Jump to content

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


1911 replies to this topic

#341 Tekadept

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,290 posts
  • LocationPerth, Australia

Posted 15 April 2014 - 11:24 PM

Thanks for the response, I've noticed lots of different approaches to lag compensation with all sorts of franken-hybrid approaches.

I have never looking into valve stuff but will have a good read, but I am familiar with Glenn Fiedlers blog, I learnt a lot from referencing his posts over the years :)

Edited by Tekadept, 15 April 2014 - 11:31 PM.


#342 SnagaDance

    Member

  • PipPipPipPipPipPipPipPip
  • Little Helper
  • Little Helper
  • 1,860 posts
  • LocationThe Netherlands

Posted 15 April 2014 - 11:56 PM

View PostEast Indy, on 15 April 2014 - 06:57 AM, said:

Any interest in offering players a secondary weight class choice if they prefer shorter queues? Or queue-length estimates to help players gauge what's in demand and what's gridlocked?


Any comment on this post Karl?

To me it seems this would help shorten wait times.

I'd even like a system where I can 'offer for service' up to 4 mechs (1 of each weight class max.). This way it could be far easier for the MM to get that 3/3/3/3 together. And when friends really want to drop in a particular configuration they just place a single mech in the 'drop pod'.

#343 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 16 April 2014 - 12:07 AM

View PostSnagaDance, on 15 April 2014 - 11:56 PM, said:


Any comment on this post Karl?

To me it seems this would help shorten wait times.

I'd even like a system where I can 'offer for service' up to 4 mechs (1 of each weight class max.). This way it could be far easier for the MM to get that 3/3/3/3 together. And when friends really want to drop in a particular configuration they just place a single mech in the 'drop pod'.


This one is a bit tricky, given that multiple weight classes, each with their own Elos, would essentially mean us queuing one user multiple times in the matchmaking queue. Given the manner in which the matchmaker is currently implemented (it's a highly threaded process), this would be difficult. If design asked for this, we would find a way to make it happen, to the extent of rewriting the matchmaker if required. The turnaround time on this feature would be high though, and design would definitely take this into account.

#344 GalaxyBluestar

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,748 posts
  • Location...

Posted 16 April 2014 - 12:51 AM

View PostDimento Graven, on 15 April 2014 - 04:53 PM, said:

Another quick question:

On the map splash screen as you're connecting, in the upper left hand corner there's a little upside down guy that occasionally pops up.

What is that?

View PostKarl Berg, on 15 April 2014 - 09:41 PM, said:


No idea? Is there a screen cap of this anywhere you can link to?


I think he's refering to this...
Posted Image

#345 Tekadept

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,290 posts
  • LocationPerth, Australia

Posted 16 April 2014 - 02:18 AM

View PostGalaxyBluestar, on 16 April 2014 - 12:51 AM, said:


I think he's refering to this...
Posted Image

easter egg? :)

#346 9erRed

    Member

  • PipPipPipPipPipPipPipPip
  • Overlord
  • 1,566 posts
  • LocationCanada

Posted 16 April 2014 - 04:34 AM

Greetings all,

For Karl: Has there been enough improvement in the missile hit detection code to now allow for new and additional weapons and warheads?

I'm looking at the Arrow IV systems (available in this timeline) and all the various types of munitions it could bring into the game. But this requires hit detection and splash damage to be well implemented for the system to function as intended. (sometimes TAG or NARC dependent, but can be dumb fired at a location.)

Side note: why are the current LRM's that are dumb fired not being attracted by an active NARC?
- Do we need to change ammo to NARC capable LRM's?
- They should be changing course and actively seeking the NARC beacon if they loose there initial target or were fired in anticipation of a NARC lock.
(the launcher unit "should not have to wait to acquire a physical lock on the target", that's the NARC's whole purpose, to pull the missiles in.)
- Even SRM's are attracted to NARC's and they don't have standard guidance.
Ref: NARC Pod - causing friendly SRM and LRM missiles (even those that do not traditionally have homing capabilities) to lock on to them.
http://www.sarna.net..._Missile_Beacon

- Or should that be directed at Paul or Matthew, could you pass that on if you have the time, thanks.

9erRed

Edited by 9erRed, 16 April 2014 - 04:55 AM.


#347 Klappspaten

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 1,211 posts

Posted 16 April 2014 - 05:03 AM

View PostKarl Berg, on 16 April 2014 - 12:07 AM, said:


This one is a bit tricky, given that multiple weight classes, each with their own Elos, would essentially mean us queuing one user multiple times in the matchmaking queue. Given the manner in which the matchmaker is currently implemented (it's a highly threaded process), this would be difficult. If design asked for this, we would find a way to make it happen, to the extent of rewriting the matchmaker if required. The turnaround time on this feature would be high though, and design would definitely take this into account.


I got a question regarding 3/3/3/3.
Will there be any way to display which weightclass will be needed at the given moment?
It would help to keep waitingtimes low if we could see which weightclass the matchmaker needs the most.

Fir example: If there was a display based on percentage.

Light: 24%

Medium: 22%

Heavy: 28%

Assault: 26%


The players then could chose Lights or Mediums, so they wouldn´t have to wait as long for a match.

Edited by Klappspaten, 16 April 2014 - 05:04 AM.


#348 East Indy

    Member

  • PipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 1,244 posts
  • LocationPacifica Training School, waiting for BakPhar shares to rise

Posted 16 April 2014 - 06:19 AM

View PostKarl Berg, on 16 April 2014 - 12:07 AM, said:

This one is a bit tricky, given that multiple weight classes, each with their own Elos, would essentially mean us queuing one user multiple times in the matchmaking queue. Given the manner in which the matchmaker is currently implemented (it's a highly threaded process), this would be difficult. If design asked for this, we would find a way to make it happen, to the extent of rewriting the matchmaker if required. The turnaround time on this feature would be high though, and design would definitely take this into account.

This makes sense from the can-it-be-done perspective, although the functional benefit is obvious, especially if there's concern over queue times.

And while there are many permutations for a group of four, a parallel queue for solo launches (84% of launches per your data) while probably process intensive, couldn't be as complicated. I don't know; I'd be a little surprised if design hadn't considered this.

(Oh, and thanks for reiterating, Snagadance!)

Edited by East Indy, 16 April 2014 - 06:20 AM.


#349 Mawai

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,495 posts

Posted 16 April 2014 - 06:28 AM

View PostKlappspaten, on 16 April 2014 - 05:03 AM, said:


I got a question regarding 3/3/3/3.
Will there be any way to display which weightclass will be needed at the given moment?
It would help to keep waitingtimes low if we could see which weightclass the matchmaker needs the most.

Fir example: If there was a display based on percentage.

Light: 24%

Medium: 22%

Heavy: 28%

Assault: 26%



The players then could chose Lights or Mediums, so they wouldn´t have to wait as long for a match.



My personal opinion is that if folks have to wait more than 2 minutes for a match then the match maker is broken. A match is typically 10 minutes and if the match making algorithm used doesn't adapt the 3/3/3/3 model in real time to the queues that are forming then wait times will become ridiculously long for the over-abundant mech classes ... and many folks will NOT have the choice to change weight class since they may only have one weight class of mech or may only enjoy playing one weight class.

The matchmaker should be designed so that the wright class composition is a specifiable parameter. It starts off at 3/3/3/3 as long as queue times are under 2 or 3 minutes for the worst case. When the queue time hits the limit it adjusts to
3/2/3/4 or similar where the most abundant mech requires more and the least abundant less. It then continues (adapting as needed) until queue times are less than 1 minute (or some lower bound) at which point it switches back to 3/3/3/3.

The key feature though is that whatever mech distribution ... it needs to generate matches with the same weight composition and close to total tonnage on both sides.

If the matchmaker can't adapt to changing queue sizes then there will be a lot of complaints somewhere down the road ...

#350 Klappspaten

    Member

  • PipPipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 1,211 posts

Posted 16 April 2014 - 06:50 AM

View PostMawai, on 16 April 2014 - 06:28 AM, said:

If the matchmaker can't adapt to changing queue sizes then there will be a lot of complaints somewhere down the road ...


These are the MW:O forums, there allways will be a lot of complaints about everything.

#351 Cimarb

    Member

  • PipPipPipPipPipPipPipPipPip
  • Caladbolg
  • Caladbolg
  • 3,912 posts
  • Twitter: Link
  • Twitch: Link
  • LocationA hop, skip and jump from Terra

Posted 16 April 2014 - 07:05 AM

View PostKarl Berg, on 15 April 2014 - 10:34 PM, said:


My apologies. I've been getting to these pretty late at night, and it's a little difficult to keep up with the volume of posts. :) I scanned back to find the post you were referring to, and I most certainly remember reading over your post before; with respect to changing the cool down and damage levels on AC weapons.

The impact on game performance is something that would have to be carefully tested, and would also be driven by design because of the obvious impacts on play experience, and the requirements for them to rebalance the modified weapons.

Similarly changing damage levels on AC's would, I assume, be a very tricky change to make.

I highly appreciate your input, so I don't mind a delay!

You aren't really able to address specifics, so I won't keep bothering you about it, but it would be extremely beneficial to the forums if someone that COULD address specifics was able to communicate like you have been. I know that's asking for a second miracle, but I have to at least try.

Any chance you could drag someone from the weapon balance side of things in here for us to torture...uh, interrogate...er, talk to?

View PostMawai, on 16 April 2014 - 06:28 AM, said:

My personal opinion is that if folks have to wait more than 2 minutes for a match then the match maker is broken. A match is typically 10 minutes and if the match making algorithm used doesn't adapt the 3/3/3/3 model in real time to the queues that are forming then wait times will become ridiculously long for the over-abundant mech classes ... and many folks will NOT have the choice to change weight class since they may only have one weight class of mech or may only enjoy playing one weight class.

The matchmaker should be designed so that the wright class composition is a specifiable parameter. It starts off at 3/3/3/3 as long as queue times are under 2 or 3 minutes for the worst case. When the queue time hits the limit it adjusts to
3/2/3/4 or similar where the most abundant mech requires more and the least abundant less. It then continues (adapting as needed) until queue times are less than 1 minute (or some lower bound) at which point it switches back to 3/3/3/3.

The key feature though is that whatever mech distribution ... it needs to generate matches with the same weight composition and close to total tonnage on both sides.

If the matchmaker can't adapt to changing queue sizes then there will be a lot of complaints somewhere down the road ...

I like this idea, but I think if you started having "adjustable" weight distributions it would allow people to game the system by all queuing as a certain weight class to basically force the matchmaker to adjust to meet their weights.

I honestly think the matchmaker would have immensely less work and be much more accurate if it just picked paired opponents one-for-one until the match was full. For instance, if I queue up solo in a 75-ton mech with a 1500 Elo, the matchmaker finds the closest weight and Elo combination to mine and then puts them on the other team. It then repeats this for each person currently in the queue, and launches when full. If a group is in queue, it does the same - matching the group as closely in numbers as possible.

Doing this would allow any number of groups, any size of group, and be as equal as possible given the population currently queued. The more extreme your group size and/or Elo, the longer your wait time, but if you give good, balanced matches, people won't mind longer waits in those cases.

The current matchmaker (if it's rumored existence is even true), regularly makes extremely lopsided matches, whether it is an Elo disparity, bad matchmaking, or whatever. That is actually the downside of Elo: as long as the roflstomps happen equally to both sides over time, it never corrects itself, but a roflstomp is actually the worst possible outcome for either team, IMO, regardless of which side of it I am on.

#352 DEN_Ninja

    Member

  • PipPipPipPipPipPipPipPip
  • The Blade
  • The Blade
  • 1,097 posts
  • LocationCrossing, Draconis March

Posted 16 April 2014 - 12:03 PM

View PostMawai, on 16 April 2014 - 06:28 AM, said:



My personal opinion is that if folks have to wait more than 2 minutes for a match then the match maker is broken. A match is typically 10 minutes and if the match making algorithm used doesn't adapt the 3/3/3/3 model in real time to the queues that are forming then wait times will become ridiculously long for the over-abundant mech classes ... and many folks will NOT have the choice to change weight class since they may only have one weight class of mech or may only enjoy playing one weight class.

The matchmaker should be designed so that the wright class composition is a specifiable parameter. It starts off at 3/3/3/3 as long as queue times are under 2 or 3 minutes for the worst case. When the queue time hits the limit it adjusts to
3/2/3/4 or similar where the most abundant mech requires more and the least abundant less. It then continues (adapting as needed) until queue times are less than 1 minute (or some lower bound) at which point it switches back to 3/3/3/3.

The key feature though is that whatever mech distribution ... it needs to generate matches with the same weight composition and close to total tonnage on both sides.

If the matchmaker can't adapt to changing queue sizes then there will be a lot of complaints somewhere down the road ...



One game for ya. Warthunder.



Edit: Yes. I've had legitimate queue times lasting a full 5-6 minutes. At least the taskbar icon flashes when I finally launch into a game while I browse the interwebs.

Edited by Tichorius Davion, 16 April 2014 - 12:06 PM.


#353 9erRed

    Member

  • PipPipPipPipPipPipPipPip
  • Overlord
  • 1,566 posts
  • LocationCanada

Posted 16 April 2014 - 01:50 PM

Greetings all,

Question for Paul:
Reference a disarmed Mech and it now being mission ineffective.

Quote

Q: It would not be unreasonable to consider a mech without weapons mission-killed for purposes of establishing a victory, of course. Since there is/was a global multiplier, that could be something interesting to try out on a test server weekend.

A: Interesting. Again something best brought up with Paul. I'll ping him and see if he has some time to stop by this thread. Obviously, we're still making fixes to HSR. Both with the patch that went out today, and we have another larger change coming on the 29th patch with Launch Module.


With the "soon" introduction of the Clans, and possible combat between strictly IS and Clan factions on some Planets. If there is a disarmed IS Mech near the match end, would the Clans (from Lore) not destroy the Mech but take it as a captured trophy and the Pilot as a "Bondsman". As the Lore and canon states the Clans detest wasting resources, and capturing "Bondsmen" was a common thing.

- And at one time the only way for any IS MechWarrior to get into the Clans.
- Testing and trials to earn his "rite" as a MechWarrior.
- Bound by honor, not shackles

Q: How will the match/game resolve "capturing" a defenceless Mech and ending the match.
Q: Will we have an option to "concede the battlefield" to the enemy and exit the area.
(common occurrence in Lore with the Clans if they deemed the loosing IS Pilots worthy)

Q: Will we eventually see a method to "eject" from a nearly destroyed Mech, and causing the Pilot to exit the Battlefield.
(possibly part of the discussed "respawn" mechanism, and a method to "re-enter" the battle in your next Mech from the Pilots DropShip)

Currently the only option we have in-game to end a mission/match is to destroy every Mech left, even if they are disarmed and without weapons. Seems rather wasteful as through out the canon storyline MechWarriors are a prized commodity and were commonly captured and ransomed back.
- suggestion for a specific "contract" on a mission, capture a Pilot from the Enemy team. Leading to nearly all the Enemy Companies destruction, identifying the Mech you need, then stripping down it's armament to the point of capturing it. Or separate from the herd and do the same thing. (difficult to do but pay's greatly)

Additionally if there is only one Enemy Mech left, and that Pilot runs away to hide and shut down, have an option for the Company Commander to "Offer to the Enemy Pilot" the "Concede the Field" to the victors option. Thus ending the match.

This would give you more options that just kill everything, as some missions may be to actually capture areas and locations, and any Mech's you can.

9erRed

Edited by 9erRed, 16 April 2014 - 03:33 PM.


#354 Cimarb

    Member

  • PipPipPipPipPipPipPipPipPip
  • Caladbolg
  • Caladbolg
  • 3,912 posts
  • Twitter: Link
  • Twitch: Link
  • LocationA hop, skip and jump from Terra

Posted 16 April 2014 - 01:57 PM

View Post9erRed, on 16 April 2014 - 01:50 PM, said:

Greetings all,

Question for Paul:
Reference a disarmed Mech and it now being mission ineffective

With the "soon" introduction of the Clans, and possible combat between strictly IS and Clan factions on some Planets. If there is a disarmed IS Mech near the match end, would the Clans (from Lore) not destroy the Mech but take it as a captured trophy and the Pilot as a "Bondsman". As the Lore and canon states the Clans detest wasting resources, and capturing "Bondsmen" was a common thing.

- And at one time the only way for any IS MechWarrior to get into the Clans.
- Testing and trials to earn his "rite" as a MechWarrior.
- Bound by honor, not shackles

Q: How will the match/game resolve "capturing" a defenceless Mech and ending the match.
Q: Will we have an option to "concede the battlefield" to the enemy and exit the area.
(common occurrence in Lore with the Clans if they deemed the loosing IS Pilots worthy)

Q: Will we eventually see a method to "eject" from a nearly destroyed Mech, and causing the Pilot to exit the Battlefield.
(possibly part of the discussed "respawn" mechanism, and a method to "re-enter" the battle in your next Mech from the Pilots DropShip)

Currently the only option we have in-game to end a mission/match is to destroy every Mech left, even if they are disarmed and without weapons. Seems rather wasteful as through out the canon storyline MechWarriors are a prized commodity and were commonly captured and ransomed back.
- suggestion for a specific "contract" on a mission, capture a Pilot from the Enemy team. Leading to nearly all the Enemy Companies destruction, identifying the Mech you need, then stripping down it's armament to the point of capturing it. Or separate from the heard and do the same thing. (difficult to do but pay's greatly)

Additionally if there is only one Enemy Mech left, and that Pilot runs away to hide and shut down, have an option for the Company Commander to "Offer to the Enemy Pilot" the "Concede the Field" to the victors option. Thus ending the match.

This would give you more options that just kill everything, as some missions may be to actually capture areas and locations, and any Mech's you can.

9erRed

A lot of that is just role playing, but having an option to concede the match would definitely be a good improvement.

#355 Killashnikov

    Member

  • PipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 187 posts
  • LocationSydney

Posted 16 April 2014 - 06:28 PM

Perhaps a better way would be to blur the edges of the weight classes. Allow light heavies like the dragon to count as EITHER heavy or medium - any mech on the edge of a weight class boundary can count in either class for matchmaking?

#356 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 16 April 2014 - 10:22 PM

View PostGalaxyBluestar, on 16 April 2014 - 12:51 AM, said:


I think he's refering to this...
Posted Image


I completely meant to follow up on that today, but it was so busy that this slipped my mind. It's certainly possible it's an easter egg of some sort.

Thanks for posting the image!

#357 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 16 April 2014 - 10:26 PM

View PostKarl Berg, on 16 April 2014 - 10:22 PM, said:



I completely meant to follow up on that today, but it was so busy that this slipped my mind. It's certainly possible it's an easter egg of some sort.

Thanks for posting the image!
BUT I MUST KNOW!!! IT IS DRIVING ME WHACKO BECAUSE IT DOESN'T SHOW UP WITH ANY SORT OF CONSISTENCY!!!

ARGGHGGHASDGSDFGHF!!!

Sorry... Raging debate...

Some people have said it indicates you are on an 'excrement' list and have had PGI set you up for additional 'difficulties' (HSR removed, 50% damage dealt reduction, in the case of any tie you automatically lose)...

Others have said it indicates the ELO has guestimated your team should lose...

But know one knows for sure.

I see it about 50% of the time I drop I think...

#358 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 16 April 2014 - 10:38 PM

I've followed up on the colour-blindness issue, and got some answers back. This is definitely a serious concern, and it is still in the product backlog. The reason for colour blind tweaks not being scheduled sooner comes down to highly constrained UI and QA resources. Taking UI engineer time to address this now would push back certain critical path features even further, and the need for added regression tests on QA to verify all colour blindness fixes would also increase pressure on a constrained portion of our development pipeline. So the intent is to revisit this post CW. Sorry for the delay :P

View PostDimento Graven, on 16 April 2014 - 10:26 PM, said:

I see it about 50% of the time I drop I think...


You mean it's not consistent? Curiouser and curiouser.

#359 Kageru Ikazuchi

    Member

  • PipPipPipPipPipPipPipPip
  • The Determined
  • The Determined
  • 1,190 posts

Posted 16 April 2014 - 10:39 PM

View PostDimento Graven, on 16 April 2014 - 10:26 PM, said:

BUT I MUST KNOW!!! IT IS DRIVING ME WHACKO BECAUSE IT DOESN'T SHOW UP WITH ANY SORT OF CONSISTENCY!!!

Weird ... shows up for me every time.

Used to be just a black square with a little grey dude doing a headspin ... now (like the above screenie) it looks like an indicator of some sort, half full of glowy blue/cyan.

#360 Karl Berg

    Technical Director

  • 497 posts
  • LocationVancouver

Posted 16 April 2014 - 10:47 PM

View Post9erRed, on 16 April 2014 - 04:34 AM, said:

For Karl: Has there been enough improvement in the missile hit detection code to now allow for new and additional weapons and warheads?

<snip>

- Or should that be directed at Paul or Matthew, could you pass that on if you have the time, thanks.

9erRed


Definitely a determination for design to make. Consider it passed on. :P





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users