Regarding Lag on the server

[Tibsun]

I'd like to see a stress test if rubberbanding occurs on a high player counts when AA and backgunners are disabled.
So here the question to the community:

Question 1:
Do you consider to test the main ATAG server without AA or AA ai disabled to see how much lag occurs?

Question 2:
If it's successful lag-free at 80+ player counts, do you consider having it disabled permanently?

[pitt]

Its not only that rubberbanding for me,sometimes it stutters like a slide show(even with Lan)....

[PanTast]

Some missions make no sense without AA.

All bombers make no sense without rear gunners.

So my answer is no

[Tibsun]

Some missions make no sense without AA.

All bombers make no sense without rear gunners.

So my answer is no

I'd play on a AI server or SP if I wanted to fly against AI.

Yes, AA is nice, but if I had to choose between little rubberbanding with AI and lag-free, I'd choose lag-free

If population is really high there would be no AI needed, and a every human in AA is worth 10 of them

But that's my opinion and I'm probably staying alone with it... connection or any other kind of lag on MP drives me nuts.. and many others.

[Spicysauced]

Thats a difficult question.
If I understood it right you mean no AI in turrets in bombers too?
Then folks who want to fly them are completely vulnerable ..
On the other had I REALLY would like to enjoy a lag free experience.

Conclusion: If this proposal would be ONLY temporary I could live with it. But as I said, temporary only, until the cause of the problem on a deeper level is fixed.

Regards,
Spicy

[farley]

Don't want to start a new thread so will ask here.

What is the Server #2 for?

Is it for overflow (for eg now when Server #1 if full at it's new limit of 60)?

Or in case people want to fly a different mission?I know that it has been around forever, but just never bothered to ask before.

Thanks for you r reply to this (even if it is a well deserved smart ass comment)

[Tibsun]

Don't want to start a new thread so will ask here.

What is the Server #2 for?

Is it for overflow (for eg now when Server #1 if full at it's new limit of 60)?

Or in case people want to fly a different mission?I know that it has been around forever, but just never bothered to ask before.

Thanks for you r reply to this (even if it is a well deserved smart ass comment)

I'd gladly abuse it as testing server I'm asking in OP or as AI free 2nd server

[=vit_unit=]

Disabling AA and backgunners (especially backgunners) is pointless for bomber to be on the server. So it is going to be the purest dog-fight server at all.

My suggestion is to disable trees colission somehow. I know everybody waited this colision. But it seems to be a less blood.

[Tibsun]

Most players do anyway only fighters and the few that are into bombers are usually sitting anyway in the back...
Saying AI is saving your arse would be also a lie, you get shot down and in the end you would jump into the back anyway.
Only good thing it does is calling out enemies, no control needed for calling out.
Escort is mostly the important factor.
If you do a big human raid, you probably also have escorts and some human gunner preferable for people flying only with keyboard (lol)

There's another thing that stresses server. AI controled planes. Reason of bomber locked at <30 player
And abdoned planes:
Player takes chute -> plane server controlled (ai takes over and crashes it by having a wing damaged).
Player exit plane or spawn new one after being killed or other reason -> plane server controlled (ai takes over and crashes it by having a wing damaged).
Player hits ground -> destroy plane server controlled.


Anyway my OP question was more about the amount of AA and to test it if is affecting MP at all, which I have no idea.
AA looks great AA is fun, but everybody ingores it and hates to get hit by it out of nowhere.
Not saying it's bad but as said, I'd rather have a completely smooth server than hundreds of ai controlled flak.

[ATAG_Flare]

Most players do anyway only fighters and the few that are into bombers are usually sitting anyway in the back...
Saying AI is saving your arse would be also a lie, you get shot down and in the end you would jump into the back anyway.
Only good thing it does is calling out enemies, no control needed for calling out.
Escort is mostly the important factor.
If you do a big human raid, you probably also have escorts and some human gunner preferable for people flying only with keyboard (lol)

There's another thing that stresses server. AI controled planes. Reason of bomber locked at <30 player
And abdoned planes:
Player takes chute -> plane server controlled (ai takes over and crashes it by having a wing damaged).
Player exit plane or spawn new one after being killed or other reason -> plane server controlled (ai takes over and crashes it by having a wing damaged).
Player hits ground -> destroy plane server controlled.


Anyway my OP question was more about the amount of AA and to test it if is affecting MP at all, which I have no idea.
AA looks great AA is fun, but everybody ingores it and hates to get hit by it out of nowhere.
Not saying it's bad but as said, I'd rather have a completely smooth server than hundreds of ai controlled flak.

Saying that AA and AI gunners are useless is in my opinion a ridiculous idea.

Many pilots like using bombers. Without AI gunners, fighters would be free to sit on a bomber's six without worrying about getting hit. Believe me, AI gunners may not shoot down tons on planes, but they do do a pretty good job of forcing attackers to make more difficult approches. Also, I did get PKed by an AI blenny gunner the other day. No AI gunners would be awful. There wouldn't be much point in flying a bomber. And bombers turn the maps.

Same thing with flak. You say that nobody likes randomly getting killed by flak, but that's the point. Without flak, there would be a lot more vulching - nobody likes getting vulching! It should be risky to fly over enemy territory, no flak would take away a lot of the fun.

Your changes you suggest would turn ATAG into a pure dogfight server. It would not be as fun.

I'd take a bit of rubber banding over no flak or AI gunners any day.

[PanTast]

Especially eliminating flak requires a lot of editing the missions themselves. My missions simply would be not able to be kept, because many flak installations are mission targets. So I guess the best thing would be, you just sit dorn and open the FMB and make your dream mission by yourself. When you're done, we can test it on server2 someday. And as far as I can tell the gunner ai requires a patch to be changed.

o7

[Baffin]

Especially eliminating flak requires a lot of editing the missions themselves. My missions simply would be not able to be kept, because many flak installations are mission targets. So I guess the best thing would be, you just sit dorn and open the FMB and make your dream mission by yourself. When you're done, we can test it on server2 someday. And as far as I can tell the gunner ai requires a patch to be changed.

o7

Is there any way to reduce the frequency of Flak shots, if this would reduce the stress on the system? Less is better than none if this is possible.

[Tibsun]

Personally I'd like to have AI that are not in the vicinity off players disabled or in sleep mode, so when not used, server wouldn't use processing power for aiming, looking around and doing nothing.

It's really weierd how a few abdoned twin engine planes with pilots and gunners abdoned to the ai server control can cause that much lag by just sitting and doing nothing...

[Spicysauced]

Im happy to report, that at least for me the lag was not existing today.
I was on ATAG 1, with the mission on the Isle of Wight, no idea what thats called
I could play like before the patch, no rubberbanding or lag, no stutter, nothing.
The server was relatively full, fluctuating from 47-ca 70 players.
So if you guys did something - it worked.

[PanTast]

Thx for the nice comment, Spicysauced

I have the suspicion that some of the lag is caused by the graphics and network settings of the players. I guess many think "what the hell its an old game and my computer is quite good, so I can max out everything"

But your comment seems to proove that Flak isn't the cause of rubberbanding. NewDawnFades is the map with the most Flak.

[Spicysauced]

Thx for the nice comment, Spicysauced

I have the suspicion that some of the lag is caused by the graphics and network settings of the players. I guess many think "what the hell its an old game and my computer is quite good, so I can max out everything"

But your comment seems to proove that Flak isn't the cause of rubberbanding. NewDawnFades is the map with the most Flak.

Yes, that seems to be right.
However, I just came from the same mission and saw that you raised the player limit up to 200. Imo thats a bit much, I was not able to achieve consistant 60 FPS anymore, on the Isle of Wight it was more like 30 or less, but it got better once in the air.
This has nothing to do with network caused lag though, that is still completely gone.
If I might suggest something, it would be that you lower the limit on server 1 to 90, maximum 100 and that you let server 2 run the same set of missions as a backup, if #1 might be full.
I was on server 2 myself two times and noticed labels were on and other weird stuff, so as I said - also for future preperation for when TF5 comes out which will draw even more players into the game, Imo it would be
good if server #2 ran the same mission set as #1.
Would that be possible? I think that would make more sense than ever before.

Have a good day!

[PanTast]

The playerlimit was set by Colander because he found a workaround for the lag issues. Currently he is also in charge of server2.

[Baffin]

Observations made during the "Big Test", 12/23/2017:

1845Z - 700' AGL, westbound, 5 miles west of RAF Shoreham. My Spitfire on the right wing of an Hurricane about 40' away experienced Rubber Banding fore and aft of about 3 airplane lengths. As I forced my way in closer, the rubber banding stopped, but then the "Warping" started where the Hurricane smoothly accelerated forward about 1/4 mile. As I sped up to catch him, he warped backwards to a point near to where he was originally. Unacceptable for formation flying. I don't know how many players were in the game.

1910Z - 2000'AGL, Eastbound, along the shoreline at RAF Shoreham. I joined a Spitfire with no rubber banding, but as I closed in to fingertip, at about 20' the "Warping" occurred several times. Another player advised me that the player count was 99 at the time.

0300Z 12/23 - 50 players online: Formation is OK with minor jiggling once you are "IN". (Very much like v4.312) Getting there (10' or closer) still requires working through some Minor Rubber banding. ACM was a lot of fun, having run out of ammo, I chased a BF-109 around at very close range with no bad effects.

Unscientific, subjective observations when compared to v4.312 described as "Good":

Player count at 60 last night yielded "Very Good" performance, bombing, pursuit and close formation

Player count at 101 yielded "Below Average" performance for pursuit and formation. Bombing not tested.

Player count at 200 yielded "Poor" performance for pursuit and formation. Bombing not tested.

Recommendation: Limit the server to 60 players Maximum until the solution is found.

[Vlerkies]

seemed pretty good to me the whole arvie/night.

100+ people online, arguably 4.3 was certainly no better.

[Baffin]

12/24 2020Z Fall Rot v2.3d Observation for about an hour...

Player count at 69 yielded "Below Average" performance for pursuit and formation. Rubber banding and warping within 50 ft of Blenheim from Spitfire Mk 1. Not terrible, but not good.

Performance was noticeably better earlier, when 50 players were present in the same map

EDIT: 12/27 2100Z - 74 Players, Brothel Map. Unable to dogfight with a BF 110 at 2000' over RAF Rochester due to warping and rubber banding. When I got in close, he'd just shift around. I left and watched a movie...