[April 12] Initial Development Manifesto Feedback

"

pdbee wrote:

No matter how you turn it around the sync solution of PoE is plain bad and will kill it in the longrun.

To me comparing this to D3 is of essence. The D3 is more dead to me than PoE. Desync is not pleasant, but unwillingness of D3 developers to do such a simple changes as really introduce some good end-game or opt-out from auction house system is much more heart-breaking. Not to mention absolutely awful community of ppl playing D3 (with some bright exceptions).

I'm still convinced that PoE can overcome these "game-breaking" problems at last. While for D3 I'm convinced there is no way for Blizzard to save the situation. :-)

Moreover GGG can ask the community of gamers to help them with this desync issue. I'm myself very skilled C++ programmer and I'm sure there is more and that there will be some expert in network communications kind enough to help them to resolve what they need to do to fix this desync issues.

is it possible to just add an option in the options interface to use the /oos command every 1/5/10seconds?
would also be a nice temp solution( wont help with desync exactly but it would lower deaths from desync)

"

What we're planning to do to improve synchronisation

...

Have you thought about having the server action calculation taking into account the current delay (ping) between client and server? So in that walking-around-the-rock-example using the position the client should be at the exact point of time the new action is started to be calculated on the server instead of the position where the last received packet says the player was.

Of course it would require the server to have the whole movement trajectory for the predecessing movement action calculated as soon as that action request reached the server. But I guess your servers do that anyways, as it is required for pathfinding too.

"

bartekltg wrote:

"

Pinchyskree wrote:

Why would anyone enjoy watching the combat stop every time something valuable drops and every self centred person crowd around it like flies on shit hoping the assigned person doesn't show up? Stupid.

Who tell you combat would stop?

No one TOLD me, it actually fucking happens! almost every dock run.

The smaller the space, the more likely desync will happen.
For some reason, hit boxes on the server seem inconsistent with the client. Get surrounded by squids in normal on a high level and try to move through, generally you will.

I heard about the desync conversation about stun mechanic, chance to hit, etc. that made D3 so easy to do. There is still desync but there aren't a lot of factors that can cause it.
It really is true that when you look at all the mechanics that D3 got rid of, it really made desync easier to fix. It allows the code to be simple and not check for a bunch of different what if scenarios. But you get an extremely simple game in the end.

A big skill to note for desync is actually using frenzy on a ranger. The delay will become noticeable because of the delay between getting the charge and when the attack speed applies on client and server both.

Trying to type this on my iPod so bit messy but ill probably expand on this later

After reading the thread about upcoming race events, I don't think I saw any mention of progression-focused races. This has been a request of players since before open-beta. I even plan on making a lengthy post in Suggestions about it because I want to see it so badly. A race where the winner is dictated primarily by his content progress and secondarily by his experience ammount would be more fun and hectic even on normal mode.

Thanks for adding a non-ninja system! I've been playing with a friend and asking him to not take loot that rolls to me so I can relax a little bit (while playing with a friend), but due to some extreme, deep seated autism, he refused to play outside of the loot system you designed, because "that's how you're supposed to play the game". So thank you for adding a new loot system for all of us here with autistic friends!

This may have already been suggested, but I wrote a university paper on this very issue a few years ago and came up with a solution that worked fairly well.

Most of your current implementation would remain the same. The only change would be that if the client was out of sync (and the server knew that the client was out of sync), and that discrepancy was within some allowable range (something reasonable but small, so that speedhacks and the like wouldn't be worthwhile), then the client actually corrects the information that's on the server - not the other way around. This would make it appear to the player that nothing changed, and the resync action goes completely unnoticed.

(Disclaimer: Obviously this solution is intended for non-critical things like movement, combat, etc, and a client should never be trusted for things like trade.)

These are some of the best posts I have ever read on any developer's boards ever. Excellent job, you write much better than Brad Wardell, whose similar posts I respect.

Doubting you are going to be able reading 100+ pages of feedback, I will post mine as briefly as possible nonetheless.

Your post about desync was divine; as a non-technical person, I found it both very informative and easy to understand. One small question I was unable to induce from your post:

1. Why do some entities seem to rubberband more than others, particularly Gneist and other rolling monsters?

I enjoyed your other posts as well. Particularly interested in solutions to the discrepancy in success and danger between Ranged and Melee. While I believe it is acceptable to close the gap without it ever being entirely equal, simultaneously avoiding the pitfalls you mentioned, I am curious as to why you did not mention one possible solution:

2. Isn't there one way to attempt to address the issue without attempting to tweak any of the skill gems themselves (which will be difficult) or the passive tree (which will be nearly impossible, because as you correctly point out anyone can take anything)? Why couldn't you, for example, tweak things that ONLY melee users utilize -- e.g. the base damage output of melee weapons? What if you were to simply increase ALL of the damage output of all melee weapons by x% across the board: wouldn't this be a benefit which would require (relatively, comparably) little balance adjustment to ensure that it did not inadvertently benefit ranged users?

"

Halcion wrote:

This may have already been suggested, but I wrote a university paper on this very issue a few years ago and came up with a solution that worked fairly well.

Most of your current implementation would remain the same. The only change would be that if the client was out of sync (and the server knew that the client was out of sync), and that discrepancy was within some allowable range (something reasonable but small, so that speedhacks and the like wouldn't be worthwhile), then the client actually corrects the information that's on the server - not the other way around. This would make it appear to the player that nothing changed, and the resync action goes completely unnoticed.

(Disclaimer: Obviously this solution is intended for non-critical things like movement, combat, etc, and a client should never be trusted for things like trade.)

Actually you are far underestimating the abilities of hackers. WoW has been out for nearly a decade and still has teleport, fly, and speed hacks, all because they trust the client. Once you give them an inch they will find a way to exploit it to take a mile. Trust me, hackers are geniuses and have LOTS of time on their hands. It's a never ending losing battle and in most cases you cause far more harm to legit customers than you do to punish the few bad ones. Plus in a F2P game its as simple as make another account.