Jump to content

Paging Karl Berg...karl Berg, Please Pick Up The White Courtesy Phone...


1911 replies to this topic

#1781 Deathlike

    Member

  • PipPipPipPipPipPipPipPipPipPipPipPipPipPip
  • Littlest Helper
  • Littlest Helper
  • 29,240 posts
  • Location#NOToTaterBalance #BadBalanceOverlordIsBad

Posted 22 December 2014 - 12:58 PM

View PostWintersdark, on 21 December 2014 - 10:43 AM, said:

I'd like to see an official confirmation, but my theory (which has some grounding in research) is:

Matchmaker picks a territory for the match. The current status of that territory determines whether this is hold/attack/defend/counterattack.

So, then, if your side holds 6+ territories, you are statistically likely to get a hold mission rather than an attack. Likewise, if you're defending a world that the opposite faction has a 6+ hold on, you're statistically likely to get a counter attack.


I think there's a die roll involved somewhere to how the MM picks its territories. :(

#1782 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 22 December 2014 - 04:20 PM

View PostDeathlike, on 22 December 2014 - 12:58 PM, said:


I think there's a die roll involved somewhere to how the MM picks its territories. :(
I think so too. I have some totally unfounded theories related to which side initiates the MM instance, but you could just assume it picks a random territory. It doesn't impact my above post, though: your chances of attack/hold defend/counterattack are basically friendly vs enemy held territories.

#1783 Goose

    Member

  • PipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 3,463 posts
  • Twitch: Link
  • LocationThat flattop, up the well, overhead

Posted 22 December 2014 - 11:42 PM

This is gon'a sound funny, but "Is there now enough time between Big Pushes for Teh Devs to revisit some most of the Backburner List?" Seems to me we have ~ three dead skill unlocks, at the very lest, to look over …

#1784 Alaskan Nobody

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 10,358 posts
  • LocationAlaska!

Posted 22 December 2014 - 11:43 PM

View PostGoose, on 22 December 2014 - 11:42 PM, said:

three dead skill unlocks, at the very lest, to look over …

Three?

Other than Pinpoint what else does not work?

#1785 Goose

    Member

  • PipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 3,463 posts
  • Twitch: Link
  • LocationThat flattop, up the well, overhead

Posted 23 December 2014 - 01:31 PM

Arm Transverse, IIRC; I forget if the other heat-centric one does anything …

#1786 Void Angel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Marauder
  • The Marauder
  • 6,601 posts
  • LocationParanoiaville

Posted 23 December 2014 - 06:32 PM

I'm pretty sure Cool Run and Heat Containment both work - Smurfy says they do. Arm reflex, I couldn't tell you - arms move so fast, it's kinda hard to tell, so whether it works or not, it's still kinda less than useful.

#1787 Alaskan Nobody

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 10,358 posts
  • LocationAlaska!

Posted 25 December 2014 - 03:33 PM

View PostVoid Angel, on 23 December 2014 - 06:32 PM, said:

...Arm reflex, I couldn't tell you - arms move so fast, it's kinda hard to tell, so whether it works or not, it's still kinda less than useful.

Arm does - and is noticeable on my Assault mechs such as the Victor and Gargoyle

#1788 o0cipher0o

    Member

  • PipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 353 posts
  • LocationItaly

Posted 25 December 2014 - 04:13 PM

I'm sure cooldown is a placeholder skill. I remember some dev mentioning it a year or two ago, and surely i'm not getting any cooldown reduction upon unlocking it.

#1789 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 25 December 2014 - 06:20 PM

View PostShar Wolf, on 25 December 2014 - 03:33 PM, said:

Arm does - and is noticeable on my Assault mechs such as the Victor and Gargoyle

As do the twist increasing ones. Very noticable on heavy assault mechs.

And for that matter the braking and acceleration skills work too.

#1790 Void Angel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Marauder
  • The Marauder
  • 6,601 posts
  • LocationParanoiaville

Posted 25 December 2014 - 11:01 PM

Twist is certainly noticeable to me - Arm Reflex was not.

#1791 Arnold J Rimmer

    Member

  • PipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 892 posts

Posted 26 December 2014 - 04:08 AM

View Posto0cipher0o, on 25 December 2014 - 04:13 PM, said:

I'm sure cooldown is a placeholder skill. I remember some dev mentioning it a year or two ago, and surely i'm not getting any cooldown reduction upon unlocking it.

Cooldown is no longer a placeholder skill - but its 5% bonus is pretty paltry in the face of all the quirks we got. Pinpoint, however, IS still placeholder. What I wouldn't give for convergence to be a game mechanic again...

#1792 o0cipher0o

    Member

  • PipPipPipPipPipPip
  • FP Veteran - Beta 1
  • FP Veteran - Beta 1
  • 353 posts
  • LocationItaly

Posted 26 December 2014 - 10:02 AM

View PostArnold J Rimmer, on 26 December 2014 - 04:08 AM, said:

Cooldown is no longer a placeholder skill - but its 5% bonus is pretty paltry in the face of all the quirks we got. Pinpoint, however, IS still placeholder. What I wouldn't give for convergence to be a game mechanic again...



Oh, didn't notice the difference, haven't done timed tests in months, and that 5% is totally unnoticeable while playing... Well, good to have one less placeholder skill then!

#1793 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 10:41 AM

View Postp4r4g0n, on 23 November 2014 - 06:28 PM, said:


Thanks for the response Karl. I've actually been following been following posts on HSR (amongst a few other issues) quite closely and just wanted to clear up that little point. So if I can demonstrably show that no damage is being done despite pointing right at the target and being below the latency threshold, I should let you guys know, right?


Yes sir, please do so.

#1794 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 10:45 AM

View PostTheCaptainJZ, on 25 November 2014 - 08:11 PM, said:

Karl, back in the day when repair and rearm were introduced, a mech's weapons would do reduced damaged as it's health was lowered from damage. Is this still true or does a weapon with 1 health left still deal it's full damage? Or did it never work this way and I am remembering wrong?


I don't think it ever functioned in that specific manner. We stored (and still store) a health value for every single piece of equipment in your inventory. That health value was propagated into game as the internal health of the component in question. So if you brought a weapon with 0 health into game, you simply wouldn't be able to use that weapon. If you brought in a weapon with 1% health though, I'm fairly certain it would still hit for full damage.

#1795 Jody Von Jedi

    Member

  • PipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 1,551 posts
  • LocationNorth Carolina

Posted 26 December 2014 - 11:11 AM

He lives!

I thought this thread was dead.

JK, I know you all have been really busy with CW. Thanks for keeping the thread alive, Karl.

Jody

#1796 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 11:32 AM

View PostLi Song, on 26 November 2014 - 01:39 AM, said:

Hi again Karl!

I followed your link to the explanation of HSR which you gave earlier. And noticed this (emphasis mine):

Immediately my algorithm sense went off the charts. Something didn't sit right with not rewinding the shooter so I sat down and drew this up:
Posted Image


As far as my understanding goes, the server sends (some subset of) the authorative world state to the client periodically (say 30Hz). The client performs a prediction (extrapolation) from the last packet to give a smooth visual experience. And smoothly melds the positions when the next world state is received to avoid "popping" where the mechs snap to the correct position. HSR keeps a history of positions for each rewindable entity.

Consider the above diagram, two players run on a parallel course, their positions being perpendicular to each other and the course. Player A aims and shoots at Player B just before she is to run in behind cover. First of all, if only the target position is rewound, then in this case Player A's shot would miss, because the players would no longer be perpendicular for the sake of the hit registration. Second, if both players are rewound only by 1 ping, then the shot would hit the obstacle as the time it took for the servers true state to be transmitted to the client wasn't accounted for. This would support the observation that slow mechs can hit fast lights even with high ping but light dogfights suffer hit-reg issues as both the shooter and the shootee is moving fast. Especially supports the feeling I've had that PPC shots from a spider some times go through the target even though it was a clean hit on my screen.

So, to me it seems like the correct thing to do is to rewind all objects, including the player by one round trip time (RTT), i.e. 2*ping. This is because the player is reacting to content that is delayed by 1 ping and it takes 1 ping for the reaction to reach the server.


Nice diagram! :)

Were the system to function as you described, then yes there would be a problem. There is a critical point that allows all of this to function in a sensible manner. The way multiplayer algorithms work, in almost every twitch based FPS, is by allowing time to desync between the various agents.

(please excuse my programmer art)

Lets say we start on the server with the following state at T0:
Posted Image

The server is going to transmit P2's position to P1, and P1's position to P2. It takes ~ half of P1's ping for P1 to receive P2's position. Similarly it takes ~ half of P2's ping to receive P1's position. Assuming P1 has slightly lower ping than P2, here are some hypothetical client states after receiving their relevant world updates:
Posted Image

Posted Image

To keep things simple, lets consider P2 shooting directly at P1 given the above frame. P1 is occluded, so we expect the server to fire a trace at P1's position and hit the displayed obstacle.

Posted Image

So P2 sends to the server two pieces of info. His move input state for the specific net frame we're simulating, and the fire command, which includes a bunch of weapon activation info. Due to latency, the server isn't going to receive P2's packets for about another P2 ping / 2 seconds, so by the time the server receives P2's packets, this is what the server world state looks like at time T1:

Posted Image


We're going to hit the obstacle in question, but P1 is clearly out of position for this weapon trace, since he's continued sending movement updates to the server this entire time. Had the obstacle not been present, this would have been a clear miss.

However, P2 knows that he was shooting at P1's position from P2 ping seconds ago. In fact, there is an authoritative server timestamp associated with that position; so when P2 took the shot, he actually told the server 'I'm shooting against server time T0'.

So, the server rewinds the world prior to performing the trace:

Posted Image

And now the server view for this fire trace recreates the client state as seen by P2 when he took the shot locally on his machine.

Hopefully that clears up how things work a bit.

#1797 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 11:40 AM

View PostVoid Angel, on 26 November 2014 - 01:20 PM, said:

Hey, that HSR analysis suggests a related question to me:

Karl, I recall an explanation a while back as to why delayed convergence was taken out of the game - my understanding was that convergence created an unequal playing field because people with high latency weren't able to take shots as efficiently as those with lower ping. However, all this predated HSR, so my question is whether we can use HSR to support delayed convergence - and whether it's something you'd want to pursue if it were feasible.


It should be possible, but would likely come at the cost of increased bandwidth, more complicated hit-reg logic, and potentially increased CPU cost for additional weapon traces. If design came to us and asked for delayed convergence mechanics, I'd grab a couple other engineers and have a nice long sit down with design about all the various implications first, prior to committing to any work.

View PostHeffay, on 26 November 2014 - 06:12 PM, said:

And a quick follow-up question, if I may: Is it possible to have weapons in different areas have different convergence? Arm mounted weapons currently work as is (instant convergence), but torso/head weapons have a [fixed/user defined] convergence that they set prior to dropping?


Again possible yes, but the best case implementation would come at the cost of extra data needing to be transmitted, both upstream and down, and increased CPU pressure to simulate everything, both on the server and simulated proxies.

#1798 p4r4g0n

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,511 posts
  • LocationMalaysia

Posted 26 December 2014 - 11:40 AM

btw Karl,

Merry Christmas and a Happy New Year :)

#1799 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 11:55 AM

View PostMawai, on 27 November 2014 - 07:07 AM, said:

Hi Karl,
There are still quite a few anecdotal concerns about hit registration.

Some time ago, Russ posted that the server tick had become messed up which was causing hit registration issues but that problem should have been fixed.

I was wondering if you have server hardware monitoring to go along with the match data tracking. In particular, depending on your hardware, some machines can adjust clock speed with load or use features to boost clock rate on specific cpus (so called turbo boost features). There are also server characteristics like the server game tick rate. Do you track these in real time? Could there be a bug that would re-write the tick rate during a match which would lead to worse hit registration under unpredictable circumstances? How about the number of matches on each server? Does that affect game performance or hit registration?

I think we have all seen poor hit registration from time to time ... and I think it might be useful for "you" (meaning PGI) to record some key server state variables throughout matches to see if there are systemic issues that might cause it.


Hey Mawai,

We do have some monitoring in place, but nothing particularly fancy beyond resource utilization monitoring at the hardware level, and some customized telemetry datamining graphs at the application level. The tick rate issue was caused by an OS update that adjusted the default resolution of the windows scheduler, so that any blocking system call would cause that thread to sleep far longer than prior to the update. From what I can tell, Microsoft made this change to improve battery usage on Windows tablet devices, and decided it would be great to propagate that change to their server OS.

This was a fairly significant failure on our part. The knock-ons caused by the software update were non-fatal, but we failed to detect this issue prior to rollout, despite the fact that we had the capability to do so. During postmortem, both the resource utilization graphs showed evidence of the issue, as well as the application level graphs. We're paying significantly more attention to our lower level server metrics as a result of this failure.

View PostJman5, on 02 December 2014 - 08:41 AM, said:

Hey Karl, back in April you made a comment about the reasons for setting up the restrictive private lobby system. I think most of us are sympathetic to the real life costs it takes to run the servers, but at the same time there is a lot of concern that the current scheme is too restrictive. From a customer value point of view, he's paying $7 per month so that he can maybe possibly run a custom private lobby game if he can find someone else who is also paying $7 per month. In my mind that devalues the worth of premium time if you can't even ensure the feature you're paying for.

You guys might not realize this, but it also creates all sorts of headaches in the league scene when you have to:
A) Find two guys on opposing teams with premium time. (can be really tough in 1v1 and 4v4 leagues)
B ) ensure they are group leads (screw up here and you often have to reform the entire group/lobby)
C) Force them to stick around the entire series (might be several hours)

Finally, it would give experienced players a safe place to let their new friend explore the game against more than dummy mechs. This would improve the "new player experience".

I don't think it's unfair to say that reducing the restriction from 2 players to 1 player with premium time would be a huge boon for the playerbase.

I know you're not exactly in charge of this area, but your players would appreciate it if the team at PGI would reconsider this current onerous system for something that makes a little more sense. In my opinion this is one of those legacy decisions that really could use changing.


Hi JMan5,

I'll forward this on to Russ; thanks!

#1800 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 26 December 2014 - 11:59 AM

View PostTKSax, on 03 December 2014 - 07:50 AM, said:

Hey Karl,
Would it be possible to see how much mech variety has changed since the introduction of the Quirk System? I know I see a lot of mechs in the game I used to never see and also see more variety in mechs on my team that before,


It certainly is possible for us internally :) Russ would likely share these stats publicly as well if asked.





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users