Lag Reduction Suggestion: Converting SPA 383 Procs

Problem with that is it can then potentially be interfered with, which is why it's server-side.

Can't they just enable some cheat-detection software, and enforce their anti-cheating rules?
I assume DBG has functional cheat-detection in their other titles?

The EULA allows it: https://www.daybreakgames.com/terms-of-service?locale=en_US#section10

LOL like that works on the afk bots.

Im sure it works, that they don't enforce it for EQ1 is another matter. My point is that there is a "relatively simple" solution for the lag problem, and all they have to do is to prevent it being exploited.

Isn't their other titles relatively cheat free? Their stance on cheating in EQ1 is holding them back when it comes to lag-problems, not technical issues...

Thick client + thin server + aggressive anti-cheating, is the way to go for optimal mmorpg architecture, imho...


On the original topic, I would love it, if my spell-casts wouldn't spam my combat-log so much, combining procs in the spam-text would ease read-ability a lot!

Not about whether I want to see it or notm, the question was, how much lag does thousands of meaningless web connections and spam cause?


Lag thread is about lag.

None. Find a new dead horse.

They really need to focus on fixing the lag before focusing on class balance tbh. While yes there are overpowered and underpowered classes we really have no idea to what degree until lag is fixed.

This is an interesting idea. However it seems like it would take extra effort for them to implement a new stacking SPA.

Couldn’t the same thing be achieved by just removing all sympathetic procs and replacing them with absurd Spell damage/Heal Amount/Clairvoyance values? For example the Type 19 Casting Ice you mentioned remove it and add 5000 Spell Damage to that one augment?

I’ve always thought those Mod3s were vastly underutilized. One could imagine a world where they introduce new AA lines to impact the effectiveness of these mods. Amplifiers to total Spell Damage/Heal Amount values, increased chance for Clairvoyance to trigger, activated short AA lines that make Clairvoyance trigger 50% of the time or 100% of the time, activated AA lines for priests that cause huge negative spell damage for huge positive heal amount and vice versa. Etc.

It seems like the goal of adding base damage to each cast or some percentage of casts could just be balanced out already through massive spell damage values on specific items that already have sympathetics.

Also while on the topic if you could replace the Druid 6-minute proc aura with instead a normal permanent aura that amplifies spell damage that would be great!

So: reducing serverside cascading and potentially interdependent events, by moving calculations to either the event of equipment change, or the event of swing/cast.

If this is correct, it could scale quite well: a character could have 10x the number of current augs with the new SPA with no noticeable impact to server performance, because the former proc complexity is condensed into a single number, calculated on equipment change.

This seems to make future game design a bit easier, with more options.

An apparent advantage from an execution perspective could be that the old proc-system could be left in place, while the new SPA immediately starts to benefit gameplay, without having to redo all historic content in one go.

By redoing the proc effects selectively, the most likely lag-culprits could be targeted and converted.

It could look like this: ask 3 end game guilds to make a prioritized list of top 20 proc effect items they believe are the most proccing and cascading. Convert these items to new SPA, eg 5 pr development cycle. Continue until list is exhausted, then ask for the next 20. I suspect that there would be some sort of 20/80 effect, where 20% of procing items drive 80% of the related lag.

Just layman speculations...

Read original post. Saw lots of math stuff. Went way over my head. Saw it was written by Sancus. Clicked like button.

@Jumbur,
A lot of interactions happens between the toons, and also the game has to stay consistent, even when there is lag, etc. These reasons alone weighs heavily against client-side combat resolution.

It would require a LOT more traffic being sent to the client, and as a rule of thumb, internet traffic costs a lot more time than processing a computation on the same machine. A barrier for making distributed computing is exactly this, which means that if you want to gain an effect by distributing computation over the internet, it has to be a problem that can be broken down into tasks that demand huge amounts of computation, but almost no traffic, relatively speaking.

Gut feeling: the lag is driven by cascading interdependencies, stuff that happens for a spell or toon has to be checked agains all other toons, that then triggers another intervonnected cascade, or perhaps that there are some bottlenecks between interdependent checks.

If its spells and spell procs that are the problem then it needs to be applied to all spells not just DD spells. A lot of spells these days proc heals, slows or other things off them not just DDs. That doesn't include items procing spells left right and centre.

They targetted caster procs, from which I have seen little to no reduction in the lag. I suspect the real culprits for the lag are the zone trash, swarm pets and raid mechanics checking too many things at the same time.

When the database server migration happened last year, the lag became unbearable.

While spell/proc is a likely contributor to the problem, I wonder if the source of the lag is the DB server migration. I don't know if the DB server itself performs more poorly or if some new middleware is in place contributing to the poor performance. IMHO the entire infrastructure and data service stack needs to be looked at - my gut tells me the fix to this issue is going to need more than a SPA tweak.

@ Iblis
True, my idea was based on the assumption that all the information needed to compute the damage, was already present on the client in the form of buffs/songs etc... It will obviously not work if there are some hidden "behind the scene" calculations that can only be resolved server-side.

A simple example, familiars: My guess is that familiars are currently handled by the server as a mob, where pathing is computed server-side. But there would no problem in moving it all to the client, and just handle it as a client-side GFX effect instead(who cares if 2 players disagree where the familiar is?)
Obviously, familiars are not a huge server load(it was only an example).

Unfortunately, the "combat-round-mechanics" are not publicly known, but Im guessing some of serverside computation could be moved to the client, if the devs dared to fiddle with it.
If it would be worth the effort is a good question...

Raid lag has been a thing for like at least 6 years at this point, it just gets worse and worse every expansion, so it's not like any hardware change since then went from lag free to lag filled

It certainly made it worse instead of better. Hardware changes in this day and age are usually upgrades and improve performance. Not take it backwards. So yes, I stick to my guns that the DB migration was a downgrade.

Just to answer a few posts that really don't belong in this thread.

1. Darkpaw has discussed that they are dealing with the "lag" problems for over six months now. Originally, it was, "We are upgrading our server/network hardware to fix these issues", and then more posts were made later about how they are aware of it and were working toward fixing it. Note that "lag" can cover many different issues that create lagging gameplay for players.
   
   So, they have made a number of posts on the lag issues and that they are working on it. They have not provided details, but they have said they are aware and working on it.
   
   Whether you believe they are giving it enough priority or not, they do know at least something and are doing something.
   
   
2. Having the client (your computer) handle calculations always opens the door to cheating. There is no lock that cannot be picked. I supposed if you threw millions of dollars at the issue and added public/private key encryption together with memory scanning / file scanning software on each PC (which would open the door for massive lag as your PC struggled with this stuff, as well as possibly open holes for malware). you could do it. Maybe.
   
   I mean the media companies went as far as creating an entire High-bandwidth Digital Content Protection system, which required specially designed computer video cards and monitors to protect DVDs. And it still was cracked.

A bit better than good. Maybe you are too out of the loop to really be engaging in this kind of discussion if you have to ask if monks are good.

Yes cannot be expert on every class now, like you

Question/Suggestion:
What if they started adding much higher delays on melee weapons while also adding much more damage to the weapons as well. Less swings = Less lags? More importantly, would that make SKs less fun to play because they would be healing less often but more when they do hit?