Discussion on Damage Bonus, Rules, and the direction of the Emulator

[ChaosSlayer]

regarding the implementation.

I havent looked ta actual code but in order to minimize extra calculation in mind it would look like this:

You store ALL values in a table which is accessed by parameters of Level and Dly

When weapon strike you simly add X(a,b) to the final damage, where X(a,b) points the table with pre-calculated values

As far no DMGB Rule goes - you simply fill the table with ZEROES and every time X is retrived its retrives a 0 =)


NOW to make this even MORE custom, we should make yet ANOTHEr small Rule which would describe a FORMULA for linear DMGb growth, and then when server loads the table will be filled with values based on the formula specified by server admin in Rules

hows that? =)

[renoofturks1]

Another RULE? I see no need for multiple rules.

[ChaosSlayer]

Originally Posted by renoofturks1

Another RULE? I see no need for multiple rules.

1 rule gives LIVE-like effect
1 rule turn effect off
1 rule- alows you create per server custom effect on the fly

all this can be incorporated into a single rule of course which you simply set to 0,1 or 2, or whatever =)

[trevius]

There really isn't much of a need for a rule for adjusting damage bonuses. I think a much better solution for custom servers would be a rule for adjusting the main damage multipliers. So you would increase or decrease normal damage instead of messing with damage bonuses. This is something that I think KLS has already considered adding at some point.

Also, along with that rule it would be nice to have an additional one for adjusting how often hits land for PCs and/or NPCs. Both of these would just add an additional multiplier to the equations that calculate them. Then if the rules are set to 100, it would mean no adjustment, 50 would be half and 200 would be double, etc.

Whether those rules get added or not isn't really related to this thread though and should probably be discussed elsewhere IMO. I am pretty sure there is already a thread somewhere talking about them.

Actually, while thinking about these 2 rules, it made me think of a cool new idea for a quest command. Maybe additionally to these rules, a quest command could be made that would adjust the entire zone damage and hit rates on the fly for all NPCs in it. Then, you could make zones that automatically scale in difficultly based on player level. Though, another rule to scale hps would probably help too

[ChaosSlayer]

Originally Posted by trevius

Also, along with that rule it would be nice to have an additional one for adjusting how often hits land for PCs and/or NPCs. Both of these would just add an additional multiplier to the equations that calculate them. Then if the rules are set to 100, it would mean no adjustment, 50 would be half and 200 would be double, etc.

YES!! I been asking for "TO HIT" modifier since last year . Technicly I have been asking for rule option where DEX would determien how often you score a hit AND a modifier to determine the ratio of how much DEX will be affectign that. Via such option such classes who inherently suck at to hit ratio (like paladins and Sk) can at least do something to improve via in game means- such as building up DEX, rather than to pop up to hit bonus for all classes

Along with that I was also asking for a rule and modifier where AGI will affect the dodge

I probably asking too much. but imho the combat engine would have been if it was writen to mimic eq2 model rather than eq1. cause in eq2 there is never too much stats since your ability to hit and to dodge directly compared to mobs level and mobs ability to hit and to dodge, and X dex and agi which is uber at say 51 will not be as efficient vs say lev 53 mob. This way there is ALWAYS room to improve your stats wihout hiting any sort of cap, since as you level up the worth of your previous stats continusly diminishes vs equal level mobs while making you much tougher vs lower level mobs

[Cantus]

Originally Posted by ChaosSlayer

I probably asking too much. but imho the combat engine would have been if it was writen to mimic eq2 model rather than eq1.

Trevius and Chaos have some awesome ideas.

Maybe I'm just "corrupted" by happily playing for so long on PEQ The Grand Creation (a server that strives to be as legit and Live-like as possible), but I tend to think that the primary goal of the EQEmu engine should be to duplicate Live's way of doing things as much as possible... even if we occasionally think that Live's way of doing things isn't the best way!

But I absolutely agree with ChaosSlayer, and folks like him, that we shouldn't ignore the desires of folks who would have a blast playing on highly customized servers!


I know I’ve only recently started contributing and participating on these forums, so it may not be my place to voice my opinion. But I can’t help thinking that we shouldn't clog up the current combat functions with a lot of rule checks. As Derision pointed out, there's little overhead in each check... but adding a buttload of these checks, and the conditionals they're embedded in, may actually start making a difference in server speed (considering that these functions are invoked each time a melee hit is made... which is frequent!)

Also, it'd make the combat code fairly difficult to understand and maintain.


Here's a possibility: Make a copy of the combat code (for example, copy Attack.cpp and make CustomAttack.cpp). Add "Custom_" before the function names in these new files. And then have a single rule check where either the "Legit" or the "Custom" version of the attack functions are called for each hit.

For speed, we'd create a single static variable that’s a pointer to a function. We’d set it to point to either the “Legit” or the “Custom” attack functions based on a master rule read either at zone load-time, or when rules are manually re-loaded. Then when a hit occurs, we simply use the () operator on this variable to invoke the selected version!

If it points to the "Custom" functions, then a ton of other rules are queried to allow the custom server administrator quite a bit of control over how combat is carried out. If none of these rules are set, "Custom" will operate just like "Legit" (just a bit slower due to all of the rule checks sprinkled throughout it).

On the plus side, it allows the “Legit” code that will be used in most cases to remain lean and mean, so that it runs fast and is easily comprehended and maintained.

But it also gives custom versions of these functions that are extremely customizable by non-coders.

The downside is that any changes to attack handling in the “Legit” versions probably would have to be mirrored over to the “Custom” versions of the functions.

I’d suggest that the highest priority right now probably should be to make sure that the current functions work as well as possible, and are as close to Live as we can make them. Once that’s accomplished, we have the perfect opportunity to duplicate them and add all kinds of customizability!

[ChaosSlayer]

Originally Posted by Cantus

Trevius and Chaos have some awesome ideas.

Maybe I'm just "corrupted" by happily playing for so long on PEQ The Grand Creation (a server that strives to be as legit and Live-like as possible), but I tend to think that the primary goal of the EQEmu engine should be to duplicate Live's way of doing things as much as possible... even if we occasionally think that Live's way of doing things isn't the best way!

well thats where I whole heartly disagree.
I don't see any point to blindly follow SOE footsteps. Over years SOE has altered so many things in the game - how can anything posibly qualify as "cannon" ? And SOE WILL keep changign things, and what you today belive is "live-like" tomorrow will not exist.

Why you insist not only mimicing EQ1 features, but its FLAWS?
And whats even worse- hard coding these flaws into server code. So it would be harder to undo them when one day SOE turns around and implement something what today Cavedude likes so much to call "not LIVE-like" - tomorrow will be "LIVE_like" and you guys will then need to recoded the whole server cuase SOE had a change of heart?

Don't find this.. umm ridiculus at best?

You code a server in such way so it can be tunned to support ANYTHING.

Not to mention a EVRY important part- why do all these people are here at Emu? Do all of them so poor they can't afford $15 a month for stable top of the line tech OFFICIAL server?
I very much doubt that- they here cuase they TIRED of eq1 "full time job" apparoach - they don't realy want LIVE-like things- they simply want to see what they never would on LIVE cuase of difculty of the game and move on.

Not even PEQ is LIVE-like, starting from the fact that it shoudl be imposible to solo anything, up to the poitn there is no freaken way on LIVE you can kill NTOV dragons in Velious gear with 20 people. Yeah I know the emu simply cannto suport 70+ peopel raids - the zone will crash - so raids were tuned down.. But wait- thats not live-like! So we had no problem ALTERING the Grand Vision when suited us. Even if emu COULD support 100+ people in a zone and run flawlessly - you REALY think there would have been 70+ people raids on PEQ when server population around 150 men total and MOST of then are casual solo players? Whats LIVE-like about that? Absolutly nothing

You also need to realise that MAJORITY of players DO NOT play on PEQ - they sick and tired of all those "cannon" things back when then played on LIVE. They play on CUSTOM servers, and its the custom servers that last the longest (exept of course PEQ which is the base server for the project).
Take SoD/Winters Roar - they count what - 600+ players? Yes they doign it illigaly, but heck its an EQ emulator nevertheless with CUSTOM content- and they have 3 times more people than "wanna-be-cannon" PEQ alone.

This should tell you what people REALY expect from an Emulator.
Why emu software is even open for download? Why not just stop giving it out, and only have PEQ which is dumped down eq-live.

NO, the IDEA was that EVERYONE can run their CUSTOM server and change thigns the way they want.
the idea NOT to mimic live and all ist flaws- the idea to make YOUR OWN EverQuest - and to do that all the "live-like cannon" attitude needs to be thrown out the window - cuase it will KILL the project.

To me Emulator is a bunch of assembly blocks - what I make out of it does not have to be in anyway resemble eq-LIVE. So it does to MOST players and private server devs here. They simply not interested in "excatly how it was on live" specialy it was broken and flawed like hell.
Some people simply do not understand or have own creative potential to develope game content - thats why they keep copying what SOE did.
Don't be like LIVE - be BETTER.

And thats why you WANT a HIGHLY customizeable server structure, then you can set few variable and turn everquest into doom or anything else if you realy want to.

Whiel I admire Cavedude dedication, I cannot agree with his "vision" of copying down SOE "vision", and every single "custom" sugestions is treated like unholy blachemy (ok true, there are CRAP LOAD OF custom features allready that were coded by the devs which allow GREAT customization, and I am TRULY thankfull for it, but the iron grip attitude realy needs to be reviewed)

The goal should be cretaing server engine so FLEXIABLE that it can be fien tune to do ANYTHING on the fly (which includes btw mimicing EQ-LIVE)
Creating such ultra-flexiable engine as much benefit to hard core fans as to custom content people, cuase again- you gives GREAT deal of control over the server, cuase as I allreday said before- tomorrow SOE gona turn aroudn and chance soemthign in the game which for years was "will never happen" (yeah like I never heard that song before sicne 1999) , and guys will have to - re-hard-code the server again, and again and again.
WHy would you want to do that?

[Cantus]

Originally Posted by ChaosSlayer

I havent looked ta actual code but in order to minimize extra calculation in mind it would look like this:

You store ALL values in a table which is accessed by parameters of Level and Dly

When weapon strike you simly add X(a,b) to the final damage, where X(a,b) points the table with pre-calculated values

Hi, Chaos!

What you're describing is a brute-force lookup table... just one that's built on-the-fly (which would, in all likelihood, be slower than a comprehensive, pre-built static lookup table.)

There are three options for calculating damage bonuses:

• A brute-force table like what you describe
• A pure-formula approach where damage bonuses are calculated mathematically
• A hybrid approach where formulas are used in cases where they're simple and fast, and tables are used for all other situations

I tried all three possibilities, and the third option -- the hybrid approach -- is what provided the best performance.

That's what's implemented here in this thread. As I mentioned above, testing (that you can easily re-create) proves that this function is a very efficient and fast way of calculating these values, and that it will have an absolutely unnoticeable impact on server performance.

Please don't think I'm shooting down your idea! It's a great one! I'm just pointing out that it was considered, tested, and found to be not quite as efficient as the solution presented in this thread.

[ChaosSlayer]

Originally Posted by Cantus

What you're describing is a brute-force lookup table... just one that's built on-the-fly (which would, in all likelihood, be slower than a comprehensive, pre-built static lookup table.)
.

Hiya!

I am not sure why you saying that this method is build on the fly
You create the table and populate it with values when server loads- and it NEVER changes - the values are very static.

You may very well put them directly into code. You simply pre-calc the values ONCE and then just add them to final dmg when/if needed

I am not arguing btw - I just not seeing the "complexity" in creating the table at server start up and keeping it static (just like you want it) until server shuts down

The bonus part of course is that BEFORE server loads you can populate the table with values based on ANY formula- again only once - and they stay static for whole duration

[Cantus]

Originally Posted by ChaosSlayer

I am not arguing btw - I just not seeing the "complexity" in creating the table at server start up and keeping it static

If you take a look at the code that's posted, you'll see many of these pre-calculated tables already included.

There'd be no reason to build the table at run-time, even if just when the server starts. Better to populate them at compile-time.

You could do a full, brute-force lookup table, but my point is that testing proves that the hybrid approach listed here is the most efficient implementation.

It works, it's fast, and it'll soon have (as requested) the ability to return damage bonuses for toons level 81+, and also an option for a custom, linear damage bonus progression.