missing info Missile guidance desync between shooter and all other aircraft in MP

[TAW_Blaze]

There seems to be strong a desync of missile guidance between shooter aircraft and everyone else in multiplayer. Track and ACMI recordings from server, shooter and target are all available and time synced for easy comparison. I recommend using HUD view on the missile to very easily see whether it's guiding or flying ballistic.

 

We tested it in MP with a 3rd person hosting so we are only fighting client vs client. The issue does not seem to exist in SP.

 

There are 2 main cases:

 

1. Shooter drops lock immediately after launch. In this case:

- shooter perspective: missile flying ballistic, when it reaches around pitbull range it snaps to target

- everyone else (including target, server, etc): missile continuously tracking despite no support from launching aircraft, at all ranges. Missile flies past you and when it hits you according to shooter perspective you explode, even though on your screen the missile is miles behind

 

2. Shooter keeps supporting with radar

- only lofting is desynced, otherwise guidance is matching between shooter and target perspectives

 

It seems to behave similarly with the 120 and not just the 54 but due to the ranges this issue is the most obvious with the 54. In the 2nd 120 shot you can similarly see how for a while it is going ballistic and when reaching pitbull range it snaps to target

 

Overall the testing procedure was:

- fire 54 at 30-40 miles, turn cold

- with the 120 fire between 20-25 miles and immediately radar off after launch

- after radar is not supporting, target turns 40-50 degrees to the right

 

*last engagement shows 54 shot being supported by aircraft radar until impact.. every other scenario the launching aircraft immediately aborts supporting the missile

 

ACMI attached in a zip. Trk are too large, but I uploaded them here: https://www73.zippyshare.com/v/ZFMLn2M7/file.html

 

Now this raises some questions like..

- Why are other clients than the shooter seeing as if the missile is tracking them all along, when it is in fact not?

- What is the primary source here? To me it seems like everything is calculated based on shooter client.. if the missile hit you on the shooter's end, you die, if the missile hits the ground on the shooters end, it disappears from the sky..

- From shooter perspective what is causing the missile to reacquire at pitbull range? Is it the overdone seeker FoV or something is triggering it to look there (data from server side)?

- Is missile data sync event based and not continuous?

 

@chizh, GGTharos if you could comment on this that would be great. Unfortunately without deep DCS software knowledge it's difficult to make the right conclusions for the things that happen.

guidance_desync_acmi.zip

Edited March 27, 2019 by <Blaze>

[Falcon_S]

There seems to be strong a desync of missile guidance between shooter aircraft and everyone else in multiplayer. ..

 

Ping?

 

I'm not NASA scientist and my calculation is wrong maybe but...

 

Ping: 200ms (i don't know... maybe just for one client or sum of both) = 0.2sec (a fifth of a second)

Missile Speed: 2 Mach = 680m/s at the moment when missile pass/miss Blaze

1sec/0.2 = 5

2sec = 2000ms (2 sec after missile miss you and then hit ...even behind)

 

 

so...

680/5=136; Missile speed is 136m in 0.2sec

 

136m*2000=272000m/200ping=1360m distance hit behind you

 

On attached pic distance is 1600m but because missile losing speed. Even i'm not sure in my science :) i'm pretty sure that all this "desync" problem is in relation with ping.

 

Question is: Is it possible to synchronize two objects at all if one does not exist on some place in some/current time?

 

Calculators links that i use:

https://www.grc.nasa.gov/www/k-12/rocket/machu.html

http://www.unitconversion.org/time/milliseconds-to-seconds-conversion.html

[TAW_Blaze]

I had 100 ms. But I doubt it's related to ping. How would you explain multiple seconds worth of delays based on pings that are < 200 ms? I'm pretty sure these cannot accumulate over time.. otherwise nothing would work in MP.

[Falcon_S]

...How would you explain multiple seconds worth of delays based on pings that are < 200 ms? I'm pretty sure these cannot accumulate over time.. otherwise nothing would work in MP.

 

...any ping will give same result here (this math surprised me :joystick: )

 

Ping:10 = 0.01sec

Missile Speed: 2 Mach = 680m/s

1sec/0.01 = 100

2sec = 2000ms (2 sec after missile miss you and then hit ...even behind)

 

 

so...

680/100=6.8; Missile speed is 6.8m in 0.01sec

 

6.8m*2000=13600m/10ping=1360m

 

Maybe problem is somewhere here:

2sec = 2000ms - why this 2 sec??? What happens in that 2sec and what causes it??? (in this case)

[TAW_Blaze]

I don't know.. these numbers don't add up. There is another scenario where the missile passes me at M1.5 and explodes 1.4 nm behind which doesn't work with the math you're presenting.

 

It also doesn't explain how the missiles in the 2 perspectives are flying a completely different flight path both in terms of altitude and angle tracking. The side effect of these 2 is also a major speed difference usually resulting in the target seeing the missile fly past him and then explode with a delay.

 

I think the first question is whether the shooter's data is taken as a ground truth and synced to the server? Currently this is what it looks like.

[BlackPixxel]

It is interesting that the missile reaches the manouvering target on the targets tacview earlier and with higher speed than on the shooters tacview, even though it does not do a loft for the target. But since the target sees the missile guiding towards it the whole time and not just when it goes pitbull it can do a smoother turn and not loose as much energy (more direct path + less energy wasted for snapping to new direction once pitbull).

If the target would fly straight towards the launch position, the missile would be slower for the target and further away, while for the shooter it would have already reached the target thanks to the more efficient lofted path.

 

I hope ED fixed this as quick as they can.

[Falcon_S]

@Blaze

I do not know, I just guess...

 

how inaccuracies work

https://www.pcgamer.com/netcode-explained/

 

This guy knows what he's talking about:

Interesting explanation at 6:11.

Edited March 28, 2019 by Falcon_S

[mrkline]

If it was only due to latency, you wouldn't see completely different flight paths from the shooter's & everyone else's perspective. That shouldn't have anything to do with whatever lag compensation is being applied.

 

Hopefully this is addressed soon - it's a huge issue in Multiplayer air-to-air currently, as folks have to try to defend Phoenixes that are in completely different places than what their client is showing them.

[Schmidtfire]

At least the missile loft de-sync seems to be a known issue:

 

https://forums.eagle.ru/showthread.php?t=235489

[BlackPixxel]

Missile synchronisation is still not working.

 

This is a very big issue for multiplayer, as it gives aim120 and phoenix an even higher advantage. They are impossible to notch when they are not where the game shows you they are, and it frequently happens that you get killed by a missile that is dead and 2km away on your PC.

 

I do not understand how this can even be a thing in a game that is about air combat, it is very frustrating.

[GGTharos]

There are two events that are synced: the missile launch command, so that all clients can simulate the missile being launched, and impact.

 

The clients simulate the flight path without any synchronization, and that's why incongruities caused by latency add up.

[BlackPixxel]

Lofting is not synced, and loss of datalink guidance before the missile goes active is not synced. The missile tracks all the time for the target, creating a very different flight path.

 

The lack of loft causes the missile to be much slower and further away from the target, and also causes a different angle to the target if it starts maneuvering. If the target just flies straight, it will get hit by a missile that is slow and dead on his side, with RWR showing that it is still far away.

 

For a "simulation" this so wrong, but also from a gameplay perspective!

And of course the issue only benefits AMRAAMs and phoenixes, making them even more powerfull.

 

ALL major events during the life of the missile have to be synced, and there should be position and vector updates in slow intervalls to make sure that the error in missile flight between shooter and target is as small as possible.

 

To notch a missile in multiplayer, does the position of the target to the missile on the shooter's PC matter, or is notching calculated on the targets PC?

[GGTharos]

Only TWO events are synced, launch command and impact. Nothing else.

 

Shooters missile is the ground truth, all other representations (ie all other client representations) are attempts at simulating that same shot.

 

Therefore you want to defeat whatever missile is on the shooters client, and your going that what you see on your client accurately represents that.

[BlackPixxel]

Yes, that's why I said all has to be synced, and even missile itself should be synced from time to time to remove any error.

 

If the shooters missile is the only thing that matters, does that mean that the shooter also calculates the notching?

[GGTharos]

Yep, that is correct.

[Big Nuts]

Interesting, is it unreasonable to expect this syncing method to change?

[GGTharos]

So far there had been no mention of changing this.

[Lurz]

Seems that's a known issue. Something weird happened to me last two days on multiplayer server. Three times I was blown without a touch of the missiles. They was just still fly.In that server there are only Fox 1 and Fox 2 missiles so obviously this affects all types. Btw my ping is usually above 100.

 

Some screeshots and tackview track:

 

 

https://www.youtube.com/watch?v=J0cNKJIfMDM

 

 

 

 

[Coxy_99]

Seems that's a known issue. Something weird happened to me last two days on multiplayer server. Three times I was blown without a touch of the missiles. They was just still fly.In that server there are only Fox 1 and Fox 2 missiles so obviously this affects all types. Btw my ping is usually above 100.

 

Some screeshots and tackview track:

 

 

https://www.youtube.com/watch?v=J0cNKJIfMDM

 

 

 

 

 

 

happened to me also a few times you watch the missile fly past then 5 seconds later you explode.

[Lurz]

Aaand another one: https://www.youtube.com/watch?v=aMJODUsy-ak&feature=youtu.be

 

 

This is definitely very frustrating and demotivating me to fly online, but offline is a crap with all these stupid AIs :noexpression:

Edited July 27, 2019 by Lurz

[Mareno]

@NineLine @BIGNEWYthis seems to have gone under your radar. This is still a huge problem.

[Vähäkylä]

I don't think the missiles receive any mid-course syncing. THis is testable in multiplayer, as the missiles can be tens of nautical miles away from each other for 150nm+ cruise missile shots.

[Kosi]

Is there any hope of this ever getting fixed?
Competitive PvP is pretty much pointless at a high level if 50% of the missiles desync.

[Flappie]

There is no tracks attached to this thread.

Please provide short tracks, one from the server, another from a client.

Missile desync is often caused by latency. Tracks will help determine the cause of the issue.