I'm sorry but I do not follow you at all. Without antilag and 13ms ping you can aim LG using your crosshair, and antilag isn't supposed to change that in anyway.

So same result as I got.. about 5% increase.
Full story at http://www.quakeworld.nu/forum/viewtopic.php?id=4567&p=2

Even with 13ms the bolt is always a bit delayed in relation to the crosshair (the 13ms ? ), from what you say (or i understood), antilag makes it hit without delay so you dont need to compensate not even a bit to aim. If this is true then it is an issue.
cl_fakeshaft 1 makes bolt model stick always in the middle without delay, but it doesnt mean you hit always where you aim with the crosshair.

I've explained this a few times already, but here we go again: The bolt you see with cl_fakeshaft 0 is NOT an accurate presentation about how your shaft hits. This is because the players you see on your screen when you see the bolt have been predicted. So even without antilag when you have 13 ms ping you should aim with your crosshair. And with higher pinger without antilag you should aim somewhere between your crosshair and the beam, depending on your ping.

If my patches get applied then the antilag feature will NOT change how you shaft with 13 ms at all. The current implementation on the servers does change it slightly.

Just remove everything that has with "antilag" to do, forget you ever heard about it, problem solved.

Medar sure knows what he is talking about, but this so called "feature" is just a big joke.

One guy decides to change qw gameplay, nice done.

If you have any actual arguments for that opinion, I would be interested in hearing them.


I agree that Renzo's actions pushing this to servers without players knowing were really bad.

But like I have said a couple times: Properly implemented antilag shouldn't change the QW gameplay with 13 ms ping at all, and on higher pings it should only move it closer to the gameplay that we have on low pings.

As answer to your first quote, I bolded your own statement.

How is that not drastically changing gameplay when ping no more matters the way it did before?

Different feeling/gameplay on different pings is a part of the game, has always been and should always be.

To me it's not an important part of QW that gameplay with high ping sucks. It's an unfortunate reality.

To me the real QW gameplay is what you get when everyone has 13 ms ping.

When players have higher pings, we should do our best to provide as good gameplay experience as possible despite those circumstances.

Let me get this straight? If it would be possible to make everyone's game work just like they were on LAN without any side effects, you would STILL not do it? Just because it didn't USE to be like that? Wow.

Lets see that patch working. Players are saying they have 5% better LG so far, could be that "servers does change it slightly".


This is the main issue.


Fog, i disagree with you, Medar is right when he says "if not side effects" and all could have same gaming experience no matter what ping.

Let's get back to reality shall we?

Well I'll ask you then: Is there a way to do it without side effects? (The answer is ofcourse no due to the law of physics)

Matrix: Of course it would be nice if it was possible.

But that's not what you said. You didn't mention anything about side-effects or anything at all. You just said nothing should be changed.

So, let me ask you again: Why is this feature such a "joke"? If you would tell me your reasoning we might have a discussion about it.

If medar can iron out the side effect and it's sure that it's not easier to hit with higher ping than it is with low ping then I guess it's ok. We tested with sailorman and with 80ms I got 48%, with 25ms I got 45% and with antilag off I got 42%. But of course part of that can be just random variance.

Yeah sorry, I took for granted people would understand that if it was possible for everyone to "play with 13ms" with no side effects, I wouldn't have anything against it.
I would welcome it of course.

I'm just thinking like: A player with 13ms and a player with 100ms playing, only way to make it actually be like real 13ms vs 13ms is to do 100% accurate prediction of the guys, right? The higher the delay, the harder to do prediction? Wrong predictions is a really bad thing?

As far as I know, you can never get rid of the delay, it will always end up somewhere, if it so be movement, player prediction or whatever.

There could perhaps be an acceptable solution with a static max delay set, as of course, the lower the ping difference, the lower the in game difference in prediction/movement/etc. Haven't seen such a discussion here though.

No one has ever claimed such a thing to exists. The idea isn't to make it perfect 13 ms play of course, but to improve on the previous situation.


You can think of it like the delay is transformed from the shooter to the victim. Instead of the player shooting seeing "a wrong thing" when he is aiming his victim will see "a wrong thing" when he is getting hit. The reason why this is a good idea is because a small difference in where you need to shoot at makes for a lot trouble, but a small difference where you see the shot that hits (or misses) you will make considerably less trouble.

Also one thing to keep in mind, is that when the pings are reasonable and the player movement speeds are reasonable, the difference is the hitbox positions won't be very big.


Yes there might be a need for that, but I would like to first test my patches without such limitations.

Thing to keep in mind is that when you have a huge ping antilag will lead to weird things, but the alternative is totally horrible too. And the bigger the ping, the worse the disadvantage that player has based on other things too.

Are you just trying to keep this going? It was you who quoted me asking if I stated that I would be negative towards a solution with no side effects making it possible to play with "LAN-ping" no matter real ping. I have clearly stated: I would gladly accept such a solution.

I haven't said that such a thing exists nor claimed anyone else have said it either, I only answered your question where you stated ".. with no side effects..".

With the difference that playing with high ping and antilag off only affects yourself, while antilag on makes your ping affect your opponents gameplay. (Please, get my point on this one...)

Medar, could your patch be implemented on a server (or a single port) for some betatesting so we all could try it out and see what it feels like? That might clear things out a bit.

So when using antilag there is no such thing as fakeshaft 0, remove it and force it to 1... Also fix enemy bolt model to straight one. Althought it doesnt fix the problem that you cant see when lg is hitting you?

As well i think gl_shaftlight, alpha and possibility to change bolt colour should be disabled. Shaft is becoming too powerfull and it kills speed of the game so let's use the original bolt model

The current patch for the SVN version is available http://kirjasto.org/medar/tmp/antilag_predict_etc_2.diff, if someone wants to compile MVDSV and try it out. I do not have the resources to run the servers.

I think Renzo will look at the patch in the near future, so after that it might find it's way to some of his servers.

i wont change my config(i tried but some weird shit happened), and my whole aiming style sry, i want use my stupid bolt only but i also want to play some late night mixies, and im a shaftbased player. Anyway, why did u ban straightshaft, or shaftaplha, if you dont need the old bolt? I dont want to shit talking about ur attempt to help the high ping players, i just want to play my games in my way

Yep, we will take a look at it, and I'll update servers the moment something is committed.

Increase fakeshaft to a point where it is roughly equivalent to 13ms, as I do. This allows you to continue aiming using the bolt without having to compromise anything in the way of accuracy.

Thanks to Renzo my patch now in use on his servers:

quakeworld.fi, qw.intarweb.dk, troopers.fi and wargamez.dk

I would appreciate any feedback.
At least pangela.se still has the old version of antilag, in case you want to compare to that.
Many other servers like qw.foppa.dk do not have antilag enabled at all.

The changes compared to the previous implementation are what are listed in the earlier post (plus some technical details), but the 6.5ms value there is actually now 13ms and the 13.5ms value is now 20ms.

EDIT: Looks like there is an unfortunate bug that might cause some problems, it's getting fixed at the moment.

i increased it, but it changed my whole gameplay still

I think there must be some fundemental problem here.
The lag itself cannot be removed only moved.

So essentially you are taking the lag from the attacker's aim, and giving it to the victim's physics.
That's why it is still possible to shoot ppl around corners, and shaft infights with low ping vs high ping become very ... weird.

The attacker sees enemy on screen, fires, and server then checks "where was enemy xx ms ago? ok we move player back to there and see if it hits".
And enemy already around the corner thinking "i got away" now dies and goes "WTF?".
And this don't even have to be some insane 3000ms scenario blaze posted, this actually happend for javve in a game against fusion where javve was around the corner but still got owned by shaft. I will try and find the demo and see if its real or just a myth though =)

I don't know if u can somehow tweak this feature.
To me it seems that antilag gives ALOT of effect to the attacker and some effect to the victim. Maybe there is a factor there.

Maybe limit the antilag somehwhat or make it less effective, and it could elliminate the physics changes effected by this, or at least make them basically unnnoticable?

Also I think the effects in real world of antilag has greater consequences and effects than developers envisioned when creating this idea.
They did not think how it would affect ingame fights, physics etc.
You have to think that stuff that tampers with physics and aim is noticed much more in QW than any other game, since it's the fastest game around.

I've been kind of paranoid about this in games (a couple of times on voice I've commented that I've killed some guy through a wall, lol) but checking the demos it usually isn't actually the case. When you view mvds in slow motion it usually turns out OK. I specced some Fusion game once and in realtime it blatently looked like Skillah shafted a guy in dm3 water THROUGH the high bridge (he was backing away to avoid discharge) but in the demo you could maybe say the crosshair (viewheight -6) was just about on the edge of the bridge. It's certainly felt like you've been able to shaft people through the corners at dm3 sng tele though (even before antilag, it's probably just more extreme now).

The teleporter bug is very interesting. Sometimes I've got surprised and pressed fire BEFORE I die, but never got my shot off. I thought it was just down to getting old and slow reactions / lcd lag or something, but maybe not

Well I suppose without antilag in 'server reality' you could never kill around the corner. It could seem that way on the client side because of player movement prediction, but if you would check the demo that represents how the server sees things, you'd see that was not the case in reality. With antilag though, this is possible also on the server. Perhaps you could add checks, that first check if there is a line of sight at current server reality, and only if that is the case, then check if there was a hit "back in time".

from this old drama-thread.

At the moment you press +lg enemy is not behind the corner in server reality. when this messege reach server enemy may have taken cover. With antilag it seems qute honest that enemy get damaged.

On the other side if you have high ping you may have taken cover in your reality but after that you find youself dead and simply not reached that cover. And it is true both with and without antilag.

I hope there will be unlagged (13ms) beam made with antilag and cl_fakeshaft 0.