Technical solution to eliminate desync in single-player sessions

"

SkyCore wrote:

"

qwave wrote:

I promise that allowing the client to calculate combat is not a million dollar project. Hundred thousand dollar project, yes.

"

qwave wrote:

Even if the core game rules are tightly coupled with their server architecture, this sort of code is almost always extremely easy to move to the client. I'll leave it at that.

So you went from 'i promise' earlier on this page to 'almost always'. The number of times you have contradicted yourself in this thread is absurdly high. How can anyone who does this be taken seriously. It is a complete invalidation of your credibility.
I will agree to this; PERHAPS it may cost hundreds of thousands. Likely it will cost millions.
Regardless, it is not your decision. You do not have the data available to correctly weigh such things even if it were. Who knows what GGGs finances can or not afford.

SkyCore, your constant attempts to make qwave look like a fraud and like he's wasting everyones time are pointless and only make you look like a jerk.
Its pretty clear he's only got the best intentions and is trying to help.

Wait who is Jason now?

I like this topic so please... stay on topic. Save the hair-pulling for PMs or something.

Ok now, regarding the fp rounding problems, found a pretty interesting article (well not that hard to find, one of the first topics on stackoverflow) with comments from many game developers regarding how they managed (or not) to solve the problem:
http://gafferongames.com/networking-for-game-programmers/floating-point-determinism/

I always thought the excuse against desync was a pretty poor one. My own pet peeve on the subject is that sometimes an action can show you where a desynced mob is without updating the mob's position. For example, throwing a trap at a desynced mob will throw it to the actual place the mob is (and detonate on them). If my client is being told where it is, why the hell doesn't it update that position?

Seems like a lot of situations where the client could at least be smart enough to spot a desync and automatically /oos. If I cleave a mob at point blank and the server doesn't send back a hit/evade, the client should be able to guess that it's desynced.

"

LogoOnPoE wrote:

Ok now, regarding the fp rounding problems, found a pretty interesting article (well not that hard to find, one of the first topics on stackoverflow) with comments from many game developers regarding how they managed (or not) to solve the problem:
http://gafferongames.com/networking-for-game-programmers/floating-point-determinism/

This is cool thanks for this.

"

found a pretty interesting article (well not that hard to find, one of the first topics on stackoverflow) with comments from many game developers regarding how they managed (or not) to solve the problem:
http://gafferongames.com/networking-for-game-programmers/floating-point-determinism/

Yeah this is a pretty famous article on the topic, with pages and pages of quotes from major developers.

"

Wait who is Jason now?

The poor moderator trying to sift through this thread. ;)

"

Seems like a lot of situations where the client could at least be smart enough to spot a desync and automatically /oos.

Yeah, they could probably re-sync a lot more frequently, but they mentioned that it would cause a lot more rubber-banding. Ugh, the whole thing just makes me cringe.

Back on the topic of pathing (I believe this to be a pretty big cause of desync), could we not analyse the points of desync and map the x,y to a set path?

"

Baxta wrote:

Back on the topic of pathing (I believe this to be a pretty big cause of desync), could we not analyse the points of desync and map the x,y to a set path?

I guess you missed my video on this like 30 pages ago. Anyway, if you look through my posting history you'll find a good example of pathing/desync. I didn't get any responses on it yet tho!

"

whrsmycoffee wrote:

"

Baxta wrote:

Back on the topic of pathing (I believe this to be a pretty big cause of desync), could we not analyse the points of desync and map the x,y to a set path?

I guess you missed my video on this like 30 pages ago. Anyway, if you look through my posting history you'll find a good example of pathing/desync. I didn't get any responses on it yet tho!

Thats a great example of what im talking about. So far they have said that player input is mirrored by the server and cause of desync can be when the server determines a different path around an object compared to the client.

I was wondering if they are able to catalogue these hotspots and have conditional constant path route.

I'm trying to stay out of the whole "make just single player a better experience" thing (although really informative).

D2 relied on client processing and had lots of hacks.
PoE relies on server processing and has a bad player experience.

Finding the right balance for today's hardware is the key. PoE went the safe way to prevent any hacks and I get that. It's just not going to be a great game unless they give up some of that control to the client and implement thresholds to limit hacks.

Not a big deal if two enemies shared the same space for a second, so don't desync and move one of them screens back. If it happens more than the threshold, then you take away control from the client on the account level. Do we really care if someone can move one enemy per minute? Or that they avoid 5% more stuns than the server says? Feel free to change the thresholds to catch them cheating too.

As it stands right now, PoE will be good for a while until the desync gets on your nerves too much. Then you move on and sometimes check the forums for posts like this.