53 replies to this topic

[Blackfang]

Genewen, on 13 June 2013 - 02:02 PM, said:

Karl talked about ping increase and HSR, but not about hit registration. So maybe you should have "bothered to read earlier messages in this thread" instead of just running your mouth.

With all due respect, that is what HSR(hit state rewind) is, it was a system implemented for hit registration, so our weapons actually hit what we were shooting at.

[Genewen]

Blackfang, on 13 June 2013 - 02:07 PM, said:

With all due respect, that is what HSR(hit state rewind) is, it was a system implemented for hit registration, so our weapons actually hit what we were shooting at.

Not exactly correct. Dead backseat riders see the server version of the game, still they see shots hit but not register. This dictates that there have to be issues aside from HSR. Karl did not say a word about that. Also, HSR not exactly applies to a close-up shut-down atlas with his broad front towards you. Ping does not matter the slightest bit in such a case and still shots don't register.

PS:
HSR = host state rewind

Edited by Genewen, 13 June 2013 - 03:26 PM.

[Hammerhai]

I can confirm that at least in Sept 2012 there were definitely hit reg issues. Maybe this has resurfaced?

[BlackBeltJones]

Karl Berg, on 13 June 2013 - 01:24 PM, said:

To be honest, I haven't actually done any deep investigations into the impact of the recent networking issues on HSR directly. I can make educated assumptions based upon what I know so far alongside user reports; but that's very different from actually turning on HSR debug mode and doing the regression tests, under production loads, with multiple release revisions of the networking layer isolated from other patch changes, with multiple network configurations, to definitively isolate behavioural changes. I wouldn't feel comfortable putting out a formal command chair post on this topic without having had a chance to do at least this level of investigation.

I can assure you though that we are not finished with these systems. The patch to network behaviour I mentioned earlier is already in our stable branch and is undergoing QA testing. We will also be revisiting HSR, not simply to address any regressions, but to improve these systems beyond where they were before.

Thanks for the feedback and info. I was curious if you know what we are experiencing? Specifically (assuming you play the game) have you experienced the no hit registration, no damage, damage but no impact issues during your own gameplay? If we knew you could also relate to these bugs outside your support capacity it would greatly enhance the gravity of your input.
Thanks.

[Splitpin]

Thanks for the reply Karl. Part of the issue here is that when HSR was introduced it was so stunningly good that when things changed back to so-so, the contrast was marked. After ~15years of online gaming @250 ping from NZ I'm well used to lag/lead shooting so much so that I didn't even think something like HSR was possible, it was a revelation and I really really want it back

[TheMagician]

Karl Berg, on 13 June 2013 - 01:24 PM, said:

To be honest, I haven't actually done any deep investigations into the impact of the recent networking issues on HSR directly. I can make educated assumptions based upon what I know so far alongside user reports; but that's very different from actually turning on HSR debug mode and doing the regression tests, under production loads, with multiple release revisions of the networking layer isolated from other patch changes, with multiple network configurations, to definitively isolate behavioural changes. I wouldn't feel comfortable putting out a formal command chair post on this topic without having had a chance to do at least this level of investigation.

I can assure you though that we are not finished with these systems. The patch to network behaviour I mentioned earlier is already in our stable branch and is undergoing QA testing. We will also be revisiting HSR, not simply to address any regressions, but to improve these systems beyond where they were before.

I had better hit registration in closed beta, if that's any indication.

[Aslena]

thanks for letting us know your working on this

[Flashback37]

Yes, thanks for the response. I just hope that the issue is resolved quickly. I mainly run YLW and if the ac20 is borked much longer I might get a little.....annoyed.

[slide]

Thanks for the update Karl.

[Wildstreak]

Past two days, I have seen my ping anywhere between 100+ to 350+.
I am usually 30-40.
Send me a PM when its fixed, no point trying right now.
At least this explains (possibly) what I have had happen.

[TheFlyingScotsman]

Karl Berg, on 13 June 2013 - 11:07 AM, said:

Hey guys; yes, as you've noticed our issues with packet latency and loss aren't completely sorted, although the graphs indicate we came pretty close. With this very last patch we're now within about 10 to 20% of the total packet rate from about a month ago, except now players won't get disconnected due to low levels of packet loss or large latency variations. We have another fix coming with next patch to try and bring us all the way back in line, which should hopefully also solve the latency problems several users are noticing.

HSR, as it is right now, is extremely sensitive to varying latency, so for those of you experiencing large latency variations due to the increased send rate, this next patch should make things much better. Slightly longer term we are discussion options for removing the need to compute latency for HSR all together.

Confidence boost +1.

[GaussDragon]

Karl Berg, on 13 June 2013 - 11:07 AM, said:

Hey guys; yes, as you've noticed our issues with packet latency and loss aren't completely sorted, although the graphs indicate we came pretty close. With this very last patch we're now within about 10 to 20% of the total packet rate from about a month ago, except now players won't get disconnected due to low levels of packet loss or large latency variations. We have another fix coming with next patch to try and bring us all the way back in line, which should hopefully also solve the latency problems several users are noticing.

HSR, as it is right now, is extremely sensitive to varying latency, so for those of you experiencing large latency variations due to the increased send rate, this next patch should make things much better. Slightly longer term we are discussion options for removing the need to compute latency for HSR all together.

This is pretty good news, because Ravens and Spiders still shrug off a tremendous amount of damage from lasers (primarily).

[BIix]

Karl,

Thanks for the replies. Can you comment on point blank hit detection misses? More than a few times myself and others I play with notice that PPC and Gauss shots on point blank stationary targets can sometimes not register at all. Which is extremely frustrating.

Thanks.

[Valore]

GaussDragon, on 13 June 2013 - 06:25 PM, said:

This is pretty good news, because Ravens and Spiders still shrug off a tremendous amount of damage from lasers (primarily).

Oh god, don't remind me. We're back to invulnerable lights. We had 6 people trying to kill one spider, including 3 PPC stalkers.

He finally died to attrition from the combined streaks of around 3 of the guys. That was the most annoyed I've been in months.

Edited by Valore, 13 June 2013 - 06:56 PM.

[Arcturious]

I got a canned response from the support guys (which is odd as they are usually better than that) saying it was probably the fact I play from AU. On the plus side though they want to look into the issue so I'll take some screens / video, show some ping tests to Calvary area etc and send it all to them over the weekend.

To those though who don't think HSR is the main part of the problem. I'm not 100% on how PGI has implemented it but the issue will be around false positive filtering.

Basically, at some point it needs to decide which hit to trust. The server, or the client. This is why sometimes when spectating you might see a shot hit, but no damage done. On the clients side the game thinks he missed.

The game now needs to decide if you missed on your screen, but hit on the server. Why should it reward you for a hit, when on your end (as far as it knows) you actually missed.

HSR in means you now can't really lag shoot or lead targets anymore. As the HSR will rewind and find that you missed, and record it as a miss even if you had the correct leading for a hit on the server.

At the moment, there is some underlying issue that is causing your actual ping, and what the server thinks your ping is to be slightly different. I believe this is partly showing up in the form of inflated, fluctuating pings. So when that happens you are effectively out of luck. You shoot and hit on your screen. HSR checks what it thinks your ping is and records a miss. The server checks and because you shot directly at the enemy you also miss on the server.

I'm hoping this net improvement coming next patch brings everything back in line. As currently, you can't lag shoot or directly shoot, or hit a small target at all thanks to this disconnect.

[StoneMason]

I'm in Aus, i've noticed +50-100 ping spikes, worse than ever before. I'm getting between 40-60% of AC20 shots not registering at the moment too and lots of lights rubber banding.

Appreciate the response from the Dev's, it was fine last month and I'm sure it will be again.

[Bluemaxx]

This really is a very frustrating situation for all players involved right now.
I understand that the game is still in beta and being developed but the continued issues with the patches is disappointing.
The patches in march havs broken the game for me so i have been unable to play since then (extremely low FPS due to, what i believe anway, the increased draw distance in game) but i have still tried to be involved with my unit as much as possible.
Reading about how the lagshield has returned does make me questions if i should return to this game once i am able to play it again.
I am also, quiet frankly, surprised people still continue to play sometime when all we end up getting is frustrated at the situation we have been given. The only reason i see people returning is that this is the closest we've come to an actual MW game in over 10 years....if this was any other franchise i would have left a long time ago and felt very bitter about the whole experience.
Other games that are also still in beta are showing far better developemental cycles and stability right now and i fear that these games, as well as others also in developement, will drag people away from MWO and make this unviable as a MMO - possibly killing it forever.
I thank Karl for the quick response from the devs so we atleast have an idea what is happenning but i can't help but ask the question:
"What will be broken in the next patch?"

[Kiiyor]

Curccu, on 13 June 2013 - 11:24 AM, said:

Kinda hard not to realize something like this I would hope you guys would tell us about things like this without lots of people demanding fix asap or quitting game. Maybe even Command Chair post?

I daresay they were collecting data to find the cause of the problem, and correctly address it.

I know everyone is used to jumping to conclusions around here, but i'm happy for a little silence if it gets me a definitive answer.

[Curccu]

Kiiyor, on 14 June 2013 - 12:02 AM, said:

I daresay they were collecting data to find the cause of the problem, and correctly address it.

I know everyone is used to jumping to conclusions around here, but i'm happy for a little silence if it gets me a definitive answer.

Also just acknowledging the issue and telling us they are working on it could give us some peace of mind. Also at this point we wouldn't have to jump to conclusions.

Personally I'm very glad Karl did post, now I know that game will be playable again soon.

[Lucky 7]

Curccu, on 14 June 2013 - 12:38 AM, said:

Also just acknowledging the issue and telling us they are working on it could give us some peace of mind. Also at this point we wouldn't have to jump to conclusions.

Personally I'm very glad Karl did post, now I know that game will be playable again soon.

Again my Wolf Spider brother Curccu speaks wisdom.

Thanks for the reply Karl, letting us know that you work on the issue gives perspective which then gives a better mood while playing :-).