People seem to be missing the point, when a game is such high
fidelity, that takes up gpu amd cpu processing power away from network
packet , physics and geometry computing. Geometry is not only a I see
it so it is there, the gpu and cpu has to compute the position back and forth
from the server to all players on screen.
Throw in network packets from 100 other players that your machine is computing.
Network
lag, you lagger get out of the server, your ping is too high leave, I
guess had nothing to do with the player according to a few posters here.
You're conflating network performance with client-side rendering performance.
Simply put, people need to not evoke the generally false argument that the engine or tech presented is innovative. It can still be a perfectly fine game for you to enjoy without trying to scrape for excuses.
What false argument?
Are you saying that MMOs don't need to tone down graphical complexity, physics sim and all that jazz as to achieve a balance of MP density vs visual quality?!
I mean the evidence it's literally in every MMO ever, even Unreal 5 with its beefy rendering improvements, with the upcoming MMO Throne and Liberty, on MP density situations such as world bosses with dozens of players, the client rendering performance is still tanking.
So it's absolutely fair to consider client performance on the MP scalability of an online game as this, because it is one of the limiting factors.
What's the point on scaling the server capacity if it the scenarios of increased density won't output a playable performance? One can't forget on SC a player is not just its character, it's that + its ship + vehicle, which is where the main simulation & rendering costs are.
Optimization passes have to kick in obviously, stuff like culling ship interiors, etc.
Which is why if you face SC against PS2 in terms of network costs of due to the amount and complexity of entities and rendering costs on higher density scenarios, it's quite different realities.
UE's challenge with networking is very well known and has nothing to do with visuals. None of the server performance has anything to do visual complexity or quality, that is 100% client side. The server couldn't care less what visual bells and whistles your client is running, it functions purely on numbers. To the server you might as well be playing as a block of untextured wood. All it cares about are numbers, IDs, states, etc. Player ID 9823673 on the mount 3487654 with the weapon 19829813 is all they process and couldn't give a toss about the complexity of the displayed assets.
The issue of MMOs being less visually appealing (at least until BDO and such came along) has nothing to do with server performance. However, the need to render potentially hundreds of players around you, especially in large hubs, does have an effect on your local machine, so there's that. But it has nothing to do with servers, those just (in a simplified form) pass on to your client the message 'player avatar A standing at position X,Y,Z, doing the picking nose emote, wearing armour ID XX' and your client then takes over actually rendering all that and showing it to you.
And then there is also the simple fact of your target audience and their machines. Many MMOs voluntarily lower their visual fidelity to accommodate a wider range of players: see WoW for example. And finally, there's the financial side - if you can get satisfactory performance and have 99% happy player population, why would you spend tons of cash on extra server performance just to get the extra 1%? Diminishing returns and all that...
You might argue that SC is adding a lot of subsystems to each ship that games previously did not bother with (although you would probably find examples to the contrary), but then those will again be reduced by the servers to their IDs and states represented by a few short variables.
Even if you take into account all that, you inevitably arrive to the point mentioned above: if the number of systems the games manage have increased over the years, so has the servers' capacity and processing power. You may argue that SC has ships 10 time more complex in terms of game systems, but today's top of the line server CPUs are probably a hundred times faster than what we had 10 years ago.
Yeah, unfortunately bacon's rant was a bit of a distraction. Graphics impacts client, yes, but graphical performance across the board has made absolutely massive leaps in the last couple decades. Nanite, Lumen, and the more commonplace leveraging of raytracing and volumetric elements all makes the rendering capability of modern engines billions of times more complex in terms of rendering capability. You could stick considerably more graphical detail into a rock on the ground in UE5 than in a full ship in the adapted Cryengine 3 being used for Star Citizen.
As for Solareus' last comment. Star Citizen isn't a mesh network for clients. You aren't getting packets from every player around you. You're only getting packets from the server. Additionally, that data is unpacked and managed traditionally by the CPU while the majority of rendering is handled by the GPU, even the physics which is handled by discrete compute processors in a GPU that are separate from the nodes used for rendering polys and visuals, dedicated directly to accuracy in computation for physics and AI use.
So I guess we got Solar and Bacon both conflating things now.
To which again, People are free to enjoy what there is to be enjoyed with Star Citizen, but these deep digs for excuses that don't hold up to technical scrutiny are just a bother to see happening all the time.
@Uwakionna it's nice to know you were involved in Firefall, it was a great game while it lasted. I spent a bit of time in it back in 2015 (? - not sure if my memory serves) and had a lot of fun. Then things changed (again), as was the case throughout Firefall's short life, and I wasn't that interested any more. But from the earlier days I have great memories (albeit fairly hazy at this point). Pity how it all ended.
Perhaps you and @lotrlore could have a chat, he was planning to do re-reviews of older MMOs and while Firefall cannot be re-reviewed in such way, perhaps an article about Firefall's development could be crafted.
@BrotherMaynard I'd only be able to comment on up to around 2012, as I left during the exodus that took place alongside the deal made with The9. That game's development was interesting for it's technical achievements for the time, but rough from the management and inability to just carry things forward until it was too late. Everything after that was more waffling and damage control from what I learned from coworkers that stayed.
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
If you are holding out for the perfect game, the only game you play will be the waiting one.
@BrotherMaynard I'd only be able to comment on up to around 2012, as I left during the exodus that took place alongside the deal made with The9. That game's development was interesting for it's technical achievements for the time, but rough from the management and inability to just carry things forward until it was too late. Everything after that was more waffling and damage control from what I learned from coworkers that stayed.
Well, I just had to check Steam, as I was curious myself - the game is still in my library and Steam says I last played on 10 August 2014. So I would guess the core of my Firefall experience was back in 2013 and up to mid-2014.
With Planetside, Tera and Vindictus it was one of the very few action MMOs out there and it had so much potential. What a shame, really, how it all turned out. The mismanagement stories are well known, of course (M. Kern visited here a couple of times a few years back to drum up his new Thumping-thumper-thumpsomething game, probably counting on people not remembering his Firefall days).
Anyway, I don't want to derail the SC discussion too much, but I do think it would fit well in the re-reviews concept, especially because for a while Firefall was a very high-profile newcomer for which many MMO players had high hopes. The story has everything: love, hope, hate, extravagance, drama...
@Uwakionna it's nice to know you were involved in Firefall, it was a great game while it lasted. I spent a bit of time in it back in 2015 (? - not sure if my memory serves) and had a lot of fun. Then things changed (again), as was the case throughout Firefall's short life, and I wasn't that interested any more. But from the earlier days I have great memories (albeit fairly hazy at this point). Pity how it all ended.
Perhaps you and @lotrlore could have a chat, he was planning to do re-reviews of older MMOs and while Firefall cannot be re-reviewed in such way, perhaps an article about Firefall's development could be crafted.
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
It's done in. And you know it. Of course, in that game design, when the elevators killed you it was intentional, not an eight year bug train.
Most good designers, when given a half billion dollars, can come up with a working game. Not using the aeternal alpha excuse.
Oh, and how's SQ404?
If you are holding out for the perfect game, the only game you play will be the waiting one.
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
It's done in. And you know it. Of course, in that game design, when the elevators killed you it was intentional, not an eight year bug train.
Most good designers, when given a half billion dollars, can come up with a working game. Not using the aeternal alpha excuse.
Oh, and how's SQ404?
So why didn't Warren make it? He got the 8 years after all...
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
It's done in. And you know it. Of course, in that game design, when the elevators killed you it was intentional, not an eight year bug train.
Most good designers, when given a half billion dollars, can come up with a working game. Not using the aeternal alpha excuse.
Oh, and how's SQ404?
So why didn't Warren make it? He got the 8 years after all...
Squadron development is ongoing and looking good
How's that SQ404 going?
If you are holding out for the perfect game, the only game you play will be the waiting one.
Even if you take into account all that, you inevitably arrive to the point mentioned above: if the number of systems the games manage have increased over the years, so has the servers' capacity and processing power. You may argue that SC has ships 10 time more complex in terms of game systems, but today's top of the line server CPUs are probably a hundred times faster than what we had 10 years ago.
What does this have to do with anything? It's the server-sided simulation costs that matter, and those... ain't cheap, CPUs are now stronger, so are the simulation requirements of entities on an online game these days.
idk why this discussion turns into "servers don't do graphics" like that's what I said. What I say and my point:
- MP density is a factor on how far can an online game, especially MMO scale, can take graphics and the server capacity on the same area.
You can have your server software that can simulate 1000 clients in a city, that is just one side of the question... Because that generates too much MP density that the player characters + their vehicles, under the context of being more complex entities and beefier assets by nature will sink client performance... Then you have a performance nightmare.
This is why GW2 WvW, when players started playing in big zergs all the time, rendering performance got awful when a client had to load those zergs in... their solution was the path of agressive optimization (you could see invisible players, just their name tags, or barebones default models until it progressively loaded everything). This also the solution for world bosses.
On my opinion I would say generally upcoming MMOs featuring high visual quality and more detailed and complex mechanics (such as physics and AI), will also feature and design for lower caps per area server/instance, instead of going many for the sake of many.
On that note, if SC does the mesh (per major planet or moon area) with the current game-server cap per area, that to me is would already be a densily populated world.
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
It's done in. And you know it. Of course, in that game design, when the elevators killed you it was intentional, not an eight year bug train.
Most good designers, when given a half billion dollars, can come up with a working game. Not using the aeternal alpha excuse.
Oh, and how's SQ404?
So why didn't Warren make it? He got the 8 years after all...
Squadron development is ongoing and looking good
How's that SQ404 going?
Still developing with Chris Roberts at the helm, How's SShock Going?
Roberts was right about their being a well of support for space games, when the big companies didn't believe it. Sadly, Roberts is good at the elevator pitch level, but terrible beyond that. It's been nearly 30 years since he last released a game. He's never successfully helmed a game outside of the structure of Origin Systems.
Having a narcissistic micromanager with delusions of grandeur as your head developer is not a great proposition. And as long as they can keep the money coming in, there's little incentive to prepare an actual release.
Not to even parse the delay of the single player SQ42. If SQ404 isn't spectacular, the house of cards could fall. Delay, delay, delay.
How's your hero's Warren Spector System Shock game going?
To his credit, Warren is not wasting people's money with vainglorious promises and cult marketing.
How are your predictions for SC/SQ404 beta and release in 2018, 2019, and 2020 doing?
So you think he worked for free lol
When have I predicted released dates for anything? I've been playing and enjoy Star Citizen whatever is available for years which should be clocking on the several thousands of hours of gaming
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
It's done in. And you know it. Of course, in that game design, when the elevators killed you it was intentional, not an eight year bug train.
Most good designers, when given a half billion dollars, can come up with a working game. Not using the aeternal alpha excuse.
Oh, and how's SQ404?
So why didn't Warren make it? He got the 8 years after all...
Squadron development is ongoing and looking good
How's that SQ404 going?
Still developing with Chris Roberts at the helm, How's SShock Going?
Still developing with Chris Roberts at the head. Yep, you'll be saying that for a long time.
Origin had to bring Spector in to handle all the logistics of Wing Commander development, as Chris Roberts wasn't capable of it. Then they had to bring in a programmer late in the game to fix all Chris's wonky ship coding. The same programmer who later got chosen to head WC2.
Now if Chris had just learned from his mistakes....
Digital Anvil 2 says hello.
Star Citizen, the Aeternal Alpha, with a half billion dollars invested, and still no server meshing or persistence. I'm sure they'll get there sometime. Ya' know, when the aliens come....
Best news for SC is that Chris is focused on other stuff; so maybe they can make some sustained progress. Or maybe there will be more updates like the last disaster.
If you are holding out for the perfect game, the only game you play will be the waiting one.
Even if you take into account all that, you inevitably arrive to the point mentioned above: if the number of systems the games manage have increased over the years, so has the servers' capacity and processing power. You may argue that SC has ships 10 time more complex in terms of game systems, but today's top of the line server CPUs are probably a hundred times faster than what we had 10 years ago.
What does this have to do with anything? It's the server-sided simulation costs that matter, and those... ain't cheap, CPUs are now stronger, so are the simulation requirements of entities on an online game these days.
idk why this discussion turns into "servers don't do graphics" like that's what I said. What I say and my point:
- MP density is a factor on how far can an online game, especially MMO scale, can take graphics and the server capacity on the same area.
You can have your server software that can simulate 1000 clients in a city, that is just one side of the question... Because that generates too much MP density that the player characters + their vehicles, under the context of being more complex entities and beefier assets by nature will sink client performance... Then you have a performance nightmare.
This is why GW2 WvW, when players started playing in big zergs all the time, rendering performance got awful when a client had to load those zergs in... their solution was the path of agressive optimization (you could see invisible players, just their name tags, or barebones default models until it progressively loaded everything). This also the solution for world bosses.
On my opinion I would say generally upcoming MMOs featuring high visual quality and more detailed and complex mechanics (such as physics and AI), will also feature and design for lower caps per area server/instance, instead of going many for the sake of many.
On that note, if SC does the mesh (per major planet or moon area) with the current game-server cap per area, that to me is would already be a densily populated world.
Might need to work on your phrasing then, as the example you put after this said it more clearly than even your statement here. Your statement here still reads like you're arguing graphics impacts servers. Though your example also seems to confuse things.
It's not the server capacity to perform that's impacted by the graphics, it's the ability for the client and client-side engine to render the entities and actions taking place.
A separate issue to that would be entity complexity in terms of components the server has to track, which was touched upon with the likes of hitboxes and physics.
Your GW2 example falls back on client side rendering performance again instead of server performance. That render cap was for making sure the clients didn't crash from exceeding the render limit and bogging the client down. The server still has to track all of those entities and their interactions all the same, regardless of how clients perceive the other player entities.
It's important to understand the distinction between these elements and how they impact the game on two different ends of the pipeline.
It's notable that your response was originally tacked onto another conversation where the topic was correcting the division of graphical fidelity versus server performance as another had made this mistake as well. It's easy to assume you are making the same argument when your address is to that topic, and you seem to make the same mistake of conflating elements.
But certainly, entity complexity on visuals does matter as far as the client side rendering capability is concerned. That does limit how many people can reasonably be rendered in an area, but is a separate challenge to server side issues around entity complexity, which is a subject of different mechanics and math going on and how effectively the server can organize and deliver data, noting ado about visual complexity.
So it's not the "server capacity" being affected there, it's client render limits. You can have more entities per server than the client's engine can render as long as you can trust the game is distributing players broadly enough that they are not exceeding the render limit.
The main reason to scale server capacity relative to client render limits is generally if one is concerned with PvP, as you don't want to be dropping models in a situation where it gives certain players a potential advantage. Hence the default/doll bodies games like GW2 and ESO will use as simplified substitutes to manage performance. That's a gameplay fairness balance due to client-side rendering limitations. Not a server-side capacity performance limitation.
If we were to tie that to the discussion you originally responded to, then we'd be examining not the hit boxes or the physics, but the raw model complexity which is a subject I addressed in my prior post.
And certainly you can have a perfectly fine game within such a scope. But then it goes back to what I had stated before, considering it falls within the scope of present technical ability based on the sliding scale of technical entity complexity versus density. Making a trade off of a few hundred less entities so you can add more polygons to them, on tech that should be a decade further along in capability compared to titles one to two decades older with higher headcounts but lower mesh complexity, is within a pretty normal scope of things going on. (EDIT: had to fix my own oopsie in there)
Hence; "People need to not evoke the generally false argument that the engine or tech presented is innovative. It can still be a perfectly fine game for you to enjoy without trying to scrape for excuses."
Comments
As for Solareus' last comment. Star Citizen isn't a mesh network for clients. You aren't getting packets from every player around you. You're only getting packets from the server. Additionally, that data is unpacked and managed traditionally by the CPU while the majority of rendering is handled by the GPU, even the physics which is handled by discrete compute processors in a GPU that are separate from the nodes used for rendering polys and visuals, dedicated directly to accuracy in computation for physics and AI use.
So I guess we got Solar and Bacon both conflating things now.
Perhaps you and @lotrlore could have a chat, he was planning to do re-reviews of older MMOs and while Firefall cannot be re-reviewed in such way, perhaps an article about Firefall's development could be crafted.
If you are holding out for the perfect game, the only game you play will be the waiting one.
Well, I just had to check Steam, as I was curious myself - the game is still in my library and Steam says I last played on 10 August 2014. So I would guess the core of my Firefall experience was back in 2013 and up to mid-2014.
With Planetside, Tera and Vindictus it was one of the very few action MMOs out there and it had so much potential. What a shame, really, how it all turned out. The mismanagement stories are well known, of course (M. Kern visited here a couple of times a few years back to drum up his new Thumping-thumper-thumpsomething game, probably counting on people not remembering his Firefall days).
Anyway, I don't want to derail the SC discussion too much, but I do think it would fit well in the re-reviews concept, especially because for a while Firefall was a very high-profile newcomer for which many MMO players had high hopes. The story has everything: love, hope, hate, extravagance, drama...
Meanwhile, can you say the same about Spectors System Shock lol? He he even on the project? Didn't they started in 2015? 8 yEaRs dEvEloPmEnt n0t rEleAsEad yeT. Duh.
If you are holding out for the perfect game, the only game you play will be the waiting one.
Squadron development is ongoing and looking good
How's that SQ404 going?
If you are holding out for the perfect game, the only game you play will be the waiting one.
https://biturl.top/rU7bY3
Beyond the shadows there's always light
idk why this discussion turns into "servers don't do graphics" like that's what I said. What I say and my point:
- MP density is a factor on how far can an online game, especially MMO scale, can take graphics and the server capacity on the same area.
You can have your server software that can simulate 1000 clients in a city, that is just one side of the question... Because that generates too much MP density that the player characters + their vehicles, under the context of being more complex entities and beefier assets by nature will sink client performance... Then you have a performance nightmare.
This is why GW2 WvW, when players started playing in big zergs all the time, rendering performance got awful when a client had to load those zergs in... their solution was the path of agressive optimization (you could see invisible players, just their name tags, or barebones default models until it progressively loaded everything). This also the solution for world bosses.
On my opinion I would say generally upcoming MMOs featuring high visual quality and more detailed and complex mechanics (such as physics and AI), will also feature and design for lower caps per area server/instance, instead of going many for the sake of many.
On that note, if SC does the mesh (per major planet or moon area) with the current game-server cap per area, that to me is would already be a densily populated world.
Still developing with Chris Roberts at the helm, How's SShock Going?
If you are holding out for the perfect game, the only game you play will be the waiting one.
It's not the server capacity to perform that's impacted by the graphics, it's the ability for the client and client-side engine to render the entities and actions taking place.
A separate issue to that would be entity complexity in terms of components the server has to track, which was touched upon with the likes of hitboxes and physics.
Your GW2 example falls back on client side rendering performance again instead of server performance. That render cap was for making sure the clients didn't crash from exceeding the render limit and bogging the client down. The server still has to track all of those entities and their interactions all the same, regardless of how clients perceive the other player entities.
It's important to understand the distinction between these elements and how they impact the game on two different ends of the pipeline.
It's notable that your response was originally tacked onto another conversation where the topic was correcting the division of graphical fidelity versus server performance as another had made this mistake as well. It's easy to assume you are making the same argument when your address is to that topic, and you seem to make the same mistake of conflating elements.
But certainly, entity complexity on visuals does matter as far as the client side rendering capability is concerned. That does limit how many people can reasonably be rendered in an area, but is a separate challenge to server side issues around entity complexity, which is a subject of different mechanics and math going on and how effectively the server can organize and deliver data, noting ado about visual complexity.
So it's not the "server capacity" being affected there, it's client render limits. You can have more entities per server than the client's engine can render as long as you can trust the game is distributing players broadly enough that they are not exceeding the render limit.
The main reason to scale server capacity relative to client render limits is generally if one is concerned with PvP, as you don't want to be dropping models in a situation where it gives certain players a potential advantage. Hence the default/doll bodies games like GW2 and ESO will use as simplified substitutes to manage performance. That's a gameplay fairness balance due to client-side rendering limitations. Not a server-side capacity performance limitation.
If we were to tie that to the discussion you originally responded to, then we'd be examining not the hit boxes or the physics, but the raw model complexity which is a subject I addressed in my prior post.
And certainly you can have a perfectly fine game within such a scope. But then it goes back to what I had stated before, considering it falls within the scope of present technical ability based on the sliding scale of technical entity complexity versus density. Making a trade off of a few hundred less entities so you can add more polygons to them, on tech that should be a decade further along in capability compared to titles one to two decades older with higher headcounts but lower mesh complexity, is within a pretty normal scope of things going on. (EDIT: had to fix my own oopsie in there)
Hence; "People need to not evoke the generally false argument that the engine or tech presented is innovative. It can still be a perfectly fine game for you to enjoy without trying to scrape for excuses."