The online racing simulator
An Investigation into Lag Accidents
After being involved in a few lag accidents in various LFS races I have many un-answered questions & thought it would be a good idea to look at lag accidents in a bit more detail (as lag is a bit of a grey area) to see if we can understand a more about what exactly happens when a player lags & re-appears/warps to another location.

From what I have learned previously many different outcomes take place when lag occurs:

1. from the laggers perspective others cars disappear from view but the race carries on as normal for them & they do not suffer or see any consequences of their lag.

2. other drivers in the race will get punted/crashed even launched into outer space as LFS cannot deal with the trajectory of an unknown/lagging car. LFS does not predict trajectory either.

3. when someone lags they dissapear from view & a timer xx.x replaces thier name tag as an indication of the amount of time they have lagged & the position they were @ when the lag started, when the timer reaches 20 seconds they are disconnected/removed from the race or virtual enviroment.

Firstly can anyone confirm or deny the above statements or more importantly have further information to add so that we may establish the truth about lag.

My questions evolve around the multiple outcomes & strange occurrences that happen before during & after lag accidents.

1. When someone lags the outcome or perspective of a race can be different on different computers. Why is this?
2. When someone lags LFS creates one or more parallel virtual racetracks where many different outcomes are played out. How many different outcomes occur & why?
3. Why does'nt the server produce the only one true outcome of what it see's/knows & make that happen on the virtual track?
4. Does the lagger ever get effected?
5. Does the green timer only start when someone lags or can someone lag/warp without the timer appearing?
6. Can LFS evolve to predict trajectory so that lag accidents maybe reduced?
#2 - garph
Quote from anttt69 :Firstly can anyone confirm or deny the above statements or more importantly have further information to add so that we may establish the truth about lag.

When you're done can you then move onto extablishing the truth about the Sasquatch?

....that all seems about right to me, nothing much more to add dunno what or why this info would help anyhting!?
you know what they say about sarcasm dont you?

This thread is to help us gain a greater understanding of lag in on-line gaming & more importantly LFS. We may even be a to reduce lag accidents which are a constant bug/annoyance within on-line racing.
#4 - garph
Quote from anttt69 :you know what they say about sarcasm dont you?

This thread is to help us gain a greater understanding of lag in on-line gaming & more importantly LFS. We may even be a to reduce lag accidents which are a constant bug/annoyance within on-line racing.

You didn't really ask any questions the first time round you kinda just made a statement which pretty much covered the discussion.....until you edited it.

I don't think you'll be able to reduce lag incidents because, well, lag just happens with out warning most of the time. The problem is the number of factors that go into cause lag, your connection, everyone else’s connection, the server, the number of users online (I mean members of the public using your service provider), is there someone in your area downloading loads of porn....the list is endless.

It might make it easier to resolve disputes between racers if you understand what lag looks like in game, but it's not always that obvious at the time that it's lag and it would be cleared up with a couple of replays anyway. As you've said before each person would see a totally different accident, one looking like being wrecked the other showing a normal race.

It would be best to just focus on what lag does to LFS (or what it looks like) rather that looking for a cure to reduce lag (cause thats kinda out of your control, except for making sure you aren't downloading stuff and things like that) and because of it's randomness it would make it almost impossible or at least very difficult to fix/predict.

Not trying to be a dick about it, it's just a very complecated issue, which I don't think we can do much about.
I understand we crossed wires a bit there as I was still busy typing . You're right it is beyond our control & many factors are involved but it would be nice to get a grasp on the situation. It may be possible for LFS to reduce lag crashes, if it is then it would benefit the entire community.
Quote from anttt69 :1. When someone lags the outcome or perspective of a race can be different on different computers. Why is this?

The client at each end predicts where each car is by calculating physics, based on the last received data from the server. This is by design to stop stutter as cross-internet connections are too slow to allow for any other eventuality.

Quote from anttt69 :2. When someone lags LFS creates one or more parallel virtual racetracks where many different outcomes are played out. How many different outcomes occur & why?

Not sure what you're on about. Think you mean the above.

Quote from anttt69 :3. Why does'nt the server produce the only one true outcome of what it see's/knows & make that happen on the virtual track?

The server itself isn't as authoritative as many of us would like. Its nothing more than a basic packet relay with some other stuff on top. Scawen has stated that it doesnt perform any physics calculations, not even basic ones.

Quote from anttt69 :4. Does the lagger ever get effected?

Yes.

Quote from anttt69 :5. Does the green timer only start when someone lags or can someone lag/warp without the timer appearing?

Lag can appear without. The timer signifies a significant loss of connection for longer than an amount of time. I don't know what the threshold is.

Quote from anttt69 :6. Can LFS evolve to predict trajectory so that lag accidents maybe reduced?

It already does, hence why when you race online the racers don't bounce from position to position.

It *might* be improved by server side calculations and ensuring that the server distributes whats going on and only corrects the "lagger", not the "lagee" (please note its been a long time since I poked at the LFS net code, so things may have changed).

anttt69 and whoever else maybe interested in joining this thread, may I suggest that you read up on even a little basic gaming network code and game traffic theory, in particular how quake did/does it is a good example and there are many, well written articles on it. It's not 100% relevant to LFS not it does give you a good understanding of whats going on and what might help avoid the same questions being asked over and over again.

It will also help to have a basic understanding between the differences of UDP and TCP.

I'm not trying to put anyone off, but if we're going to have this discussion then lets do it properly

Ultimately though, its Scawen's decision.
I've never been in a race, where the results were different on different machines....

Lag can't be helped, unless people leave their 3rd world countries
Yes, I Include myself , i live in a 3rd world country too.
Quote from BlueFlame :I've never been in a race, where the results were different on different machines....

I think you've misunderstood the "results". It was the outcome of the lag, which is then corrected, not the finishing order being refered to.

FGED GREDG RDFGDR GSFDG