Technical solution to eliminate desync in single-player sessions

"

grogs123 wrote:

That's bollocks. If it was absolute crap he wouldn't reply, why would he waste the time?

Because he is spreading misinformation about desync. GGG have been spending so much time trying to accurately explain why desync occurs, and then this guy comes out of nowhere with buzzwords and long posts and irrelevant white papers claiming he has fixed some fundamental issue in CS

GGG have already spent enough capital trying to explain why desync exists, why its a problem, and why there is no magic bullet for it

Then this guy comes out of nowhere claiming that he has some magic pixie fairy idea that fixes all of these desync problems.

GGG is right about this, there is no solution for desync apart from changing the network model of their game

Yeah, I made this entire proposal just to troll deteego. Best long-con ever. You caught me.


You're right, GGG have developed an infallible network system and no mere 'community member' has the knowledge to resolve it. None of us are professionals on the topic, we're just lowly gamers / consumers.

"

qwave wrote:

Yeah, I made this entire proposal just to troll deteego. Best long-con ever. You caught me.

You made this thread because you are delusional and think you know you are talking about when you don't

Even worse your arrogant, you claim you know more than other people on this subject (hint, being arrogant and wrong is not a very good combination, being arrogant and delusional is even worse)

(also I am not mad, why would I be?, your some random buffoon on the internet who in my eyes talks the talks, but doesn't walk the walk)

"

qwave wrote:

Yeah, I made this entire proposal just to troll deteego. Best long-con ever. You caught me.

Sarcasm detected.
Analyze post content: . . . . . . . . . . . . .
None detected.
Request thread lock.

"

HellGauss wrote:

"

The main source of desync is latency.

No, the main sorce of desync is lack of info. The client must predict server response (= big amount of data) to know the effects of its action, instead of precisely calculate these effect by itself.

The core of the idea of OP is so simple that it actually blows my mind to read how it is misunderstood, even if I must admit that OP did not explain it so well. I do not know... maybe there is a global misconception about the fact that a 'real time' simulation of the game should be on the server. This is not the case, a real time simulation should only be on the client to ensure playability.

I understand the OPs proposal. And still, if there was no latency, the client wouldn't need to predict anything, as the server would instantly know the client action, and the client would instantly know the response, so it is latency :)

"

I understand the OPs proposal. And still, if there was no latency, the client wouldn't need to predict anything, as the server would instantly know the client action, and the client would instantly know the response, so it is latency :)

The biggest misconception comes from the fact that people believe the server MUST tell the client 'WHAT' to do. In reality, the server only needs to act as a referee. The entire proposal is built on the concept that the client informs the server of what's happening, and the server does not need to tell the client the game state.

qwave, would you please do what you promised and stop responding to deteego and SkyCore, it's just derailing the thread.

Sorry LogoOnPoE, will get back on topic. =)

"

LogoOnPoE wrote:

I understand the OPs proposal. And still, if there was no latency, the client wouldn't need to predict anything, as the server would instantly know the client action, and the client would instantly know the response, so it is latency :)

OK, this is more precise. However in this case i want to point out that PoE as a 5+ sec latency, this is the frequency of updates. Client update server with 50-200msec latency, but server update [edit: fully update] clients with 5+seconds of latency, or when \oos is cast. The F1 stats only display the first latency.

"

qwave wrote:

"

I understand the OPs proposal. And still, if there was no latency, the client wouldn't need to predict anything, as the server would instantly know the client action, and the client would instantly know the response, so it is latency :)

The biggest misconception comes from the fact that people believe the server MUST tell the client 'WHAT' to do. In reality, the server only needs to act as a referee. The entire proposal is built on the concept that the client informs the server of what's happening, and the server does not need to tell the client the game state.

I'm talking about the current situation, not your proposal, when i talk about latency. I understand your proposal and how that would work. Well except of your proposed leniency for fp rounding. That one wont work imho. There just needs to be a way to eliminate the rounding errors.