I'm not saying we've done the full comprehensive version of that, language barrier is also an issue, but I have never once seen a highly variable MTR to the final hop from his line (at least not while SQM is running). He often sees weird variability in intermediate hops but that's meaningless as long as the final hop responds well.
1 Like
I imagine many a gamer would even pay to have their privacy and security violated in that way if it meant they no longer got shot behind corners...
I'm waiting for the day this becomes widely available in consumer PCs (goofy Linus face not included):
You don't even need GPS, just NTP synchronization with a vendor pool of GPS synchronized servers would be fine. Apparently the PS5 gets its time by doing an HTTP request to a Sony server... pshaw what a joke. That might be accurate to 500ms if you're lucky.
1 Like
With GPS you do not need to rely on NTP working as expected, and you can get reliable timestamps in the microsecond range, while NTP is "only" good for milliseconds... Sounds theoretical, but if you want to base your game's causality on that data*, I would opt for micro over milli. (The bigger issue is that indoors, people will need external GPS antenna close to a window or they will have no time at all).
*) in all honesty, I am the only one proposing that 
Without having looked into the details, this sounds opague and round-about... then again it should be possible to exchange two packets back to back even over http.
exactly, as soon as you need a reliable GPS signal to play, you're screwed in my opinion. Getting millisecond accuracy from NTP would be a huge step up in what they're doing. The fact is, if your game sucks to play because of causality issues, it's 100% on your lackluster caring about your customers and the fact that gamers will pay MORE to try to fix things if your game sucks and hence the incentives are all broken.
2 Likes
Then again, given how loose some gamers play with the rules, I am not unhappy if gamers stay of the under-secured NTP infrastructure until that has been better secured ( Network Time Security (NTS), a secure version of NTP with TLS and AEAD is currently a proposed standard and documented in RFC 8915)
But yeah, other than that NTP would be helpful, especially if Sony/MS would start deploying a commensurate number of open NTP servers (but I might be biased here by my wish for that to happen for other reasons).
Well, that and strict causality sucks, as it requires you to not cur corners and will limit your ability to match players significantly. IMHO not a good excuse, but then I do not need to operate game servers and make a living from doing so.
There's an inherent problem in real causality. We're playing a game where we simulate everyone is a few tens of meters apart, but the soonest that something they do can be reflected in the world is 50-100ms due to network lag and jitter buffer, so in fact, the world acts as if they are 0.05 to 0.1 light-seconds apart = 30000km. If you play a space battle with ships 30000 km apart it will all feel natural. If you play a shooter with everyone in the same town square it will feel totally weird. if you play fifa football it will feel like you're running through thick syrup. I think the solution is LAN parties with everyone on the same gigabit switch
1 Like
Sure, but you can paper over a lot of this by speculating the local display into the future based on known game state, and revising/correcting that once differing information becomes available. Not perfect, but apparently pretty decent up to surprisingly large amount of milliseconds.
Theoretically true, but not a real issue, since our useful interactions top out way below 30000km (while out vision reaches far wider, when watching the stars), and our own internal processing is not all that fast so we can accept a few dozen milliseconds of delay, especially if the immediate sluggishness is papered over by local speculation. I am pretty sure that perceptually if one would introduce the corrections always with a forced switch of in-game "camera" view (like a cut in a movie) people would notice even less (but not a real option for FPS).
Maybe to those of us that have actual rel word experience in relativistic space battles
Yes, obviously shorter delay is better, but the challenge is getting a critical mass of players so that matchmaking will work for almost everybody almost all of the time, and to allow things like, playing with friends overseas. I guess what I want tp say is, there might be external reasons to accept less tight causality.
I think we're on the same page. We have to accept less causality, we must speculate into the future and then accept that some people will feel that they "were shot behind cover" because we speculated they would continue moving out of the cover and showed that on the other player's screens, and in some sense it's better to "piss off the one guy who lost" compared to pissing off "5 guys who clearly saw that player come out from cover". I think this is the source of a lot of complaints.
1 Like
I have the problem both on server and p2p, can a friend do an MTR on my public address??? is this useful?
Yes, it certainly would be helpful, especially if the same friend is playing p2p with you. In regards to server play it is less helpful as it would not show any congestion along the path to the server...
So one of the unlikely, but possible issues might be that during game play, the egress path from you to the server might suffer from too high delay/jitter, so that all the packets you send out with excellent timing (according to your packet captures) get somehow their timing massively degraded (or even their order changed) on passage to the server, and the only way to see such effects is by measuring along ideally exactly the same path as the game packets travel.
BUT that said, I really would appreciate if you could run the EA connection quality report and post screenshots? of the following parts:
- Connection Quality Info:
- Your Connection Score
- Your ISP Connection Score
- Your Match History
- ISP Comparison Leaderboard
Maybe EA has some diagnostics that might explain your problem in the game.
Here is a link to a french page for that tool:
Votre guide dâutilisation du Rapport qualitĂ© connexion EA
Also during game play what do the in-game connection indicators show?
no indicator lights up in the game except when I play in 4g
This is a bit underwhelming, I had hoped for at least a list of all (recent) matches with the connection statistics. BUT this seems to support your clean packet-captures as well and does not seem to point at network issues. And still I fully believe your reports of delayed and sluggish game play... Maybe you need to change your mindset and play like all of your players are big oil tankers that react slowly to steering and motor output and hence require massive predictive control, that is plan ahead an try to counter the delay by initiating moves much earlier....
Or just give up on this game and only play games that don't suffer from terrible programming
Yeah.... now I understand than one picks one's games not based on singular metrics, like quality of netcode, but it would at least be interesting to learn how @segal_72's experience would be with games with supposedly better netcode. Not having followed that field much in the last ~3 decades, I have no first hand experience to offer, but I have been told that Valorant seems to have a competent modern netcode; would be interesting to learn how well that plays for the OP...
How about a video to show the issue @segal_72? I mean just taken using your phone whilst playing?
2 Likes
Agreed it would be interesting from multiple perspectives. There is a good article on Valorant and it's netcode
I haven't been following all of this but am I right in thinking that so far it is only that @segal_72 feels there is lag? If so wouldn't it be nice to see the issue on a video?
1 Like
Yes, for sure... thinking this over a bit, it might be possible to (assuming his phone allows 120Hz allow motion recordings) show the game and the finger on the controller button to estimate the time between button press and reaction in the game... (it would need to be something easy to spot so that press and action can be unambiguously correlated).