Your dynamic infantry rendering solution hit critical mass last night.

Excellent, thanks.

+1

They'll probably implement a population cap for each major base/outpost and its surrounding hexes.
-Can't spawn unless you're already inside.
-Can't enter from the outside (restricted area warning)
-Players over the cap get warned to leave.

I wrote this on Nov 13th near the end of beta and sent it to John Smedley, I did not get a response.

Mr. Smedley,

First I would like to say that I just signed up for a 6 month All-Access Pass as well as purchased $150 worth of Station Cash in anticipation for the release of Planetside 2. But I am concerned about Infantry draw distance, warping players and low FPS.

While I do have faith that eventually the crashing bugs will be fixed but it seems like the other two issues are not seeing any progress lately. With only a few days until launch this concerns me greatly that above all else the game needs to feel solid or many people will not stick around to support the game by buying station cash. FPS games, at their core, are about competition. Network and game-play 'smoothness' are critical so players feel like the game is not holding them back, limiting their ability to perform as expected. Being killed by someone who warped out of your cross-hairs who does not seem to be affected by his own lag, sudden frame drops that get you killed, seeing players vanish from your screen as they hit this Infantry draw distance or a Pilot being shot by an Anti-Aircraft MAX unit from such a distance as you cannot see him, these issues are going to cause a lot of people to rage quit.

As for the game smoothness we could really use a lot more advanced graphic settings to tweak our experience, turn off some special effects and get the frames per second under control so its not so sluggish.
I don't mean to rant and I appreciate you reading this far, I just wanted to give my opinion that these things should be the absolute top priority for launch. Thanks for taking the time to respond on Twitter and the forums.
Regards,

PS2 User: voigt

I am not completely sure if this is related to the dynamic render distance, but last night we were taking over a rather busy Amp station. Platoon storms the place, no problem. Gets inside, shots are fired, gens go down and we cap the point. Everyone is having a blast. Once its capped I (Platoon leader) make the call to fix everything and mount up to roll to the next base.

Spawned a sundy and a couple guys hopped in for a ride to the next destination. As we exit the shield on the outer wall, the sundy starts taking damage from nowhere. We check the skies, hills, ect. and nobody is in sight. Sundy gets destroyed and apparently a prowler killed the sundy.

I head back afterwards along the walls and as I come up to the point we got blown up sundenly the prowler is rendered camping the gate. Now im not sure if tanks can cloak of if this is due to the render distance (second time it rendered due to Hex activity going down maybe?) but if this start affecting vechicles then we might as well replace all the guns with different kinds of baseball bats and fit the vehicles out for a demolition derby.

I don't care what gets done as long as I can see all the enemies in the area I'm in.

It is -seriously- disorienting trying to run down a "clear" catwalk only to be greeted by an invisible scatterMAX taking advantage of the render drop.

Network wise: Less position updates (Especially for vehicles that aren't moving at a high speed) and a clientside prediction algorithm to compensate for the bigger gap in position updates.



As for clientside optimization I propose that as a temporary hotfix 2D sprites (or "impostor") are introduced to help cope with render intensity of soldiers to battle FPS dips.

What is a sprite? Take a look at the hologram icon on top a certed sundy, or the squad beacon. It's 2D, but you usually can't tell.
The same thing can be applied to infantry, IF the rendering gets tough, turn all rendered units outside of a 45 degree cone into 2D sprites (or "impostors"), and then turn the sprite into a model if it enters that specific cone. Problem solved.

Edit: Ofcourse I'd like a job.

This... I'm not quite sure how to respond. From a purely technical standpoint that's a possibility, but you're suggesting it for the wrong symptoms.

The problem is not a result of attempting to avoid client-side performance problems. The problem is a result of the server having to send updates to every player in the area for everything ever other player in the area is doing. The hardware apparently can't handle it with the current coding, so SOE's solution was to make it so that it only renders, and thus sends updates, for players in an increasingly smaller radius from you as the population in the area increases.

What you're suggesting would cut down on the number of updates the server has to send - it wouldn't have to send model-specific orientation (aiming, heading, etc.) to the client, as it would only have to send positional and attack data. Unfortunately, it would also look like complete ***, and probably turn more people off than the current problem does.

I updated the post before you replied, I think it covers the other aspect too now.
My bad, I've only had 2 hours of sleep last night.

It does. And that makes a bit more sense.

Unfortunately I still think it'd garner just as much scorn and complaints as the current problem. From a purely gameplay stance it would be acceptable, but I can't think of any way to do it that wouldn't look hella ghetto.

The other problem is clientside prediction. FPS player movement is extremely difficult to predict, and given that infantry gun hits are determined by your client, not theirs, and not the server, you'd actually magnify the current problem with the netcode not updating often enough and causing a lot of hits to register that actually shouldn't.

For now though, anything beats completely 'removing' units from existence.

I can confirm this as well. Whenever you chuck a grenade and get the prompt that says "+100 XP Kill" but the nameplate at the bottom doesn't pop up, that means that you just killed someone outside of the render distance.

One of the devs also posted about it here: http://forums.station.sony.com/ps2/...problem-in-massive-battles.52547/#post-657777

Also, here's a video of how bad it can get:

They do this and most people will stop playing. That would defeat the whole point of PS2. I would leave at once.

Does anyone know if they at least acknowledge this problem exists..? I haven't seen anything. This has got to be one of the very top priorities. It hasn't made the game completely unplayable, for me, but it does somewhat restrain how I play. I feel I've had to learn how to adapt to it.

Unfortunately, the current situation largely defeats the whole point of PS2 as well. You can't have massive battles when you're effectively unable to engage outside of a 10m circle around you and your only effective tactic is blindly lobbing grenades at where you think other players might be.

The situation I reported in the OP is a great example of this. We were massively outnumbered in the last push. We still held them off for over half an hour after the rendering distance removed our ability to fight in any other fashion than grenade spam due to two factors and two factors only:

1. I'm pro at parking Sunderers in places that make them absurdly hard to get to for attackers.
2. The attackers had only two very constrained routes to get to the point, and we had more room to avoid likely grenade landing points. Understand, we were horribly exposed to regular infantry weapons fire - without the rendering distance problem they'd have swept us clean off of our defensive positions in no time at all, but since they had no meaningful ability to engage us at range they couldn't take advantage of that.

The only reason we even lost the plant is because the sheer volume hailstorm of hostile grenades effectively made it impossible for us to survive getting to the control point without being a light assault, and eventually the superior numbers made flying over to the point instadeath.
Keep in mind here, I'm not complaining that we lost the plant. By all rights we should have gotten swept out of there LONG before it actually happened, and that's no slight on the skill of the attackers, or compliment to the skill of the defenders. It was simply that we were both crippled in our ability to attack each other.

We dont know how big the cap would be, if it is a temporary solution etc.

Either way, I'm pretty confident whatever numbers they settle on will still be "massive" when compared to any other online FPS out there.

If you're looking for perspective in the MMO industry on this problem, the only other game that comes close to PS2 in scale is GW2's WvW. They also hit a wall in how many players can be displayed in a specific area, while balancing performance on player's computers and the game servers. GW2 caps their maps around 500 players, normalizes the models of enemy players, and still in large battles you will have invisible enemies etc.

Both games use different engines, so you can't make a direct comparison. But they're running into the same problems and as far as I can tell PS2's is doing a superbly better job.

*Edit
I should mention Eve, but I don't play it so I can't speak as to how they do it there. I know they use "time-dilation" to slow things down so the servers can keep up, but I dont know how they handle client limits. I figure they just have empty target boxes or something. Either way, I don't think eve's solution would work well in FPS land.

Holy cow! AudieSmurfie that right there is irrefutable proof of this issue. That was only a few hundred or so right? And the game boasts thousands? What I saw was fluid gameplay but loss of the ability to shoot any of those guys that were just right there on that hill next to the sunderer. Scary!

For getting massive battles to work, this issue really needs to be fixed if PS2 is to have a sustained life.

Now I'm no programmer, but my brother is and he's too lazy to post here so I'm going to try to explain what he told me. Basically, every time to you add a player to a game, it almost exponentially increases the amount of computations and network traffic. The server has to update the additional player with the added information of everyone else's location, and of course add that players location and actions to the queue. Each person you add means another player who needs all that information. How Planetside 2 even manages to run at all really blows my mind. Maybe the solution isn't technical, but design based.

The mass zerging on horribly designed bases like the Tech Plant causes this kind of situation. You have two spawn points right next to each other, with a shield in between. Every 8 seconds 20 people die, and immediately respawn right next to each other. Over time, a huge zerg builds up. This is inevitable with the current spawn selection system, base design, spawn rate, capture rewards, and lack of defense rewards. There's simply no reason to split the zerg up and push players into other regions. Most players would rather pick the spawn right next to where they perished, because they know they'll be in the fight instantly. Why else would 100s of players slam themselves into a tech lab's shield, or try to get through that horribly designed back door for 1 hour? They literally pour their bodies into a grinder because it's the closest and easiest way to get back to the fight. I've seen more stupid at 1 tech lab fight last night than in all my gaming career combined.

TL;DR The solution may not be technical but a complete re-work of design in order to create incentive for spreading out, also Tech labs are stupid.

I'm sorry. What is the point of this thread?

We already have threads reporting this, we even have a response from the CEO of the company. Its being worked on. They are aware.

Do some of you really think posting this over and over will make the solution come any faster? Going by some of the posts, you guys are acting like they are ignoring the issue, and just sitting around doing nothing about it.

That is false.