What I did is store up the snapshots in an array so there should always be two snapshots there to interpolate between, you just have to figure out how much of a delay to give the interpolation to all but guarantee that there will be two snapshots there to interpolate between. Right now I’m using TCP (I know, gross) :
to send the snapshots but I’m eventually going to add a UDP layer to the network to send over the movement snapshots, and after that I should be able to decrease the latency quite a bit.
So right now there’s no extrapolation, but I’m going to implement it in certain situations like in getting hit so the client-side can guess where entities are going to be for important things like that. Right now the latency is a little bit too much for hit reactions
I’m telling you, it’ll be a success. The quality of a game doesn’t depend solely on the gameplay, it mostly depends on the graphics, because that’s the thing that draws in players and entices them. If the gameplay and logic is the head of an octopus, the graphics would be the octopus’ tentacles XD

