They could reduce lag right now if they would admit their mistakes

[IronClan]

Lag. Projectile burst is harder on the server than constant rate of fire. Sev~

So by extension increased number of attacks of all sorts that have physics checks would be harder on the server? For example speeding up attacks? Eladrin agreed with you, it's why he felt the need to remove stacking alacrity.

One of the best ways to lower projectile burst is to give ranged players options that boost per projectile damage at the expense of the number of projectiles.

You know the same way you guys could have kept Physics checks on melee attacks from ballooning by giving bonuses that created larger damage numbers instead of more alacrity increases?

At least you guys are catching up to where Eladrin was. When can we stop this charade?

[IronClan]

Once again don't take it from me take it from the guy who actually got rid of so-called "DPS lag" that was plaguing Shroud runs at the time:


These characters currently perform many attacks very quickly, and each attack is pretty heavy performance-wise.

We’re also considering introducing a new mechanic to replace some speed bonus effects, called double strike:

The Fighter Alacrity capstone, and the Paladin spell Zeal would be changed to a +10% bonuses to double strike.

For those that are interested in the “how does that help? You’re still making 11 attacks in 10 swings” – double strikes, like the proposed off-hand attacks, would bypass the additional physics detects and secondary characteristics of normal attacks such as glancing blows, while still providing an increase to overall damage over time.

Irrefutable evidence that increasing attacks over time of popular styles* is directly responsible for the Lag that not coincidentally got worse every update where they added alacrity.


* such as Swashes, SWF Barb's and Paladins, Repeaters and now Warlocks with their chain shot (SWF = 30% more than TWF, Vanguard = 25% more, Crossbows, who knows how much faster they are, but it's a lot, and I don't even want to guess how heavy a warlock is, we know that excluding them from raids has lowered lag anecdotally).

[IronClan]

In the interest of not derailing the official thread where Sev made that comment there's also this


TWF was nerfed years ago in the name of fixing lag. It didn't. Then to make up for the huge lack of dps things like alacrity and doublestrike were added. It feels like we are chasing our tail here.

Alacrity was removed in that update. It most certainly had a lag reducing effect. In my experience anyway the game went from laggy in DPS dense situations to almost lag free for many years.

Lag is not "fixed" or not, it's not a switch you turn off or on, no more than reducing darkness with candles fixes nighttime. The more you add the lighter it gets. Lag has degrees and it gets worse or better but it's never going to be completely gone. Much like those changes which reduced lag and helped make what was commonly called "shroud lag" and also "DPS lag" at the time become just an old term no one needed to use any more (until about a Year ago when Swash reintroduced it) because it was no longer a thing. Multiple changes are required to improve latency just as multiple changes has increasingly made it worse lately.

For example when they added SWF and Swash complaints of lag increased and people such as myself who RARELY experienced lag beyond the odd rubber band or short glitch, started to see more lag and complain. Then they caused a popular S&B style to have 25% alacrity with lots of average players gobbling up Vanguard in the name of making their formerly gimpy S&B characters much more fun and viable. And lag got a little worse still, more frequent occurrences, much more likely in 12 man instances. Then they speed up crossbow animations which also added more, at this point as someone who had spent 4 years playing DDO and considered the lag normal for most of that time. I grew to the opinion that they had pretty much made it into the laggiest MMO I've ever played (for more than the time it took to uninstall). Warlock seems to have added a few more nails to this coffin.

Sadly

[Drwaz99]

Yeah, I agree. Admitting mistakes is a seemingly no-no for this group of devs; they just can't seem to do it. The funny thing is, most people respect honesty and can easily sniff out BS'ing when trying to avoid it.

They also seem to like to repeat the things that cause them problems. The attack speed you mention and I have my own little theory about teleporting monsters. I have no evidence, but I do know that the shadows caused lag and were adjusted multiple times to correct lag in Temple of the Deathwyrm (and lag is still there) and yet they create a new raid and a majority of the monsters teleport all over the place (and have been adjusted multiple times).

It's like their design motto is "Do The Same Thing, Pray for a Different Result".

[Beethoven]

Once again don't take it from me take it from the guy who actually got rid of so-called "DPS lag" that was plaguing Shroud runs at the time:



Irrefutable evidence that increasing attacks over time of popular styles* is directly responsible for the Lag that not coincidentally got worse every update where they added alacrity.


* such as Swashes, SWF Barb's and Paladins, Repeaters and now Warlocks with their chain shot (SWF = 30% more than TWF, Vanguard = 25% more, Crossbows, who knows how much faster they are, but it's a lot, and I don't even want to guess how heavy a warlock is, we know that excluding them from raids has lowered lag anecdotally).

Out of context quotes based on an premise that code, once written, could not possibly ever be changed or modified and that solely the number of computations but never their complexity matters do not provide irrefutable evidence of anything except your own ignorance of the topic you are trying to discuss.

Also, going on about the same topic mindlessly idolizing one developer while lashing out at any- and everyone who may ever have disagreed with him makes you sound like a spurned lover who never got over a past affair than a professional analyst.

You omit the context where Eladrin was specifically talking about how the physics checks caused by these additional attacks made them so heavy performance wise and how these checks were reduced by making offhand attacks "piggy-pack" on mainhand attacks as well as the implementation of double-strike which too does not have a separate physics check.

So, while your correct in that every computation requires /some/ memory, your analysis fails to take into account so many relevant factors and relies so heavily on out of context quotes that it becomes entirely irrelevant to any attempt of intelligent discussion no matter how often you repeat it.

[dunklezhan]

You omit the context where Eladrin was specifically talking about how the physics checks caused by these additional attacks made them so heavy performance wise and how these checks were reduced by making offhand attacks "piggy-pack" on mainhand attacks as well as the implementation of double-strike which too does not have a separate physics check.



Yeah, manyshot I'm pretty sure only makes the one physics check, although it certainly makes separate 'to hit' calculations. Does doublestrike/shot make separate attack rolls? I kinda thought it didn't. Which would add to the effect of reducing lag, if so.

I do agree with adding to per-projectile damage too. But at least as far as AA's are concerned, isn't that exactly what's proposed anyway?

[PermaBanned]

So wrong. So, so wrong...

They could come on the forums right now, and post in this very thread that you are so so right and they effed it all up... and the lag would remain unchanged. "Admitting their mistakes" would have no immediate or appreciable effect on lag. What's more, simply ballooning the damage from a single arrow to x4 would do nothing to address the lost chances to proc things. Nor would their admission do anything to mollify the outraged players who must be nerfed to reduce lag. I also remember the Eladrin days, and the TWF nerf - what I don't remember is the throngs of appreciative players saying "Yes! The nerf is so good now that the Shroud is less laggy!"

Have they effed it up? Obviously. Will openly stating it accomplish either of the following:

• Lag disappears?
• People are willing-if-not-happy to be nerfed now?

I'm betting the answer to both is a resounding no.

[IronClan]

Also, going on about the same topic mindlessly idolizing one developer

You omit the context where Eladrin was specifically talking about how the physics checks caused by these additional attacks made them so heavy performance wise and how these checks were reduced by making offhand attacks "piggy-pack" on mainhand attacks as well as the implementation of double-strike which too does not have a separate physics check.

30% alacrity for SWF (which if you haven't noticed how many Melee's are running around with Orbs is very popular) causes 30% more physics checks than TWF, 25% for vanguard likewise in addtion to any that are added by bashes (can only speculate), ditto for increasing crossbow ROF, and god only knows how computationally intensive Warlock spam is, we only know anecdotally that removing them from DoJ eases lag according to reports on the subject.

As far as idolizing Eladrin.... you mean the guy who wouldn't listen to S&B players and went ahead and nerfed my main at the time character causing me to take a year off from the game? That Eladrin? You're just making up an ad hominem to direct the discussion personally.


BTW when first confronted with this evidence Sev said:

SWF still isn't generating as many attacks as TWF so I doubt it's that.
Sev~

He has admitted that he was wrong in that quote at the time he did not realize that TWF off hand procs do not have physics checks. So in fact TWF is 30% slower than SWF.

I like Sev, in fact I like him a great deal more than Eladrin. However this matter is about someone making a mistake that has worsened DDO's latency so despite the fact that I like him I also feel convicted that a mistake was made. If you felt 100% convinced that someone made a mistake and THOUSANDS of DDO players are suffering poorer gameplay because of it would stand by your conviction? Or because you like Sev do you just say nothing as people lag out far more often and far worse than a year or so ago when they made Swash and SWF popular and at the same time Lag got noticeably worse? Personally I stand by my conviction, even with people such as yourself trying to insult and denegrate me for being stubborn or repeatedly bringing this up. I know it sucks to consider the possibility that a mistake might have been made that inadvertently worsened the game but it does happen... Everyone makes mistakes sometimes maybe I'm making them when I post about this... I'm willing to be wrong but I am really quite sure I'm not.

[IronClan]

Have they effed it up? Obviously. Will openly stating it accomplish either of the following:

• Lag disappears?
• People are willing-if-not-happy to be nerfed now?

I'm betting the answer to both is a resounding no.

Okay hold on, if they admit that adding alacrity was a mistake they don't even have to nerf anyone to fix it... just replace the alacrity with a comensurate amount of something that has less network traffic footprint such as DS which as we know was invented just exactly to replace alacrity with something less impactful, +W, Melee Power, higher Strength Mod... Any number of ways to replace it.

No nerfs are being asked for... hell you should know me better than that but then again you didn't even get my forum nick right so maybe you don't:confused:

Here's how I see it, you seem to agree that alacrity has worsened lag (and why not increasing network traffic should cause peak traffic situations to be more destabilizing) but you don't want to agree with me because you don't like the messenger... Which one is the facepalm worthy of I wonder? You disagreeing with me even though you agree, or me harping on a subject that might be a lost cause due to Developer hubris?

[Axeyu]

DPS lag was caused by some articial limit to bandwith or packages sent/received (I don't remember the details) and was fixed by removing/changing said limit. It never existed on the EU servers.

[Knobull]

We did not have this problem pre-MOTU. When the physics engine was changed and the stuck-in-wall-bug appeared it all went bad. We still have the all mobs are incorporeal until they attack problem too (usually after they pass through you then turn around). The lesson should be: do not use libraries you have not written yourself in your software projects, for the quality and design purpose of the code is not in your control. If you want it done right, you have to do it yourself. Also, it is not "spaghetti code" it was tight and efficient functional programming. Get off the object-oriented library using lazy bandwagon, and you will be on the road to recovery.

/Me steps down from soapbox.

[IronClan]

DPS lag was caused by some articial limit to bandwith or packages sent/received (I don't remember the details) and was fixed by removing/changing said limit. It never existed on the EU servers.

Indeed that was part of it, might have been the greater part of Shroud lag, but clearly it didn't go completely away as even CitW and FoT had it's bouts of DPS lag.

Lag gets worse or better it's about degrees, and IMO it got degrees worse when new alacrity upped the physics checks that happen in big fights and high traffic situations. The correlation is too strong, lag complaints have gone way up in the last year, and we know that at least one Dev removed stacking alacrity because he felt it contributed to the bottleneck that happens in high DPS situations.

[PermaBanned]

Okay hold on, if they admit that adding alacrity was a mistake they don't even have to nerf anyone to fix it... just replace the alacrity with a comensurate amount of something that has less network traffic footprint such as DS which as we know was invented just exactly to replace alacrity with something less impactful, +W, Melee Power, higher Strength Mod... Any number of ways to replace it.

No nerfs are being asked for... hell you should know me better than that but then again you didn't even get my forum nick right so maybe you don't:confused:

Here's how I see it, you seem to agree that alacrity has worsened lag (and why not increasing network traffic should cause peak traffic situations to be more destabilizing) but you don't want to agree with me because you don't like the messenger... Which one is the facepalm worthy of I wonder? You disagreeing with me even though you agree, or me harping on a subject that might be a lost cause due to Developer hubris?I'm disagreeing with you because admission is irrelevant to correction. Admitting they screwed up fixes nothing, just as fixing the problem requires no admission.

And yes, unless my understanding of % proc chance abilities is way off, fixing it will require nerfs. Does proc'ing a DoubleStrike/Shot create two charges of Blitz? Does poc'ing a DS/S create two charges of Adrenaline? Does a DS/S proc create separate opportunities for Colors of the Queen? Shield bashes? Stunning? To the best of my (in this case hopefully flawed) knowledge, the answer to the above is generally no. So unless I am indeed wrong, simply increasing damage or DS/S to compensate for fewer attacks doesn't really compensate for fewer attacks now, right? Less attacks would be a nerf just about any way you slice it, but less attacks really is the only way to reduce the lag generated by too many attacks - with or without any admitting of their mistakes.

As I said the first time, just admitting their mistakes would fix nothing. The very next post in this thread could be the admission you desire and the lag would be the same 5 minutes, 2 hours or a week from now. Nor would admitting it make people ok with the nerf. In fact, it would probably just make the whole thing worse, 'cause people would just rant and fuss along the lines of "You screwed up, and now you have to nerf us to fix it? What the {expletive}!"

[PermaBanned]

Oh, and I didn't get you nick' wrong, it was more of a punish thing as I actually envisioned the facepalm with a gauntlet clad hand ;)

Fwiw: My problem is with the message and not the messenger. Sometimes I agree with you, others not. This time - not ^^

[slarden]

I am not getting crushing lag all the time in the game. I get lag in specific quests/raids.

There are many things in the game that changed besides alacrity
1) more auras
2) more effects in general going through the combat section
3) larger monster spawns
4) more instances of monsters that always used to cause lag and now appear in greater numbers (horned devils)
5) more mobs teleporting around
6) more cleave abilities and the ability to chain cleave non-stop
7) Turbine's service provider Cogent became embattled with all the major ISPs due to Netflix bandwidth usage and disputes over congestion/cost of upgrades. Cogent itself even acknowledged throttling bandwidth.

If there was an easy fix Turbine would have made it.

I think most of have been in a raid that is lagged out for several minutes. One guy DCs and suddenly everything is better. It seems unlikely that is due to alacrity to me. I don't know what causes it, but I doubt your silver bullet is the complete answer.

[UurlockYgmeov]

They have admitted they made mistakes, but it is bad business to admit it publicly. Rather they have embraced Kaizen.
http://thumbs.dreamstime.com/t/kaizen-15798652.jpg

OOOO! another weekend popcorn thread! :cool::p

https://lh3.googleusercontent.com/-4foONJeV_kA/VgC_aVTuYEI/AAAAAAAAwBY/9TlNgEyPRcQ/w346-h194/15+-+1

[Deadlock]

I'm disagreeing with you because admission is irrelevant to correction. Admitting they screwed up fixes nothing, just as fixing the problem requires no admission.

This whole thread is just a repeated attempt by someone desperate to have their pet theory proved right. It has nothing to do with lag or improving the game.

No matter how many times evidence that things like aura and particle effects are the biggest contributing factors is presented, they're not interested in it because that doesn't fit their pet theory.

If you deign to point this out, you'll be accused of "ad hominem" attacks :)

You're basically arguing with a brick wall, it's completely pointless. It's not a debate or discussion when one side is so entrenched in their opinion that they cannot conceive of the possibility that they are wrong.

[Xionanx]

Actually I would look at it like this, its not an "either or" situation, its an "also does" situation.

Physics Checks - Cause Lag - NO MATTER THE SOURCE.. Be it "Swords" checking to see if they connect or "Arrows" checking to see if they hit or "Aura's" checking to see if the enemy is in range.. Its a PHYSICS CHECK

The question then becomes, which is checking more often and how can the devs go about fixing it:

Aura's - Cover and Area and have to check that area fairly often to see if an enemy enters/exits the Aura. The number of Aura's in effect + the effect that is applied increases this check exponentially. IE. a Warlock having 3 Aura's is effectively doing a 360 degree "Cleave" every "time increment" multiplied by the number of Aura's the Warlock has.. this causes Lag

Arrows - Cover a "Line" shaped area and either damage the 1st enemy they encounter, the enemy that is targeted, or every enemy in that line. This "line" if effectively "infinite" in length as bow range only seems limited by view distance. But.. its not actually a "line" its a "traveling marker".. Anyway, "manyshot" and "doubleshot" multiplies that marker by x amount.. when what it SHOULD BE DOING is mimicking double strike. Where you still only shoot ONE arrow, but upon contact that ONE arrow has a chance to do 1,2,3,4,5 arrows worth of damage.

Swords - we already know how this works.

There is also the "lag" introduced by persistent AoE spells, which for all intents and purposes should be equal to the lag of "Aura's".

Plus the "Mob AI Lag" which supposedly was a problem at one point which is why we now have dungeon alert. But is also a problem with "Teleporting" enemies. Every quest, raid, etc that has ever had teleporting enemies has had lag. Shroud/Tower/etc.. The enemy AI engine IMO needs a "Cooldown" flag placed in it where:

1. AI Decides to Teleport (NO DESTINATION SELECTED)
2. AI Ceases all other actions and goes into "Teleport Mode"
3. AI Monster Dissappears
4. AI Script "Pauses for 1 second"
5. AI Script Selects a Destination
6. AI Monster Re-appears
7. Teleport goes on a 30 second cooldown
8. AI Resumes actions

As it is now, I get the feeling that the source/destination are being calculated at the same time + the AI is trying to move/attack while its also trying to teleport. We have all seen enemies that are mid swing hitting you then suddenly they teleport and appear somewhere else only to "slide move" because they are still mid swing.. its just a lot of needless calculations going on. Basically the "teleport" action needs to be the ONLY action that a monster is taking until the teleport is over. I get the feeling tho, the reason monster and hireling AI are so... how they are, is because when the game was originally under developement they "Contracted Out" the "AI Scripting" part of the game.. I forget where I read that, but I'm fairly certain my memory of that is accurate. Anyway, due to that, the AI's have always been BAD and they have never been able to get the to work quite like they want them to. (if it were me, I would gut the entire AI system and rewrite it from scratch)

Basically.. EVERYONE IS RIGHT. Now we just need the devs to be consistent in their implementations of lag fixes.

[fatherpirate]

They need to run combat attacks like Vegas runs slot machines.

What you see, you drop money, pull handle
lights and sounds
first reel stops
second reel stops
third wheel slows to tease you
then stops
you get paid if you win.

What the machine sees
coin drop stops random number generator at a particular number
outcome fully decided - with prewrote script on how to show result.

got burst, client does ONE random number check - that decides the results of all the rolls as a package deal
might be 12 2 18 whatever, but ALL possible roll outcomes are assigned 1 number each...the program checks that.

[Jiirix]

I run DDO on an up-to-date gaming PC and have a good connection speed. I didn't notice any lag with manyshot so far. But I did notice lag in fights with mass spawns when AOE effects like Acid Cloud where up and the mobs spawned into it. There is definitely not "the lag" but many different forms of lag. Biggest Difference between Console and PC games is that PC games run on millions of different configurated machines. Therefore I understand that games are not 100% perfect when released but I don't understand why customer feedback on quality issues is handled the way it is.

I work in pharma QM and like every PC is a bit different from the next every patient is a bit different too. In both cases getting standardized and analyzable customer feedback is important for further development. Know your customer. .. that slogan works in nearly every industry. In the case of online games a “report lag” button sending a “performance log” could help. The log would have all the data the devs need to evaluate the cause of lag and be transmitted on logout? With a bit of data mining one should get good reports on the circumstances that cause the customers to experience lag. The attempts to fix lag in DDO always seemed like a stab in the dark to me.

@ UurlockYgmeov
Kaizen/CIP is a good basic start but so 90's. It still fits product development techniques like the “Stage-Gate-Model” IMHO. But those old systems are to linear, not adaptive enough to environment shifts and don’t encourage creative innovations the way you need it for game development. Kaizen and co. make boring products with less/no errors left, just look what happend to Sony.

[Eth]

DPS lag was caused by some articial limit to bandwith or packages sent/received (I don't remember the details) and was fixed by removing/changing said limit. It never existed on the EU servers.

Problem is if you break a certain number on said bandwith (at around 35k bps) it does start lagging extremely.
For everyone in the same instance as you.

And that's only for one person. It's not uncommon to have 3k+ on a single character during a fight with several mobs. I don't know how it exactly it works in groups, but if a single character can lag out an instance at 35k+, it wouldn't surprise me if 10 characters at "normal" 3k can lag out an instance just as much. Having no cap on the bandwith doesn't help if the server is only able to handle a certain amount.

If it works that way, having a limit would actually be better, because getting attacks lost isn't as bad as lagging out completely.

[stoerm]

What is lag?
Baby don' hurt me
Don't hurt me
No more

It's great to be able to throw around pet theories and call the other side dishonest, incompetent, whatever while they are bound by rules that restrict them from discussing the details.

While I don't doubt that players are able to observe patterns, you might want to consider the cases where lag *doesn't* happen in similar circumstances. Oh, except you can't because that information isn't available to you.

For starters people should abandon the term "lag" because it can mean just about anything.

If I could make a plea it would be to please stick to describing observed behaviour and the circumstances it happened under. It does not help your case to attribute developer actions to character flaws. As a life lesson I've found it most fruitful to assume the best of people, not the worst. It might work here too.

You may want to act like a crack addict does toward a pusher. Up to you.

[Jiirix]

You may want to act like a crack addict does toward a pusher. Up to you.

Isn't it more acting like a addict because the crack was diluted/watered down?

[bsquishwizzy]

In the interest of not derailing the official thread where Sev made that comment there's also this

Alacrity was removed in that update. It most certainly had a lag reducing effect. In my experience anyway the game went from laggy in DPS dense situations to almost lag free for many years.

Lag is not "fixed" or not, it's not a switch you turn off or on, no more than reducing darkness with candles fixes nighttime. The more you add the lighter it gets. Lag has degrees and it gets worse or better but it's never going to be completely gone. Much like those changes which reduced lag and helped make what was commonly called "shroud lag" and also "DPS lag" at the time become just an old term no one needed to use any more (until about a Year ago when Swash reintroduced it) because it was no longer a thing. Multiple changes are required to improve latency just as multiple changes has increasingly made it worse lately.

For example when they added SWF and Swash complaints of lag increased and people such as myself who RARELY experienced lag beyond the odd rubber band or short glitch, started to see more lag and complain. Then they caused a popular S&B style to have 25% alacrity with lots of average players gobbling up Vanguard in the name of making their formerly gimpy S&B characters much more fun and viable. And lag got a little worse still, more frequent occurrences, much more likely in 12 man instances. Then they speed up crossbow animations which also added more, at this point as someone who had spent 4 years playing DDO and considered the lag normal for most of that time. I grew to the opinion that they had pretty much made it into the laggiest MMO I've ever played (for more than the time it took to uninstall). Warlock seems to have added a few more nails to this coffin.

Sadly

Ummm, what EXACTLY is the point of this?

Aclarity increases attack speed. That means more attacks per minute. This means more information packets going out per second, with means more load on the communications line and server. A packet gets dropped (something not uncommon with UDP across the Internet) and you potentially have a delay as the server tries to resolve a missing shot / swing / whatever.

And “physics checks” on attacks? That happens EVERY attack regardless of its modifier. That’s how they do not allow you to cleave stuff through doors. Duh.

As for what is causing the lag, it could be:
1) Their communications stack.
2) The various linkages between the client and the server,
3) The physics engine,
4) Whatever calculations they use to calculate damage and update stats,
5) A combination of all of the above.

Unless they rolled their own physics and / or communications engines, they usually have ZERO CONTROL over modifying these. Unless they purchased the source (or its community stuff), and that’s often a ****-shoot to resolve.

Oh, and NONE of this really explains how being in a public area – where no combat is going on – you get that stutter lag that sometimes happens.

And unless you know their code, your comments are speculative at best.

So, again, what’s your point?

[Chai]

So by extension increased number of attacks of all sorts that have physics checks would be harder on the server? For example speeding up attacks? Eladrin agreed with you, it's why he felt the need to remove stacking alacrity.

One of the best ways to lower projectile burst is to give ranged players options that boost per projectile damage at the expense of the number of projectiles.

You know the same way you guys could have kept Physics checks on melee attacks from ballooning by giving bonuses that created larger damage numbers instead of more alacrity increases?

At least you guys are catching up to where Eladrin was. When can we stop this charade?

Physics checks on melee attacks still haven't ballooned to pre U5 levels, largely due to offhand being a proc with no physics check, where it had its own physics check in the past.

They could code many shot to fire one shot, one physics check, and deal damage 4 times per hit (just like doublestrike deals damage twice per hit when it procs) for the duration of the burst. I bet the players wouldn't even notice the change.

But instead people ask for changes which will likely be rioted over, because you know, its a change, so lets riot. :p

[dunklezhan]

They could code many shot to fire one shot, one physics check, and deal damage 4 times per hit (just like doublestrike deals damage twice per hit when it procs) for the duration of the burst. I bet the players wouldn't even notice the change.



I like this idea better than all the other ideas they've had. Kill the doubleshot penalty and make many shot just do an appropriate damage multiplier where now it would add an extra arrow (only a smaller multiplier than a straight 'extra arrows' worth - manyshot can miss the odd arrow from time to time and it should whereas your idea will miss an entire batch worth at once). Maybe add half an arrow's worth almost twice as frequently as we'd currently add an extra arrow.


lets riot. :p

Indeed. I call dibs on the storefront windows!

[Mirta]

So by extension increased number of attacks of all sorts that have physics checks would be harder on the server? For example speeding up attacks? Eladrin agreed with you, it's why he felt the need to remove stacking alacrity.

One of the best ways to lower projectile burst is to give ranged players options that boost per projectile damage at the expense of the number of projectiles.

You know the same way you guys could have kept Physics checks on melee attacks from ballooning by giving bonuses that created larger damage numbers instead of more alacrity increases?

At least you guys are catching up to where Eladrin was. When can we stop this charade?

It doesn't seem to me like our damage calculations are causing these issues. We can start getting lag as soon as monsters get aggroed in raids like DoJ so it seems like it is aggro related. Too many monsters calculating aggro too often against too many targets. Since the lag shows up most often in raids this is what my best guess would be.

However, we also get this lag randomly (but much less frequently) in other places. I've seen it everywhere from the bank, marketplace, and Gianthold to individual quests at low level like waterworks. I would guess this is caused by instances being randomly divided between processing servers and if you get thrown on the same server where a raid is causing lag you will get hit, too.

Damage calculations seem very unlikely given all the past changes discussed in this thread. I'd like to see the effects of them changing the aggro mechanic to either make the mobs check less often who they want to attack, or toss all the different factors that add extra calculations to aggro level. For example: don't beeline for the caster just because they're a caster or the healer just because they threw a heal.

[Kompera_Oberon]

It doesn't seem to me like our damage calculations are causing these issues. We can start getting lag as soon as monsters get aggroed in raids like DoJ so it seems like it is aggro related. Too many monsters calculating aggro too often against too many targets. Since the lag shows up most often in raids this is what my best guess would be.
Have you run DotJ recently? I ask because after they added the extra walls and reduced many mob's aggro area the lag has almost entirely vanished. And what lag is occasionally seen is typically a transient patch and not the wipe-inducing, raid group crushing lag seen before they made those changes. It's like night and day.


Defiler of the Just

The monsters that spawn in the trap room are now split into two groups, and a small delay has been added so that they do not spawn at the same time.
Teleportation behavior of devils has been changed so they now have a longer cooldown between attempts, and will not teleport upon spawning.
Made some adjustments to monster spawning in the Fallen Archons/Reinforcements area.
Abishai Executioners will now stick closer to the Fallen Archons.
Pit Fiends and Horned Devils have had their detection range lowered.
Tieflings have had some of their behaviors adjusted.
Some monster pathing has been improved.

[Eth]

Oh, and NONE of this really explains how being in a public area – where no combat is going on – you get that stutter lag that sometimes happens.


Simple. Two examples:
1. Reset your destiny
2. Buy a stack of 100 whatever

This will cause some lag for everyone.

[JOTMON]

Simple. Two examples:
1. Reset your destiny
2. Buy a stack of 100 whatever

This will cause some lag for everyone.

Have never noticed these as causing lag for others in a group.

[JOTMON]

Have you run DotJ recently? I ask because after they added the extra walls and reduced many mob's aggro area the lag has almost entirely vanished. And what lag is occasionally seen is typically a transient patch and not the wipe-inducing, raid group crushing lag seen before they made those changes. It's like night and day.

The combination of changes have improved lag for Normal/Hard Runs..
EE runs are still screwed...
The white-dotting Executioners have just added aggravation.

[Severlin]

So by extension increased number of attacks of all sorts that have physics checks would be harder on the server? For example speeding up attacks? Eladrin agreed with you, it's why he felt the need to remove stacking alacrity.


On two weapon fighting, which is already the most expensive style, yes. On single weapon fighting which throws out many fewer numbers of attacks it's okay to add some since it still doesn't exceed a rate of attack threshold.



One of the best ways to lower projectile burst is to give ranged players options that boost per projectile damage at the expense of the number of projectiles.


Correct.



You know the same way you guys could have kept Physics checks on melee attacks from ballooning by giving bonuses that created larger damage numbers instead of more alacrity increases?


No. Melee attacks don't generate projectiles. In other words, every missile attack not only does a physics check, but it also generates a new entity in the world (the missile) that has a number of collision checks as it flies through the air. The entity creation is also expensive.

Also, doublestrike generates another strike after the physics check is already made. Doubleshot, on the other hand, is implemented in script and actually generates a completely new missile which has its own physics check, and also creates a new entity in the world which fires and has constant collision checks.

Apples and oranges.

Sev~

[Thar]

why is the lag worse now then when server populations were more than double? is it a limited instance/population issue? even if everyone had double the attacks, it shouldn't be causing more server stress than in years past?

[Severlin]

why is the lag worse now then when server populations were more than double? is it a limited instance/population issue? even if everyone had double the attacks, it shouldn't be causing more server stress than in years past?

We've always had reports of lag in raids. Always.

As to the likely causes of lag right now, we are looking at several things.

~ Excessive projectiles. We've talked about this in other threads.

~ Auras. We've increased the number of auras in game, like Warlock's Enlightened Spirit or Bard's Warchanter. Although we used essentially the same tech as previously, the number of auras might be causing overall issues. We've been looking into ways to create new, more efficient aura tech and doing some redesign on some of the auras to help this. You won't see changes along these lines until after the holidays.

~ Lag spikes. There is still some server wide issue that can lag out individual servers with long lag spikes we are tracking down.

~ We don't want to start ripping other stuff apart right now because we'll be moving to a new datacenter at some point and the increased bandwidth and increased CPU power will change our metrics.

Sev~

[Grailhawk]

No. Melee attacks don't generate projectiles. In other words, every missile attack not only does a physics check, but it also generates a new entity in the world (the missile) that has a number of collision checks as it flies through the air. The entity creation is also expensive.

Also, doublestrike generates another strike after the physics check is already made. Doubleshot, on the other hand, is implemented in script and actually generates a completely new missile which has its own physics check, and also creates a new entity in the world which fires and has constant collision checks.

Apples and oranges.

Sev~

So your not afraid of big work the work you did on xbows for the rogue pass shows that, why not change doubleshot to work like doublestrike then where it skips the physics and entity creation parts.

[IronClan]

On two weapon fighting, which is already the most expensive style, yes. On single weapon fighting which throws out many fewer numbers of attacks it's okay to add some since it still doesn't exceed a rate of attack threshold.Sev~

Off hand attacks in TWF are riders with no physics check. As far as I can tell that should make TWF main hand attack close to if not exactly the same speed as someone swinging 1 weapon in terms of physics checks. While the animation is different and it's hard to test; they appear to be similar when you discount off hand (again no physics checks) attacks.

If we are talking about physics checks SFW has +30% alacrity thus should be ~30% more physics checks than TWF... after compensating for the untestable (for a player) differences in animation speed.

People always complain about lag, but never as much as the last year. People like myself who have been playing Multiplayer internet games for 20 years suddenly went from "there's no unusual lag in DDO" (you can find quotes of me saying this) to feeling that this is a very laggy game based on how many more times ENITRE PARTIES complain in voice chat about it. I used to defend your game, Sev, now I'd have to call it perhaps the most laggy networked game I've ever played and not based off my lag, but how many more raids and quests end up with 6 people saying: "anyone lagged", "yep frozen solid", "me too" etc.

Sorry man but this is important to me and too important not to clarify, As Robert A. Heinlein said I'd rather be wrong than uncertain.

So you're saying TWF even without off hand physics checks causes physics checks more often than SF with +30% stacking alacrity?

And you're also saying that SFW with +30% alacrity did not increase the overhead during peak DPS situations which are 90% of where the lag is encountered in my experience? And neither did Vanguard alacrity (25% more physics checks than old S&B players contributed) and neither has speeding up crossbows a lot, and neither has Warlocks chain blast spam bouncing around entire rooms?

All of this is just pure coincidence that in a year the lag has ramped up along with these things?
Why did we have less lag a year ago with more players than we do now with less?

[Chai]

On two weapon fighting, which is already the most expensive style, yes. On single weapon fighting which throws out many fewer numbers of attacks it's okay to add some since it still doesn't exceed a rate of attack threshold.



Correct.



No. Melee attacks don't generate projectiles. In other words, every missile attack not only does a physics check, but it also generates a new entity in the world (the missile) that has a number of collision checks as it flies through the air. The entity creation is also expensive.

Also, doublestrike generates another strike after the physics check is already made. Doubleshot, on the other hand, is implemented in script and actually generates a completely new missile which has its own physics check, and also creates a new entity in the world which fires and has constant collision checks.

Apples and oranges.

Sev~

This further emphasizes how making many shot fire only one projectile, but report 4 hits per shot fired for the duration would be "less expensive" on processing, but likely be the least amount of change on they user experience. I guess the question now is how resource intensive is it to make that change. What kind of people hours are needed?

[Chai]

Off hand attacks in TWF are riders with no physics check. As far as I can tell that should make TWF main hand attack close to if not exactly the same speed as someone swinging 1 weapon in terms of physics checks. While the animation is different and it's hard to test; they appear to be similar when you discount off hand (again no physics checks) attacks.

If we are talking about physics checks SFW has +30% alacrity thus should be ~30% more physics checks than TWF... after compensating for the untestable (for a player) differences in animation speed.

Nowdays yes. We still haven't come anywhere near reaching pre U5 physics checks per time unit, because at that time offhand was its own physics check, as well as being 100% for all TWF.


People always complain about lag, but never as much as the last year. People like myself who have been playing Multiplayer internet games for 20 years suddenly went from "there's no unusual lag in DDO" (you can find quotes of me saying this) to feeling that this is a very laggy game based on how many more times ENITRE PARTIES complain in voice chat about it. I used to defend your game, Sev, now I'd have to call it perhaps the most laggy networked game I've ever played and not based off my lag, but how many more raids and quests end up with 6 people saying: "anyone lagged", "yep frozen solid", "me too" etc.

The lag freezes which occur happen many times when no one is even fighting in the same instance the lag is occurring in. Im still wondering if it is possible that increased physics checks in one instance can cause performance issues in another instance, when people are located in the same server instancing. If our EE raid with a bunch of melee and/or many shotters is in instance i2049 for instance, and someone elses quest is using i2049 as well, during the massive DPS phase of a raid, are other people questing in i2049 experiencing issues with lag?


Sorry man but this is important to me and too important not to clarify, As Robert A. Heinlein said I'd rather be wrong than uncertain.

So you're saying TWF even without off hand physics checks causes physics checks more often than SF with +30% stacking alacrity?

And you're also saying that SFW with +30% alacrity did not increase the overhead during peak DPS situations which are 90% of where the lag is encountered in my experience? And neither did Vanguard alacrity (25% more physics checks than old S&B players contributed) and neither has speeding up crossbows a lot, and neither has Warlocks chain blast spam bouncing around entire rooms?

All of this is just pure coincidence that in a year the lag has ramped up along with these things?
Why did we have less lag a year ago with more players than we do now with less?

confirmation on this would be good I agree.

[Seikojin]

Off hand attacks in TWF are riders with no physics check. As far as I can tell that should make TWF main hand attack close to if not exactly the same speed as someone swinging 1 weapon in terms of physics checks. While the animation is different and it's hard to test; they appear to be similar when you discount off hand (again no physics checks) attacks.

If we are talking about physics checks SFW has +30% alacrity thus should be ~30% more physics checks than TWF... after compensating for the untestable (for a player) differences in animation speed.

People always complain about lag, but never as much as the last year. People like myself who have been playing Multiplayer internet games for 20 years suddenly went from "there's no unusual lag in DDO" (you can find quotes of me saying this) to feeling that this is a very laggy game based on how many more times ENITRE PARTIES complain in voice chat about it. I used to defend your game, Sev, now I'd have to call it perhaps the most laggy networked game I've ever played and not based off my lag, but how many more raids and quests end up with 6 people saying: "anyone lagged", "yep frozen solid", "me too" etc.

Sorry man but this is important to me and too important not to clarify, As Robert A. Heinlein said I'd rather be wrong than uncertain.

So you're saying TWF even without off hand physics checks causes physics checks more often than SF with +30% stacking alacrity?

And you're also saying that SFW with +30% alacrity did not increase the overhead during peak DPS situations which are 90% of where the lag is encountered in my experience? And neither did Vanguard alacrity (25% more physics checks than old S&B players contributed) and neither has speeding up crossbows a lot, and neither has Warlocks chain blast spam bouncing around entire rooms?

All of this is just pure coincidence that in a year the lag has ramped up along with these things?
Why did we have less lag a year ago with more players than we do now with less?

I will just bring up the last part, since that is a huge portion of the lag beast from combat:
The more dice rolling in a single 'attack' the more work the server and client have to do. Period. The more work, more data in that attack, the larger that piece of data being sent is, the longer it takes for the server to crunch, make a feedback, and again, large data coming back... These all lead to lag.

One of the biggest threats to lag in the day, was frame size. Frames are a term for transport packages over networks. And the frames used to have a fixed size. So they couldn't be more than X amount of bits. And if a single piece of data was too big for a frame, a table had to be sent at the beginning describing the data size in frames, so no loss of data occurred. And on the game I worked on, that single issue cause 95% of the gameplay problems in our product that could not be fixed easily.

Back when TWF was being 'nerfed', I brought up condensing the packet size for combat. Doing something to make it smaller so TWF wouldn't need a change. Return on Investment is a big decision maker in living products like MMO's. And it takes 10 times longer than it originally took to fix or work on something that already exists. So if it took 3 months for combat traffic to be made, tested, and refined before release, it would take 30 months to work on it now. So I can easily understand the resistance to my request to fix the combat packets in ddo. My only reminder is these are the things that ultimately make or break the longevity of the product. I want ddo around forever. So for me the return is meh, the investment is critical do.

[Snarglefrump]

Also, doublestrike generates another strike after the physics check is already made. Doubleshot, on the other hand, is implemented in script and actually generates a completely new missile which has its own physics check, and also creates a new entity in the world which fires and has constant collision checks.

Could you reimplement doubleshot to work like doublestrike, then? One missile that hits twice.

[guzzlr]

Are certain spells considered projectiles in a similar fashion to arrows?