These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Player Features and Ideas Discussion

 
  • Topic is locked indefinitely.
 

Server balancing through gameplay

Author
Kirluin
#1 - 2014-01-22 14:19:28 UTC  |  Edited by: Kirluin
Putting an idea that came out of a different thread into it's own:

Problem:
... Multithreading the Eve solar system code will not really solve the problem. Every time they've increased node capacity, players have piled on load until it breaks again. Eve has gone from 20v20 killing a server to 4,000 man blobs killing a server. If they increase capacity another 10x/node, players will increase load 12x and complain. Rinse and repeat.

Solution:
Unless Eve devs get something that can handle 30,000-60,000 users on one node, they are better off changing the problem and providing gameplay incentives to NOT pile blobs into a system. Something long the lines of "if we control adjacent systems, both fleets gain a better bonus than if they are all in one system." A player generated incursion style mechanic is one way to do this.

Incentive for top brass: So an alliance commander with 2,000 people at his disposal would WANT to split his forces on adjacent systems to give them all a boost. Defenders would WANT to attack each of these adjacent systems to get a similar boost or deny the enemy the boost. Smaller side skirmishes can contribute to an assault . This way you induce invasion waves spread out on a battlefront instead of every man invading Normandy on the same 10 foot wide section of beach. Players can still blob if they want to though.

Incentive for players themselves: The other reason people pile into a system is killmails. Ideally if you get on a direct killmail attacking/defending a side skirmish (i.e. you are somehow contributing to the incursion effect), you share in all the fleet killmails for the assault.

Otherwise everyone will pile on whatever system will get them the capital ship killmail anyway.

By all means devs should continue to tweak server performance, but by adding good player driven reasons to broaden the scope of conflict we can not only help the problem, but also create interesting new strategic gameplay. Alliance commanders would have new options to plan an assault more carefully and use the star map more like a chess board.

Many thanks to Corraidhin Farsaidh for making this idea better.
Icarus Able
Refuse.Resist
#2 - 2014-01-22 14:28:18 UTC
Or we do a system where sov is linked to a group of systems and they all have to be attacked/defended simultaneously
Tchulen
Trumpets and Bookmarks
#3 - 2014-01-22 14:59:29 UTC
Actually, distributed local processing would probably solve the problem if they could recode the engine to deal with it. The technology is heading towards increased number of cores rather than increase in individual core speed. The current issue is that the technology isn't increasing with player demand. If they could use multiple cores per node the technology would be improving with the player demands a lot better than it is now.

For example, for anyone who've heard of Cray, currently a Cray supercomputer wouldn't help CCP because like all supercomputers it's massive processing isn't supplied by a single super-powerful processor core but by hundreds (to thousands, I believe) of cores.

There are already systems out there with plenty of processing power. Until CCP recode the engine to be able to utilise multiple cores there really isn't a lot they can do about this. They're unlikely to try to reduce the amount of people who can fight in an individual battle considering that's one of EVE's selling points.

The other potential is to build (or have built more likely) a virtualisation platform that sits between the hardware and the OS which aggregates all the processor cores on the physical piece of hardware and passes that to the OS as a single core. This is probably significantly more difficult that just recoding the EVE engine, however.
Kirluin
#4 - 2014-01-22 18:12:45 UTC
Tchulen wrote:
Until CCP recode the engine to be able to utilise multiple cores there really isn't a lot they can do about this. They're unlikely to try to reduce the amount of people who can fight in an individual battle considering that's one of EVE's selling points.


Multithread rewrite would be a big win for sure. However that is a massive undertaking.. if i recall correctly the server code itself is written in a cooperative multitasking paradigm. So you'd not only have to thread enable the stackless python implementation, but you'd have to pretty much rewrite every piece of code at the same time for preemptive multitasking. Since coop multitask doesn't have to deal with locks or other resource contention, you'd have to add all that code and test it from scratch. That's just to multithread it... if you change to a distributed processing model spanning servers then you have to refactor for higher latency, distributed scheduling, etc which is another order of magnitude of difficulty.

Lastly you'd have to retrain all your developers on the new model, and implement LOTS of new testing to handle concurrency. Overall you're looking at drastically increasing your time to implement for new features until everyone got up to speed.

All that is kind of like converting your single engine fighter plane to a multi engine jumbo jet. And retaining the performance of the fighter jet. switching the fuel type to diesel.. without landing. The results would be sweet though.

However at the end of the day you would still run into some maximum single system limit that players would try to exceed, so something that moves gameplay away from blobness is probably inevitable. Easier on client side performance/user interface too.

You could also link this system to cap/supercap ship deployments: to gain the best benefit you may want to deploy caps in something other than a one system blob. Other strategies then become viable, like holding a blops gang in reserve to jump across the battlefront as needed for best special teams impact, etc. better performance AND interesting game time decisions.
Centis Adjani
Adjani Corporation
#5 - 2014-01-22 18:39:04 UTC  |  Edited by: Centis Adjani
And to force players to use this play style Kirluin suggest, CCP could implement following additional game mechanic.

If a system reach a critical mass of ship masses and TIDI goes below 20% (or reach 10%),
a timer for a Black Hole in the system starts.

Players now have the choice:
- stay in the System - and when timer is over and Black Hole is born, all ships will become pulled into and destroyed.
- leave the System immediate to reduce the critical mass and stop the timer

Instead of a Black Hole another solution may be a Wormhole with extreme gravity.
Which pulls the ships inside and they reappear in a random system somewhere in the Galaxy (Null Sec only or another WH).
Dolorous Tremmens
The Scope
Gallente Federation
#6 - 2014-01-22 18:56:26 UTC  |  Edited by: Dolorous Tremmens
Have a maximum tonnage for a fleet, but still keep the same max number of fleet members. Reducing the number of large ships jumping at one time to the same cyno might help curb losses, and give the server enough time to throw up warnings, A fleet of fleets would be less coordinated, and then theres FC quality and quantity.

The number of ships in a system can't be capped, that would be gamed, but the location of the jump in points could be changed. Cynos would not be able to be lit on gates, planets or moons, forcing off grid jumping. Cyno pos mods are naturally exempt.

I liked the idea of getting bonuses from adjacent systems, but it it could be modular, and cap ship based. you would need cap ships in neighboring systems with new capital leadership modules, which would work on a ratio basis. To get bonuses you need one cap ship per wing. you'd need fewer capships to boost a cap fleet in the other system, but more for the subcaps.

Leadership modules could only be used in triage mode, which I believe means outside pos shields. The bonus givers also act as a bit of a reserve, should things start to go badly next door.

There should be a focus on more organization and a reason to diffuse the blob into surrounding systems, even if just a bit.

in fact there should be another level of organization, since we have Capital ships: flotillas, or a fleet of fleets. oh, its not for mass movement, strictly for organization and another level of leadership bonuses, previously mentioned. a reserve fleet pumping out bonuses for other fleets in neighboring systems ( limited by constellation)

Get some Eve. Make it yours.

Zan Shiro
Doomheim
#7 - 2014-01-23 02:50:04 UTC
Kirluin wrote:
By all means devs should continue to tweak server performance, but by adding good player driven reasons to broaden the scope of conflict we can not only help the problem, but also create interesting new strategic gameplay. Alliance commanders would have new options to plan an assault more carefully and use the star map more like a chess board.



Thing is this can be done now, but its not.

What I saw way back in IT/NC (no dot) wartime.

NC hired PL to hit delve. this caused IT to split forces and do jc ops to protect resources there. Less people on front so push ops put on hold. Basically its on the crews to either hire or send own peeps to force the jc ops.


Or the HW- fuster cluck. Whether IT spin doctoring or planned while most of all sides cuaght up in the week of node issue crap there was another IT fleet out there crashing sov in other parts of nc space. IIRC sov was removed in a few systems to shut down a decent section of NC's jb network. Granted this did not work out well (IT withdrew eventually) but premise showed some promise.


If crews are happy to slam heads in the same system that really is a leadership issue really, not mechanics.


Also worth noting this bonus idea probably won't go very far. CCP has a long memory. Some WH empires have used wh system bonus to great advantage in the past. I forget the wh crew and what wh effect they used but it made them damn near impossible to remove even if you really wanted too. Both sides in theory have access to the same wh effect. As defenders however they had the upper hand in placement for max effect right off the bat.

I want to say something about uber range or power on sr guns like blasters that made storming that wh entrance suicidal for all intents and purposes. CCP rebalanced this effect in time. Don't think they got busted for exploit but that may be due to the fact imo ccp only pulls the exploit card when they can't fix the issue with code imo (i.e. the recent new version of pos bowling ccp advised on in recent months.....they jsut can't kill pos bowling so exploit it becomes whenever it shows its face again lol).
Sigras
Conglomo
#8 - 2014-01-23 03:00:47 UTC
It was Apature Harmonics in a WH system called Nova. It contained a level 6 magnatar effect which made all e-war far more effective

This allowed A-Harm to tracking disrupt their own ships allowing for a negative tracking number. Because of the nature of the tracking formula, a negative tracking number means that 0.5 gets raised to a negative number which always results in a number greater than 1 which means the chance to his < 100% at any range.

Essentially A-Harm exploited the e-war effect and broke the freaking game. Youtube search "Clarion Call 3" for more information.
Zan Shiro
Doomheim
#9 - 2014-01-23 03:07:57 UTC
thank you...memory gets hazy as the years pile up lol
Kirluin
#10 - 2014-01-23 17:04:50 UTC
Good to know some relevant history, thanks! I'm sure CCP can iterate to something workable as they did in that case, possibly by building off the well established incursion system.

Lin Fatale
Garoun Investment Bank
Gallente Federation
#11 - 2014-01-23 17:44:39 UTC
there are far more easier approaches to deblob
why do people blob? because its easy to do so, they have the tools to control large amount of ppl with relatively low effort
take away that comfort und there are more chances that the fight will be more dynamic

e.g.
- reduce fleet size to 50
results in more FCs needed, movement is bit harder, more communication needed, mistakes will be made, chances that fights will be stretched through more systems or at least on several grids in one system

some kind of these approaches will also balance the low vs high skilled players
if fleetzise is only 50, suddenly you need 5 times more FCs
a controled way how you comunicate etc.

a well-practised team will be good, some random team will be bad
and not who controls the most sentrys one clickfest
Kagura Nikon
Native Freshfood
Minmatar Republic
#12 - 2014-01-23 18:02:26 UTC
Lin Fatale wrote:
there are far more easier approaches to deblob
why do people blob? because its easy to do so, they have the tools to control large amount of ppl with relatively low effort
take away that comfort und there are more chances that the fight will be more dynamic

e.g.
- reduce fleet size to 50
results in more FCs needed, movement is bit harder, more communication needed, mistakes will be made, chances that fights will be stretched through more systems or at least on several grids in one system

some kind of these approaches will also balance the low vs high skilled players
if fleetzise is only 50, suddenly you need 5 times more FCs
a controled way how you comunicate etc.

a well-practised team will be good, some random team will be bad
and not who controls the most sentrys one clickfest



irrelevant effect. Anyone that has been on large battles know the FC is not even necessarily in the main fleet spot. The fleet commander is the guy that speaks on TS, the ogthers are jsut ALTS giving fleet bonuses.

"If brute force does not solve your problem....  then you are  surely not using enough!"

Agonising Ecstacy
Chaos Army
#13 - 2014-01-23 18:04:32 UTC
No - people will just use external tools instead. You cant 'limit what people can do' by capping/removing tools - instead you MUST incentivise people to change behavior.
Corraidhin Farsaidh
Federal Navy Academy
Gallente Federation
#14 - 2014-01-23 18:19:02 UTC  |  Edited by: Corraidhin Farsaidh
naturally I agree with the concept as I think I suggested it in the other thread :)

Thee has to be a real incentive to fight in smaller fleet actions to lure people away from the blob. I suggested using the incursion mechanic as it is well tested and already exists. the only code needed additionally would be to register that a force has control of gates, and then trigger the sov incursion effect when all gates into a given system are controlled by one aggressor fleet.

My suggestion was also that the controlling fleet must control the gate on the adjacent system side, thus spreading any gate control fleet combat into at least 1 other system, probably two or more.unless you lay a really good ambush and the enemy jump into a dead end system.

This would hopefully be relatively simple to code and test as the Incursion code already exists and the Sov control code could possibly form a basic framework for gate control. This could be achieved by placing a Sov structure that hack the gate which the aggressor would have to defend from the trapped fleet to retain the Sov Incursion effect
Lin Fatale
Garoun Investment Bank
Gallente Federation
#15 - 2014-01-24 01:51:28 UTC  |  Edited by: Lin Fatale
Kagura Nikon wrote:
Lin Fatale wrote:
there are far more easier approaches to deblob
why do people blob? because its easy to do so, they have the tools to control large amount of ppl with relatively low effort
take away that comfort und there are more chances that the fight will be more dynamic

e.g.
- reduce fleet size to 50
results in more FCs needed, movement is bit harder, more communication needed, mistakes will be made, chances that fights will be stretched through more systems or at least on several grids in one system

some kind of these approaches will also balance the low vs high skilled players
if fleetzise is only 50, suddenly you need 5 times more FCs
a controled way how you comunicate etc.

a well-practised team will be good, some random team will be bad
and not who controls the most sentrys one clickfest



irrelevant effect. Anyone that has been on large battles know the FC is not even necessarily in the main fleet spot. The fleet commander is the guy that speaks on TS, the ogthers are jsut ALTS giving fleet bonuses.


you have to look at it from another perspective

it does not matter if you call the person FC, alt of something or if he is in FC position or just squadmember
fact is someone has to do it, aka give you warpins, call targets, tell you what todo

its harder to get 1000 individuals coordinated if they have to do it on their own
than that one person is doing it for anyone else

1000 individuals will do mistakes
if only one player controls everything there will be no mistakes, no dynamic

its the same with the drone assign issue
with drones assigned to one player, there are no mistakes of fleetmembers because there skill is not relevant
if everyone has to shoot on his own, there is room for mistakes, delays, km whoring, whatever

and the same counts for the number of fleets
there is a diffrence if you only have to control one fleet or if you have to coordinate 50 fleets
It adds room for dynamic, chaos


ofc it will not totaly change everything but its a small step into the direction that blops dont have to be more efficient than smaller fleets

but the goal should be to add dynamic into the game and not stagnation
Scipio Artelius
Weaponised Vegemite
Flying Dangerous
#16 - 2014-01-24 02:01:05 UTC  |  Edited by: Scipio Artelius
Lin Fatale wrote:
it does not matter if you call the person FC, alt of something or if he is in FC position or just squadmember
fact is someone has to do it, aka give you warpins, call targets, tell you what todo

its harder to get 1000 individuals coordinated if they have to do it on their own
than that one person is doing it for anyone else

1000 individuals will do mistakes
if only one player controls everything there will be no mistakes, no dynamic


Sorry, but if you reduce the fleet size from 256 to 50, that will really do nothing to change the dynamic of 2000 pilot fleets vs 2000 pilot fleets (or similar size 700 v 2700, etc.).

That just means having more fleet organisations with people in different positions.

It doesn't change the size of the overall engagement and does nothing to change the effects on the server. Your only talking about putting pilots in different boxes, not reducing the total number of pilots on grid.
Kagura Nikon
Native Freshfood
Minmatar Republic
#17 - 2014-01-24 09:43:57 UTC
Lin Fatale wrote:
Kagura Nikon wrote:
Lin Fatale wrote:
there are far more easier approaches to deblob
why do people blob? because its easy to do so, they have the tools to control large amount of ppl with relatively low effort
take away that comfort und there are more chances that the fight will be more dynamic

e.g.
- reduce fleet size to 50
results in more FCs needed, movement is bit harder, more communication needed, mistakes will be made, chances that fights will be stretched through more systems or at least on several grids in one system

some kind of these approaches will also balance the low vs high skilled players
if fleetzise is only 50, suddenly you need 5 times more FCs
a controled way how you comunicate etc.

a well-practised team will be good, some random team will be bad
and not who controls the most sentrys one clickfest



irrelevant effect. Anyone that has been on large battles know the FC is not even necessarily in the main fleet spot. The fleet commander is the guy that speaks on TS, the ogthers are jsut ALTS giving fleet bonuses.


you have to look at it from another perspective

it does not matter if you call the person FC, alt of something or if he is in FC position or just squadmember
fact is someone has to do it, aka give you warpins, call targets, tell you what todo

its harder to get 1000 individuals coordinated if they have to do it on their own
than that one person is doing it for anyone else

1000 individuals will do mistakes
if only one player controls everything there will be no mistakes, no dynamic

its the same with the drone assign issue
with drones assigned to one player, there are no mistakes of fleetmembers because there skill is not relevant
if everyone has to shoot on his own, there is room for mistakes, delays, km whoring, whatever

and the same counts for the number of fleets
there is a diffrence if you only have to control one fleet or if you have to coordinate 50 fleets
It adds room for dynamic, chaos


ofc it will not totaly change everything but its a small step into the direction that blops dont have to be more efficient than smaller fleets

but the goal should be to add dynamic into the game and not stagnation



Current engagement already have up to 10 fleets on each side. Its ano brainer to double that for any of these organizations.

"If brute force does not solve your problem....  then you are  surely not using enough!"