Is this what lag looks like?

[Flitgun]

Is this multiplayer lag? I've not seen it like this before in IL2. I'm guessing it is though. It could've been very bad for me if it wasn't for that friendly.

 

 

It's bizarre how the E7 speed normalises when on my 6. And then he dives out at about 1500 km/h with no engine. It must be a coincidence. Must be. I want it to be a coincidence.

 

 

[Legioneod]

Looks like lag but idk. Ever since the update multiplayer has been buggy for me.

[Solmyr]

I'd say no. To me a lag looks like the plane is frozen for something like 1 sec about, then suddenly get back to its "right" position. Here the guy seems to be accelerating like hell when he wants to. Careful, I don't want to pretend that he's a cheater, I don't know, let's see if the devs or the community can answer to you, but to me it looks pretty suspect. Maybe I'm misunderstanding something though.

[Mitthrawnuruodo]

Doesn't look like lag. The acceleration seems strange, but it's hard to conclude anything without knowing the speeds of both (or at least one) aircraft.

 

Tacview would be very useful in situations like this. 

Edited March 26, 2018 by Mitthrawnuruodo

[56RAF_Roblex]

Lag does not always mean standing still then jumping.  There is a type of lag called 'Rubberbanding' where a plane slows down for a second then speeds up for a second to get back to where it should have been.  Basically instead of getting your normal 60 frames in a second, you only get 30 of them over a second then 90 frames over the next second. You have now had the 120 frames over two seconds that you should have but you saw the target plane go 1/2 speed then double speed.

[Flitgun]

I've never seen it like this though, have you Roblex? When I see rubberbanding it's more like this: The lagging aircraft will suddenly fly DIRECT from one point to an updated position in a STRAIGHT LINE like a ufo beaming across the sky - perhaps the aircraft will travel sideways to get there ignoring maneuverability limits, and then continue in a normal fashion. I've never seen a lagger just move faster along a curving flight path and obeying aircraft maneuverability limits (seemingly at mach 2).

[Mauf]

What you people describe is simply client side interpolation at work.

The server sends regular updates on movement data for the planes around you. Think the common term is snaps or snapshots (that's the term used back with Q3 servers). You could call it the Server FPS, don't know what it's actual value is for BOS. Basically, it represents the server time resolution that your client can use to draw stuff around you. So what does a Client that recieves maybe 21 to 30 refreshes per second do when itself runs at 60 fps or more? Correct, it spreads the snaps out across it's own rate and guesstimates the inbetween values to smooth the movement across it's own rate. That's what client side interpolation is (and it's usually pretty good so long as the info keeps coming in without much jitter).

Now comes lag. Suddenly, the stream of snaps from the server stalls and the client is missing info (or the server is missing info from a client)... what does it do? That depends on how the client is programmed but from what I usually see in MP games, the client simply uses its last interpolation data for an object and continues extrapolating on what it last knew from the respective object (direction of travel, speed, speed change delta maybe and so forth). That's when they suddenly go in a straight line for a few seconds, sometimes at a weird angles (CloD is very obvious with that for example). Then, SNAP, packets come rolling in again, lag is over. Client has its last known location (it got from its own extrapolation) and uses that point to interpolate everything back to the new good position, resulting in these weird sharp angled warps we see.

That would be long timed lag spikes. If you get several smaller stutters, you would end up with the rubberbanding effect that Roblex describes.

Through all of this, there is no sudden overspeeding or massive G loads happening on the warping client. To them, everything was just as smooth as you would have expected it to be.

 

Client side interpolation is the worst form of lag compensation except for all those other forms that have been tried from time to time...

[Flitgun]

Interesting Mauf. I had thought the term rubberbanding referred to when a client appears to be bouncing back and fourth along it's path... like a 'twanged' rubber band. 

 

So, you're saying that the E7 flight path depicted here is mostly a guesstimated fiction; the server is just 'joining the dots' for my viewing pleasure. Is that right? 

[kendo]

No, not Lag.

 

Looks like LA5 to me....

 

 

 

[Mauf]

40 minutes ago, Flitgun said:

Interesting Mauf. I had thought the term rubberbanding referred to when a client appears to be bouncing back and fourth along it's path... like a 'twanged' rubber band. 

 

So, you're saying that the E7 flight path depicted here is mostly a guesstimated fiction; the server is just 'joining the dots' for my viewing pleasure. Is that right? 

 

Kinda. Not just your viewing pleasure but rather "What else can it do in this situation?" Also remember that what I describe above is rather simplistic and BOS might employ much more refined techniques to compensate lag. Also, you want these techniques in place. Quake 3 had the option to display everything without any sort of client side interpolation. It's a teleport fest. Interpolation deals very well with small scale lags and that's what you usually have to deal with. Huge, varying lagspikes are not the norm on your usual gaming capable connection

 

Looking at the footage, I guess BOS client doesn't interpolate acceleration. So what I guess happens is: He dives and accelerates (and he was diving a bit earlier than you see on your end). All the while, lag happens too in a short stuttery fashion (maybe his Windows Update decides to download stuff while playing or somesuch).

The plane dives and extrapolation just keeps it going on the same path but doesn't factor in acceleration. New info arrives, your client is informed that plane is actually ahead of extrapolated position, interpolation moves it ahead. Viewable effect on your end: ZOOOOM. Now this happens 3-4 times in short succession. Viewable effect on your end: ZOOOOOOMZOOOOOMZOOOMZOOOM, nicely smoothed over resulting in an impression that he zipped down pretty rapidly.

 

As to rubberbanding: Usually this happens when interpolation catches some quite funky changes in speed. Aka, your client guessed your opponent faster than it actually was and thus with the new data available, it has to be moved back from the further away extrapolated position to the new actual position from the server. Viewable effect: Other plane warping back, rubberband.

 

Since we're on topic of weird lag effets: As some might know, there is also the situation where you have an opponent shooting into blank space somewhere next or behind you and you getting hit. This is probably due to a technique called "Time Rewind" on the server. What this does is it saves your positions over a certain time window and then when an opponent shoots, it compares your past positions against his server reported latency to "guess" what the other client might have seen on his side when he took the shot. So in order to call a hit, it doesn't take your current position in server time, but rather your past positions within a time window up to the other players latency. As with interpolation, this method bears strange results when big latency numbers are involved on one end.

Edited March 26, 2018 by Mauf

[Flitgun]

Thanks again Mauf. I am somewhat relieved to hear this.

 

This is the most serious issue for me in this game - 'packet loss', 'lag', 'ping', etc. It ruins the game for me. I get blasted when I thought i was safe and I get ejected back into my living room when I am about to pull the trigger. 

[Guest deleted@134347]

yes, it's "lag" or packet loss, or something in relation to the network performance. The 'dotted line' flying as I call it when the enemy is sort of skipping over frames is one of them. Sometimes the plane freezes in one spot then transports 200-400 meters to another spot immediately.. etc..

[Solmyr]

40 minutes ago, moosya said:

yes, it's "lag" or packet loss, or something in relation to the network performance. The 'dotted line' flying as I call it when the enemy is sort of skipping over frames is one of them. Sometimes the plane freezes in one spot then transports 200-400 meters to another spot immediately.. etc..

 

As said earlier, but here the plane looks to "drive" normally until he accelerates abruptely to an insane speed, but in a "railroady" way, it doesn't look like it's teleported and showing you that it's teleported, it more looks like it actually flyies to me. Just my opinion, then I can't say if it still can be a network issue or not but looks weird.

Edited March 27, 2018 by Solmyr

[Livai]

A E-7 with Me-262 behavior....... 

[senseispcc]

On ‎26‎/‎03‎/‎2018 at 2:53 PM, kendo said:

No, not Lag.

 

Looks like LA5 to me....

 

 

 

It is a La 5 and not a LAGG with 2 G.

It is slow but is the speed of the game set correctly? Lag is more object going from one point to an other instantiniously then stopping in mid air for some time ...

[Flitgun]

6 hours ago, senseispcc said:

It is a La 5 and not a LAGG with 2 G.

It is slow but is the speed of the game set correctly? Lag is more object going from one point to an other instantiniously then stopping in mid air for some time ...

 

Hello senseipcc, three things:

 

I'm almost certain his Lag/La5 comment was stated with tongue in cheek, though I hope he doesn't give up his day job in pursuit of comedic professionalism.

 

In case anybody doesn't know me well enough to believe I might be dishonest, which is quite a lot of you I suppose, I left the recording bar there so you could see I hadn't tampered with the game speed.

 

I suggest you read Mauf's posts. It appears you are as ignorant on the main topic as I was and reading him may very well bring you up to speed.

 

[kendo]

4 minutes ago, Flitgun said:

 

....though I hope he doesn't give up his day job in pursuit of comedic professionalism.....

 

 

 

No chance ...