Jump to content

Public Test - 24/jun/2014


212 replies to this topic

#201 Heffay

    Rum Runner

  • PipPipPipPipPipPipPipPipPipPip
  • The Referee
  • The Referee
  • 6,458 posts
  • LocationPHX

Posted 25 June 2014 - 09:26 AM

View PostMawai, on 25 June 2014 - 09:22 AM, said:


Actually, the total number of folks in the queue doesn't affect queue times ... it is only dependent on the distribution of mechs entering the queue.


At a certain level, it absolutely does. If there aren't enough groups to create matches, then the wait time increases. And the PTR is so low population that queue starvation is a real issue. Fortunately that's also a great way to test the expanding requirements, so at least they were able to look at that.

Mr Berg,

You mentioned there was a dashboard that allowed you to view the data. Would it be possible to see a screenshot of that? Would love to see the internal tools you use.

#202 Mawai

    Member

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

Posted 25 June 2014 - 09:43 AM

View PostHeffay, on 25 June 2014 - 09:07 AM, said:


Well, that would be a nice feature to add. Maybe even be able to fiddle in the mechlab or even drop in the testing grounds while you wait for the queue. Then again, since the constraints will open up as time goes on, the longest wait is probably still going to be under 5 minutes, so there needs to be a way to save your work in progress. It's all possible, but probably not a super high priority since queue times will be (relatively) low anyway.

But there is nothing inherently wrong with 4x3, and since the population will happily self balance, you end up with far more interesting matches where role warfare can actually thrive, instead of people just bringing the tonnage to be the hero. Which they rarely are. And since you have better matches, you have more happy people in all weight classes, retention goes up, people buy more, etc, etc.

There really is no downside to 4x3.


I think the fundamental assumption (possibly flaw) in this statement is "since the population will happily self balance".

There are a lot of folks who can play all the weight classes (I have mastered mechs in all of them) but I tend to like to play heavies and lights. I have a hero mech in each category. I also have lots of mechs to level up ... so I CAN play clan or IS in every weight class ... but most of the time I probably won't. At least some of the time I will want to play my heavy mechs because they are fun. Other people feel the same way about assaults ... so much so that they may only have assault mechs.

It would be a bad business move to penalize a substantial fraction of the player base (up to 70% appear to prefer playing heavies/assaults) with punishing queue times (if it works out that way) ... in order to increase the fraction of 3/3/3/3 matches generated.

This is where the selection of tuning values Karl mentioned comes into play.

What is a reasonable queue time limit before MM release valves kick in and break from the 3/3/3/3 frame work?
- I tend to prefer 2 minutes ... waiting longer than 2 minutes for a match that is over in 8 minutes on average would seem like too long to me. However. perhaps 5 minutes would be more of a deterrent ... BUT ... how many folks would choose not to play rather than wait 5 minutes for a match? 10 minutes? 20 minutes?

When the release valves kick in the MM will produce weight class balanced matches with a non-3/3/3/3 distribution which shortens the longest queue wait times. It will have to do this for some number of matches to reduce the queue times to some target value. This value needs to be less than the queue time limits where the release valves first kick in ... i.e. non-3/3/3/3 matches need to be run until the queues are under 1 minute for example ... then it can go back to 3/3/3/3 (or some similar algorithm). The hysteresis in the queue times is necessary or the matchmaker will not get a chance to generate any 3/3/3/3 before another queue hits the time limit and release valves kick in again.

In any case, I am looking forward to see how this works in practice. The community may be self-balancing or it might not ... my opinion is that the community will not be able to self-balance for any length of time ... maybe the first hour when everything is new ... but in the long run people will revert to playing what they like to play ... or not play at all if the queue times in that case are too long. The key is what value to set as the queue time limit ...

#203 Heffay

    Rum Runner

  • PipPipPipPipPipPipPipPipPipPip
  • The Referee
  • The Referee
  • 6,458 posts
  • LocationPHX

Posted 25 June 2014 - 09:48 AM

View PostMawai, on 25 June 2014 - 09:43 AM, said:


I think the fundamental assumption (possibly flaw) in this statement is "since the population will happily self balance".

There are a lot of folks who can play all the weight classes (I have mastered mechs in all of them) but I tend to like to play heavies and lights. I have a hero mech in each category. I also have lots of mechs to level up ... so I CAN play clan or IS in every weight class ... but most of the time I probably won't. At least some of the time I will want to play my heavy mechs because they are fun. Other people feel the same way about assaults ... so much so that they may only have assault mechs.

It would be a bad business move to penalize a substantial fraction of the player base (up to 70% appear to prefer playing heavies/assaults) with punishing queue times (if it works out that way) ... in order to increase the fraction of 3/3/3/3 matches generated.

This is where the selection of tuning values Karl mentioned comes into play.

What is a reasonable queue time limit before MM release valves kick in and break from the 3/3/3/3 frame work?
- I tend to prefer 2 minutes ... waiting longer than 2 minutes for a match that is over in 8 minutes on average would seem like too long to me. However. perhaps 5 minutes would be more of a deterrent ... BUT ... how many folks would choose not to play rather than wait 5 minutes for a match? 10 minutes? 20 minutes?

When the release valves kick in the MM will produce weight class balanced matches with a non-3/3/3/3 distribution which shortens the longest queue wait times. It will have to do this for some number of matches to reduce the queue times to some target value. This value needs to be less than the queue time limits where the release valves first kick in ... i.e. non-3/3/3/3 matches need to be run until the queues are under 1 minute for example ... then it can go back to 3/3/3/3 (or some similar algorithm). The hysteresis in the queue times is necessary or the matchmaker will not get a chance to generate any 3/3/3/3 before another queue hits the time limit and release valves kick in again.

In any case, I am looking forward to see how this works in practice. The community may be self-balancing or it might not ... my opinion is that the community will not be able to self-balance for any length of time ... maybe the first hour when everything is new ... but in the long run people will revert to playing what they like to play ... or not play at all if the queue times in that case are too long. The key is what value to set as the queue time limit ...


People were self balancing even when there was *no* impact to the queue time. Personally I'd like if if they put the average wait time for the last 12 players or something like that, so you can judge exactly how badly you want to wait until the relief valves kick in. If they are set to 3 minutes, you can either choose to do that, or insta-queue in your Hunchie. But it's the kind of thing that should be more transparent, as having more information is never a bad thing.

#204 Mechteric

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 7,308 posts
  • LocationRTP, NC

Posted 25 June 2014 - 09:52 AM

I liked the new sorting abilities in the UI, alphabetical is way better than whatever randomness it was before.


I still really REALLY want a proper list view of mechs and equipment, I really hate grid style lists because its much slower for me to find what I'm looking for.

#205 Wispsy

    Member

  • PipPipPipPipPipPipPipPipPip
  • Talon
  • Talon
  • 2,007 posts

Posted 25 June 2014 - 12:00 PM

View PostHeffay, on 25 June 2014 - 09:48 AM, said:


People were self balancing even when there was *no* impact to the queue time. Personally I'd like if if they put the average wait time for the last 12 players or something like that, so you can judge exactly how badly you want to wait until the relief valves kick in. If they are set to 3 minutes, you can either choose to do that, or insta-queue in your Hunchie. But it's the kind of thing that should be more transparent, as having more information is never a bad thing.


If people are self balancing now then why is the amount of heavies/assaults so much larger then lights/meds...

Maybe that is a surprisingly small part of the population...

Edited by Wispsy, 25 June 2014 - 12:00 PM.


#206 Bilbo

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Nimble
  • The Nimble
  • 7,864 posts
  • LocationSaline, Michigan

Posted 25 June 2014 - 12:09 PM

View PostWispsy, on 25 June 2014 - 12:00 PM, said:



If people are self balancing now then why is the amount of heavies/assaults so much larger then lights/meds...

Maybe that is a surprisingly small part of the population...

Because weight class has little to no effect on wait times at the moment and everybody just got their brand new Timberwolves and Direwolves.

#207 East Indy

    Member

  • PipPipPipPipPipPipPipPip
  • The Hammer
  • The Hammer
  • 1,242 posts
  • LocationPacifica Training School, waiting for BakPhar shares to rise

Posted 25 June 2014 - 12:15 PM

View PostMawai, on 25 June 2014 - 09:22 AM, said:

Actually, the total number of folks in the queue doesn't affect queue times ... it is only dependent on the distribution of mechs entering the queue.

It does affect them when there are literally no players available, which appears to have been the problem last night. It's also not the case when other factors like group number are being reconciled. (On that note, trying to accommodate large groups should probably be introduced after 3/3/3/3 has been operating with unlimited groups of 2-4 and solo players.)

#208 Mawai

    Member

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

Posted 25 June 2014 - 02:39 PM

View PostEast Indy, on 25 June 2014 - 12:15 PM, said:

It does affect them when there are literally no players available, which appears to have been the problem last night. It's also not the case when other factors like group number are being reconciled. (On that note, trying to accommodate large groups should probably be introduced after 3/3/3/3 has been operating with unlimited groups of 2-4 and solo players.)


Actually, I think they specifically introduced the group queue BECAUSE they found that the 1 group/side rule resulted in huge group queues. I think that (at least at certain times of the day) there are far more groups than they originally stated in Paul's Command Chair post ... as a result, they had to introduce a group queue to handle the overflow from the 1 x 2-4 person group/side limit in the solo queue. The only way to accommodate group overflow was with a group queue ... and as long as you are implementing that it might as well handle larger group sizes. Thus ... this gives the design they have implemented in the current matchmaker.

Pure speculation mind you ... but it makes sense to me and this is why you can't have 3/3/3/3 1 group/side without also having a group queue to handle the rest of the groups. (The algorithm for balancing these group queue matches by groups AND weight class must be brutal ... also, I would expect that to keep the group queue times in check they will probably have to relax the constraints fairly frequently ...)

#209 StainlessSR

    Member

  • PipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 443 posts
  • LocationSunShine State

Posted 25 June 2014 - 02:47 PM

View PostKarl Berg, on 24 June 2014 - 11:32 PM, said:


Yes exactly. I suggested they keep the urban in reserve as a PTS only mech to draw users. I was laughed at ;)


Incentiveizing playing on the test server during tests would definitly increase the response, however the reward should only be offered after 1-2 hours of playing or a specific number of drops say 4-8 with no disconnection (or afk) before match result screen allowed (or the round does not count). Also, the reward should only be available by playing during a test (kinda like warframe does then maybe offer it a month or two later as a buyable item), other than a mechbay there isn't anything that is currently ingame that would incentivize me to play on the test server (altho I do anyway when I can).


View PostMawai, on 25 June 2014 - 09:22 AM, said:

***snip***
1, Lowest fraction in queue less than 25% = 10% cbill bonus
2. Second lowest fraction in queue less than 25% = 5% cbill bonus
3. Third lowest fraction in queue less than 25% = 2.5% cbill bonus

Could also look at xp bonus but that would be much less of a motivator for the player base to switch mech classes since it is of greatest interest only when leveling up a mech.

I honestly would think that
Spoiler
would work better as you can have two queue under 25% (usually light and med)


[Rant]
Until pinpoint damage is removed from the game and weapon spread is put in (not weapon spread ala clan where some of the damage if transfered to adjacent parts) where if you fire 4 ac2's two may hit where you aimed {center torso} but the other two would hit {left torso} and {right leg} and the pinpoint modules/skills just increase the chance of all weapons hitting the same place (max of 50% for each weapon which are checked seperatly) all that fall damage and heat nerfing and charging time is going to accomplish nothing in the end, the game will still revolve around high pinpoint alpha for instakills.
[/Rant]

TLDR: remove pinpoint damage and get rid of the NON fixes you have tried the latest being fall damage


Edit: removed impolite unnessesary word

Edited by StainlessSR, 25 June 2014 - 02:50 PM.


#210 BLOOD WOLF

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Jaws
  • The Jaws
  • 6,368 posts
  • Locationnowhere

Posted 25 June 2014 - 03:29 PM

Pinpoint="absolutely precise; to the finest degree." Any weapon can deal pinpoint damage, it depends on the pilot ability to use his weapons.

#211 MonkeyCheese

    Member

  • PipPipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 3,045 posts
  • LocationBrisbane Australia

Posted 25 June 2014 - 04:18 PM

View PostMawai, on 25 June 2014 - 09:22 AM, said:

The aspect they haven't addressed yet is the carrot. I would suggest something like:
If the mech proportion in the queue is less than 25% then there will be a cbill bonus for choosing to play that class of mech.

1, Lowest fraction in queue less than 25% = 10% cbill bonus
2. Second lowest fraction in queue less than 25% = 5% cbill bonus
3. Third lowest fraction in queue less than 25% = 2.5% cbill bonus

Could also look at xp bonus but that would be much less of a motivator for the player base to switch mech classes since it is of greatest interest only when leveling up a mech.


Or you know things like proper role warfare and bonuses for doing such things for lights and mediums.

Quote

Role Warfare is a term used for applying in-game mechanics and features, which assist a player in optimizing the game for their style of gameplay.

There are 4 basic types of gameplay roles in any team based game. These are the common roles found in military conflicts between two or more forces, and are similar to a chess game in terms of how they play out on the battlefield. The four roles are:

Scouting – Gathering information as a reconnaissance/stealth unit that gets relayed back to the rest of the friendly force.
Defense – The defender holds the ground gained by the offensive forces and protects those in need.
Assault – The assault role is for the tactical forward units whose primary role is to seek and destroy.
Command – The command role is split into two different levels, lance and company. Lance command falls to those in charge of up to 3 other players and assist in coordinated attacks on key targets. The company commander utilizes the 2 other lance commanders and all information being relayed back to him to make global calls on the battlefield.

ROLE AGGREGATION
Role Aggregation combines all of the BattleMech and Pilot Skill customizations to create customized gameplay that the player has chosen to follow.

It all begins in the game’s cycle of playing matches. The player first starts off by selecting contracts or entering quick matches and playing the game in general.


As players play through matches, they earn XP for their BattleMechs. XP is gained by performing various tasks within a match and bonus XP can be gained by being successful at the end of the match. Bonus XP can be gained from the meta-game which can give XP based on faction/allegiance to a House.


After each match XP is stored in the last BattleMech used. This pool of XP can be spent on BattleMech efficiencies. Each BattleMech chassis has its own pool of XP and custom Tech Tree. At certain branch points in the BattleMech Tech Tree, a Pilot Point is awarded to the player.

Each branch under the BattleMech’s Tech Tree represents attributes that fall into the Role Categories described earlier on.


Pilot Points that are unlocked in the BattleMech Tech Tree are then assigned to the player’s pilot avatar (MechWarrior).


Now the player can use the Pilot Points to unlock pilot abilities in the Pilot Skill tree. As abilities are purchased, Modules are unlocked in the store.

Again, the branches in the Pilot Skill Tree represent attributes that fall into the Role Categories described earlier on.


After purchasing modules from the store, using soft currency (C-Bills), the player can equip them to the BattleMech of their choice. The maximum number of Modules that can be equipped is determined by the BattleMech chassis.



It is this game cycle that brings customized gameplay to the player experience. The combination of BattleMech Modules and Pilot Abilities will dramatically augment the play style of the player.

ROLES PLAYING OUT ON THE BATTLEFIELD
How the roles play out on the battlefield relies on the players’ activity. If a player customizes their gameplay to be a scout, then they should play the role of a scout and relay information back to the command players.

Scouts
Scouts are the main source of information on the battlefield. It is their utilization of Information Warfare that is the key to the success of a team. Scouts need to get to the front lines as soon as possible in order to gather information as fast as possible. They should utilize fast moving BattleMech which allow them to do so and at the same time allow them the opportunity to escape should the need arise. Advanced scout players will need to use scout Modules to enhance their abilities to detect enemies and relay information.

Suggested BattleMech Class: Light/Medium
Suggested Modules: Radar and Detection enhancements
Suggested Pilot Skills: Scout class skills
Defense/Assault
The defense and assault roles are very similar in their core gameplay. The only difference between the two is how they want to balance their skills and modules. There are numerous configurations that will prove to be beneficial to either role being played.

Defense may want to skew their attributes to heavier armor and friendly assistance modules, while the assault may want to skew their attributes to targeting and assistance modules. Both roles need to be combat effective in both long and short range abilities.

Suggested BattleMech Class: Medium/Heavy/Assault
Suggested Modules: (Defense) Assistance and Detection enhancements, (Assault) Assistance and Targeting enhancements
Suggested Pilot Skills: (Defense) Assistance of friendly units and Detection/Protection skills, (Assault) Enhanced Targeting and Assistance of friendly units
Command
Command players are a special breed that will need to be "situation aware" of everything happening on the battlefield at any given time. It is up to the command player to let all friendly units know of key operational changes in plans or actions to be played out while in combat. The command player is always at risk due to the nature of the role and their dependency on the Battle Grid to issue orders and distribute information. It is because of this, command players should be protected by other players on the same team. Movement co-ordination is the key to the survival of the command player by making sure the command player is never left alone.

It is important to note that radar/information sharing has a defined range. This means that a command player cannot simply stand at their original spawn location for the entirety of the match. They must keep moving in order to make sure all members of the team are within communications range.

Suggested BattleMech Class: Light/Medium/Heavy
Suggested Modules: Ability to call upon support units/Radar enhancements
Suggested Pilot Skills: Self-defense and Information sharing skills

THE CULMINATION OF ROLE AND INFORMATION WARFARE
There is a heavy cross over between the technologies involved with Information Warfare and Role Warfare. Role Warfare is more based on the execution of the utilization of Information Warfare equipment. It is this cross over, as a whole, which defines the new gameplay features of MechWarrior® Online™ that has not been seen in previous iterations of the franchise.

Players are now truly allowed to customize their gameplay experience to suit their play style which in itself is fairly new to on-line FPS/RPG titles.

Great care is being taken to ensure that the new systems are designed in a way that make sense for gameplay as well as to not step on the toes of the BattleTech™ Universe.


from here http://mwomercs.com/...-3-role-warfare

But sadly things like tag and narc still seem to give out really low rewards as an incentive to lose an energy or missile weapon outside of using them on your own lrm boat.

and the command console has nothing to do with commanding.

Plz fix pgi

Edited by MonkeyCheese, 25 June 2014 - 08:07 PM.


#212 ReXX

    Member

  • Pip
  • Bad Company
  • Bad Company
  • 13 posts
  • LocationMissouri

Posted 28 June 2014 - 06:29 AM

I strongly suggest permitting a little bit of wiggle room with the 3/3/3/3 set up. Many people like to roll the same weight class when teaming up with friends. A Light Swarm for hunting lone assaults. A medium team for snagging capture points or intercepting scouts. And so on and so forth. The concept of the "Steiner Recon Lance" is also something I've always loved. While I have no real qualms with forcing us to mix weight classes to keep the feel of the universe, where the factions traditionally have a wide variety of weight classes in any one force... I think strict 3/3/3/3 is just too much.

So, again, I'd appreciate a little wiggle room. Perhaps something that allows for a small variance, where one team can have only 2 of one class and 4 of another...but not more than that, or we'll just end up with 2 lights, 2 mediums, 4 heavies, 4 assaults.

The system should pay attention to tonnage to help balance that out. If one group brings 4 timber wolves, they may find they lack large assaults while the other team has a Daishi and an Atlas.

This system should not provide wiggle room for solo people who are just PUGing it. Only for four man lance teams. This is to lower wait time competition for those who are actually organized and will be a larger benefit to their team than four random hunchbacks with no coordination.

This is my suggestion for how to keep weight classes varied, but allow for those cool, uniform teams people like to roll with on occasion. Not that it's something I'd want to do constantly, but the imagery of it is cool and terrifying.





5 user(s) are reading this topic

0 members, 5 guests, 0 anonymous users