What happened exactly was that there was a death near the cap (it happened to be a 3L)... and for some strange reason.. the game (specifically the server) continued to think that the mech was capping our base... despite it being dead. Only until the other capping cap (another 3L) was at the base again soon after and it eventually died... and the "base capping" by nobody had stopped. Mind you, there was no ECM spider involved here...
Anyways, I believe this is the problem. The server for some reason occasionally does not realize to "change states". For example, if a mech has AMS going off at incoming missiles, and proceeds to DC/CTD, the server continues the "AMS" firing process... it seems to ignore the fact that no more missiles are firing at it. The same goes with DC/CTD players that have their mechs going in one direction or circles... the game uses their "current state" to dictate what the mech continues to do.
Therefore, the server isn't very good at clearing the "current state" of capping a place. If a mech had spawned and/or died at the cap location, there is an oft chance that the cap location is "stuck" because of the rules of capping a location (an enemy mech can only stop a cap... in Conquest, you cannot change a cap location's affinity).
Anyways, this is probably also why you may get an overheat death or a prolonged shutdown state. The server does not attempt to give a repeat message to the client that their state should change.. so it happens under heavy packet loss or server duress...
This may also be why the thermal/night vision turns themselves off (does this still happen?).. the server for some reason thinks you may have turned it off (or that it was "not turned on" to begin with)...
This needs to be addressed at some point. The way you would test it is to kill mechs that are capping or cap-stopping. Eventually, the server will screw the messages up and make a point uncappable or always capped (depending on the situation). This doesn't always happen, but I believe this is easily reproducible.
Edit: Also, I want to point out something else that is related...
If you have played MW4 and suffer from bad lag, you may recall that you suddenly stop into place and you kinda "reset" yourself with the speed. In MWO, packet loss causes you to "backtrack" a little, but it keeps your speed and direction/motion constant while being in a "mech limbo". These are different behaviors, and may or may not necessarily be good... but that's just for comparison sake.
Edited by Deathlike, 28 April 2013 - 12:40 AM.
