Pre-Tidi a big fight was 1200 people in local. Pre-tidi that crashed the node.

Current situation is 1500-4000 in local with extreme tidi plus possible node-crash.

It's always been thus. When CCP increase the capacity of a system, the player base crash the capacity with more people.

If in four years time CCP triples or quadruples the capacity, players will pile in five or six times as many.

I find it amusing when people describe fights before TiDi and it sounds exactly like my experience now.

Finally a good post.

The problem is the way the players are playing the game, because that is what the game rewards most.

Pre-Tidi a big fight was 1200 people in local. Pre-tidi that crashed the node.

Current situation is 1500-4000 in local with extreme tidi plus possible node-crash.

It's always been thus. When CCP increase the capacity of a system, the player base crash the capacity with more people.

If in four years time CCP triples or quadruples the capacity, players will pile in five or six times as many.

Problem isn't tidi, Its the drone clouds.

Beehive domi were nerfed for this very reason all those years ago.

Drone assist is broken, but it's not the cause of the problem.

Perhaps but 4000 drones cant help one would think.

tidi havent solved anything.
unless you consider watchign ata stillframe frozen for 4hours better then a blackscreen.

If you actually consider that progress and fun i feel sorry for you.

Right now, I am in a high sec system, at least 4 jumps from low sec, let alone null.
And we have Tidi at 77%, (was 87% a moment ago) with 33 people in local.

Yeah, this TiDi thing is working great.

Right now, I am in a high sec system, at least 4 jumps from low sec, let alone null.
And we have Tidi at 77%, (was 87% a moment ago) with 33 people in local.

Yeah, this TiDi thing is working great.

Root of the problem ....

1) No friendly fire ... guns fire through friendly ships, how nice.
2) No AoE damage from exploding ships.

You design a game for the mindless blob, you naturally get a mindless blob in return.

How do you design an MMO that doesn't facilitate players working together?

I don't think line of sight will ever be a thing because there's already quite a bit of calculation going on that the server can't handle.

I think LOS should be a thing because it adds so much to the tactical possibilities. I'm sure nVidia or someone in the graphics hardware domain has a piece of ray-tracing hardware which would make LOS calculations trivially easy to do in parallel on the same motherboard as the processor managing the combat simulation. It might even be possible to offload hit/miss and effective damage calculations to such a piece of hardware. Then all the simulator would be responsible for is physics and marshalling inputs and outputs. Then some of that could be offloaded, since vector processing belongs on a maths unit GǪ and all that's left is marshalling data and keeping track of time. Of course that's the paper-napkin plan from the person whose crowning achievement to date has been porting apps from badly written PHP to badly written Django

If you ask any hardened veteran who was around a long time ago, they will tell you that big fights were generally Not Fun because of all the disconnects, black-screening, etc. Time Dilation did a lot to fix this and made big fights a lot more bearable.

The question is, even with this why is it that nearly all fights over timers these days cause node crashes, are slowed down to their maximum TiDi percentage and are generally not a good time? Why are we headed back to the old days of lag and unplayability?

Today's meta-game has shifted towards having the biggest bloc, as weaker blocs have died out. The majority of member corps of these blocs did not go and stake their own claim. They didn't reform into a new power bloc. They just simply joined one of the existing larger blocs, because straight up fights are what most people find fun, not backroom deals or dealing with logistics which are huge investments of effort that having your own bloc requires. This results in the three gigantic sov blocs we have now.

The problem with having these gigantic blocs is that suddenly instead of the 500 to 1000 man fleet fights which were huge back in the day and can easily be handled by TiDi, now every single fight over an objective is a 2000 man slugfest.

The secondary problem with this is that this gigantic fight can kick off pretty much anywhere where you can expect any one fleet to pass through. If your fleet takes gates, CCP would have to reinforce all systems en route to make sure one fleet of 250 can pass through without getting jumped and the fight escalating. All midpoints have to be reinforced. All staging systems have to be perma-reinforced. Any small objective turns into a big one, and suddenly every single timer system has to be reinforced, which can be a problem if a large number of timers are generated.

There's no way you would be able to reinforce all of these systems so that they can handle a gigantic fight. The players can't always deliver intel to you on time for downtime, and you shouldn't throw your hands up in the air and blame us for a broken system which promotes a metagame that causes heavy amounts of lag.

This has been a slowly escalating issue since 2011. TiDi was a good bandaid solution for CCP to buy time and allow them to improve upon nullsec and the sov system to make it less bloc-centric before we got here.

But again we see ourselves at the breaking point, where players are feeling the side effects of CCP failing to make any effective systematic change to a real issue, and even with the attempted fixes that have occurred there is no end in sight. Even the arguably greatest addition you have made, the introduction of TiDi, has become ineffective due to a system that is not working properly.

All of the effort put into quality-of-life changes is much appreciated, but at the end of the day these changes don't matter if you're stuck in slow motion for 3 hours longer than you wanted to play for.

So how do we change the mechanics to go against blob warfare?

Even if Ccp would fix it so you could fight a 5000 man fight without lag people would just bring 6000 and still whine about lag. Remove the incentives for blobbing instead. You would get more intressting gameplay for everyone AND no more issues with lag. A win win solution.

I think LOS should be a thing because it adds so much to the tactical possibilities. I'm sure nVidia or someone in the graphics hardware domain has a piece of ray-tracing hardware which would make LOS calculations trivially easy to do in parallel on the same motherboard as the processor managing the combat simulation.

CCP decided that Tidi was "good enough" and started working on the graphics. The problem with that is, shiny pictures require more time to generate.

This isn't going to be not garbage until they figure out how to utilize multiple cores for node processing.

If you ask any hardened veteran who was around a long time ago, they will tell you that big fights were generally Not Fun because of all the disconnects, black-screening, etc. Time Dilation did a lot to fix this and made big fights a lot more bearable.

Adding line of sight mechanics or even aoe would increase the load on the server exponentially.

Create several targets that can be hit and disabled by smaller gangs and that are economically or tactically relevant.

Example: Make jump bridges have 200 k EHP. Put Moon mining structures outside pos shield and have 200 k EHP (before the disablign is reached). Remove POS scramblers and warp disruptors (makign easier to a small gang to hit and run).

That type of thing. Woudl add another field to cripple the enemy, withing timed and fast attacs that woudl demans several smaller gagns patrolling,k not a single huge blob to coutner it.

Some will eventually escalate, but now there is real action with strategic relevance that is not made under TIDI.


Billion EHP targets as the only relevant warfare objectives lead to the need of large fleets and long time operations, creatign the oportunity to be counter blobed, what escalates in the need of every operation to be a Huge blob of as large as possible scale.



Give us hit and run, and that will create less field battles. Players will gravitate to the most efficient way of winning the war. Just make possible for other tactics than massive blob to work, and players will use them.

Now those are the sort of ideas I think CCP should be looking into, make it more about small engagements, say 500 ships a fleet max, strike multiple target systems at the same time allow smaller gang raids to have some effect, put in penalties that apply as the fleets start going over a certain size a leadership negative that shows that over a certain limit C&C becomes a problem and make it a penalty mechanic incentivising not blobbing.

These are some of the things that should be looking into, not just multi-thread the system so we can blob more, yes multi-thread to make it run smoother but create a set of mechanics and tactics that discourage super large fleets.

Root of the problem ....

1) No friendly fire ... guns fire through friendly ships, how nice.
2) No AoE damage from exploding ships.

You design a game for the mindless blob, you naturally get a mindless blob in return.

Let's see, what we got here.

Suggesting that we fix lag by adding in anti-blob mechanics which exponentially increase the required calculations? Check.
Suggesting sharding the server? Check
Complaining that TiDi doesn't work just because it too has a limit on what it can handle? Check.
The forum's resident DBAs and software engineers telling CCP they how to fix it by reversing the polarity? Check

All in all, a great thread.

This is something that'll require a heavy mix of better hardware and much, much better software, and it would be nice for the Dev's to at least hint that it's something that's not only being actively worked on, but that it's a priority.

if there are overall way less people engaged in a fight, slightly adding the processing requirements on a per pilot basis is still an overall net gain.

Now those are the sort of ideas I think CCP should be looking into, make it more about small engagements, say 500 ships a fleet max, strike multiple target systems at the same time allow smaller gang raids to have some effect, put in penalties that apply as the fleets start going over a certain size a leadership negative that shows that over a certain limit C&C becomes a problem and make it a penalty mechanic incentivising not blobbing.

These are some of the things that should be looking into, not just multi-thread the system so we can blob more, yes multi-thread to make it run smoother but create a set of mechanics and tactics that discourage super large fleets.

It's a shame that you spent all that time writing that post when all that's there at the core is the single greatest non-solution to scalability issues.

Clearly mechanics that incentives simple structure shoots and structure reps are the way to go.

Less fighting over the structure and more just shooting or repping it.

I promise you, optimized code and compatible server hardware can run 2000 pilot fights without TiDi. There are software solutions to both optimizing the use of current hardware technologies and for optimizing the way that the EVE client interacts with the server.

I believe this is number 4 in my checklist.

Bonus points for including a version of "640 kB ought to be enough for anybody"

You selectively quote like a pro. I believe this is numb

Years of practice, and you should see a doctor about that numbness.

Here's another tidbit: I'm not the one assuming the population is capped by anything. If it was, incremental improvements to server efficiency might be worth it. However, incremental improvements have led us to a situation now for which I made a sig in 2007. It's not the answer. Lag-free fleet fights are not just over that next rise, it's time to go back down the mountain and re-evaluate.

with TiDi - great idea - but definatley a band aid solution.

Expand on it CCP!

So - System A is going to have a 2000+ fleet battle. CCP gets notified of when and where , that node is re-enforced. When a node is re-enforced it links to a box in the new rack. All combat related CPU calculations for that system are calculated on a box that's purely dedicated to doing the thinking. Have it delay or trickle information suchs as kills, deaths, yadda yadda back to the main server so it doesnt bottleneck there.

I seriously don't understand why everyone hates TiDi so much.

I seriously don't understand why everyone hates TiDi so much. In any other game so much activity from so many players would render the game completely unplayable. We would rather have a slow-mo battle than a full server crash, but people fail to appreciate how useful TiDi is. I cant stress this enough so i'm going to say it again: if large battles like those in EVE happened in any other game the game would crash, instead of run. The fact that we can even have these kinds of battles or events is truly amazing.

TiDi is not a band-aid solution. Its a CCP-doesnt-have-infinite-money-for-better-servers-and-this-never-happens-in-other-games-so-no-one-knows-a-better-solution solution that we take for granted.

OP is just a fresh wording of "buy better servers CCP!" which is the complaint that gets tossed around far too often.

No, you (and to be fair, many other posters in this thread) completely miss the point I am trying to make.

There is a more permanent solution to soul-crushing lag than just throwing processor power and optimization at the problem.

There is a structural issue in the game in that the only way to obtain progress performing specific activities causes a huge processor load in one very small area, resulting in a broken game. TiDi didn't fix the issue, the next incarnation of TiDi won't fix it and purchasing more million dollar supercomputers won't fix it.

CCP had plenty of time to do something about it before the problem crept up again, but instead they chose to do nothing and we are pretty much back to where we started.

I find it amusing when people describe fights before TiDi and it sounds exactly like my experience now.

Removing drone models altogether might help. I don't remember anyone that liked drone models.

I also think there should be more AOE in the game, and along with that, have formations that people warp into on grid.

EDIT: I reckon it would be very well received

What is that solution?

To identify the reasons why people start/get involved in these situations and to distribute the load across multiple nodes.

I swear, if you're not underemployed or a part time student, TiDi makes this game unplayable (on the occasions where you have to deal with it).

But again, a low expectations playerbase doesn't just see TiDi as a necessary evil, but talk about how much happier they are with it than without it.

It's like Internet Stockholm Syndrome - why worry about fixing a serious problem when things could be so much worse? Amirite?

CCP is pretty unique in having a single-shard universe, and it's a major feature of EVE; but there is a pretty good reason why all MMOs don't do single shard . . . can you guess why? If you said unplayable lag, you win! And what you win is unplayable nullsec fights due to lag!

If one of EVE's several defining features (massive-scale battles) doesn't work, it needs to be fixed, not bandaged. The repeated node deaths of the previous few months show that things are getting worse, not better, and asinine solutions like removing drones or forcing players to use less ships aren't going to help. This is something that'll require a heavy mix of better hardware and much, much better software, and it would be nice for the Dev's to at least hint that it's something that's not only being actively worked on, but that it's a priority.

EVE is already one of the least-interactive video games out there, adding 8 hour fights (and 2 hour travel times to get to those fights) is pretty sad. But again . . . benefits of a low expectation player base, I guess :P

Or maybe there is some other fix (other than serious, dedicated software overhaul and the addition of some even higher-end hardware) that hasn't been brought up yet. Maybe . . .

1) Subspace Warp Core Interference: when x amounts of supercaps are on the field, the combined interference from their warp drives causes some sort of pulse that straight up kills all drones in the system (electronics are too fragile for it) but leaves bombers intact (hardened electronics) as well as all ships (again, hardened electronics). It would mean that all fights that escalate to supercaps would mean drone boats get the shaft, but I'm sure the meta could adjust accordingly, and smaller fights would be unaffected.

2) Voltron Drone Assist: when a fleet has deployed x amount of drones and assisted them to one ship, the drones interlock into one giant behemoth drone-cannon. Requires a new type of drones that have better tracking/damage/optimal than current T2 drones, making it ideal for fleet fights where certain engagement ranges can (hopefully) be predicted. Instead of 5,000+ drones on the field, you have two giant drones facing off against one another, or one giant drone just murderfacing the opposing fleet unless and until it gets blown apart by opposing battleships/cruisers. Added points if EWAR drones can pull the same trick, and a monster Hornet-god can AoE ECM whole fleets with tear-inducing goodness (might be overkill on that one).

Problem isn't tidi, Its the drone clouds.

Beehive domi were nerfed for this very reason all those years ago.

I agree. Only drone boats should be able to field 3-5 drones, everyone else should be limited to 1-2 drones, with drones buffed to compensate. Would also be nice to add an incentive to using larger drones on a ship, ie. if a battleship could field either 2 medium or 2 small, it would choose the medium for more DPS or the small for better tracking. When the alternate to 2 medium is 4 small, the smalls give better tracking and dps at the same time, while also making one ship count as 5 units as far as the server is concerned.

Solution: drone hardpoints.

But can you see any simple solution except - don't swarm ?

From what i observed tidi spikes when people begin to shoot - especially when carrier drones shoot ( calculation for each drone probably, delegation, and the fact that they shoot at the same time ).
Maybe CCP go into this direction - or provide us with some simplified combat scheme during big fights - simplified scripts for shoots calculation , that will round some of the values, or use previous results.

I know this is bad solution, but it is way better than "we are going to kill ihub - job for 10minutes , see you in 5 hours.

Actually the best solution is to change the Sov mechanic from structure oriented to control oriented, but if CCP were to try to do so the same people would be here crying about CCP making the game unplayable(unplayable being a synonym for effort, strategy and thinking in this case). There is no solution for these folks, they'll complain no matter what happens.

drones need to be changed to a gun slot on your ship, and the drones replaced with some kind ammo that gives the same bonuses and it needs to auto assist to targets you are firing on. When the server is trying to deal with 2000 people it should not be doing calculations for x5 each person's worth of drones.

Its a terrible example but I think you get the idea

AoE DOOMSDAYS

Make some new ship to use these. They will be very expensive, an alliance might only have two or even three of them.

how big were fights before Tidi?

How big can they get now?


Unfortunately we need about 200 atoms and a Quantum computer to run a proper fight smoothly.

Are you implying drone assist is the lead cause of heavy TiDi?

Remember back in 2010 when ccp grayscale said there were looking into fix null sec...

Yeah so its almost 4 years later... any updates?

Problem with your idea, Frostys: people will abuse the player limit by bringing in a force large enough to reach the limit all at once, blocking out a defense force.

I suggested long time ago that you have to wait the difference between TIDI/Normal Time - to join the fight.
This is 1 timer and it is easily for ccp to manage it without any big load.

For example TIDI is jumping 1-20% and difference between TIDI system rest of eve is 20 minutes.
You arrive at the gate to this system/ use JB / Titan etc and you get nice prompt.

"You will have to wait 20min to join this fight."
When you click Yes - you get perma cloack and have to wait 20min , or you click no - and go back.

When at some point of battle tidi is gone - timer gets a reset, and all waiting people jump in in perma cloack until they move.

Edit:
One of other topic how to limit TIDI.
People from bloob fights know how things go messy, when sentry drones begin to shoot.

https://forums.eveonline.com/default.aspx?g=posts&m=3921265#post3921265

The reason why this is a problem is because 90% of the people are spineless clowns who would rather hide in numbers than actually have some backbone and self respect, they're grind bears. Don't blame EVE or CCP for the massive ****** coalitions.

Been in 0.0 lately?
Try undocking 100 people at the same time, massive tidi, something that didn't used to be a problem. Run 100 man gang thru some gates to a fight and watch the tidi build with each jump.
CCP decided that Tidi was "good enough" and started working on the graphics. The problem with that is, shiny pictures require more time to generate. Jita used to handle 2700, they lowered it to 2k, same reason.

Tidi is also not the great balancer that CCP would have you believe either.
I had two alts in that Black legion Super welp, the one in a Dread could not cycle the siege module for the entire fight. My other character in a subcap ran slow, but all the modules actually cycled, at the expected time based on the tidi.

Tidi is a temporary fix, the problem is, at the pace CCP works, that could be years before they come up with something better.

In order to be involved with Sov warfare now, you must have no life, because any fight could turn into a 10 hour yawn fest

You don't know what you're talking about.
But that's hardly surprising.