Jump to content

Balancing The Alpha Strike With A Reactive Reticle


387 replies to this topic

Poll: Poll (348 member(s) have cast votes)

Do You Agree with the OP's Suggestion?

  1. Yes (276 votes [79.31%] - View)

    Percentage of vote: 79.31%

  2. No (60 votes [17.24%] - View)

    Percentage of vote: 17.24%

  3. Other (Explained in Post) (12 votes [3.45%] - View)

    Percentage of vote: 3.45%

Vote Guests cannot vote

#301 Otto Cannon

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 2,689 posts
  • LocationUK

Posted 13 July 2013 - 04:12 PM

View PostTank Boy Ken, on 13 July 2013 - 02:47 PM, said:

Also if you really suggest single fired weapons fire into the Center. Did you even think about what that would do? This will make using single BIG guns the best thing; So 2x Gauss or 2x Gauss and a PPC, all fired after each other will be the best with this solution.


Durrr... That's not a problem, genius. Light mechs use smaller weapons but heavies or assaults are supposed to use single big weapons where possible. You think a big mech is better suited to an AC20 or 28 small lasers?

The only halfway sane point was movement convergence. I think that would be better if it was more affected by target speed than your own movement. That way in your example with the sniper the tables would be turned, and lighter mechs would benefit most.

Talking about random cones just shows you didn't even bother to read the post before drooling nonsense over the thread. Hitting the same location repeatedly with single fire takes skill. Compensating for predictable, clearly marked non-random divergence would also require skill. The childishly simplified pinpoint we have now does not. It's a scrubcrutch because anyone can high alpha kill with a single click as easily as he clicked the icon to start the game. The game is not intended by PGI to have pinpoint instant convergence but we're stuck with it for now because of technical problems with hit detection.

At least Paul isn't a tryhard wannabe goon troll. Squawk to you too.

Edited by Otto Cannon, 13 July 2013 - 04:46 PM.


#302 DocBach

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 4,828 posts
  • LocationSouthern Oregon

Posted 13 July 2013 - 04:13 PM

How does a system which makes gaining convergence to achieve a high-damage pinpoint alpha strikes favor it anymore then the system we have currently? You understand that in the proposed system if you torso twisted away to avoid damage you would lose convergence and have to re-establish it again, probably at a higher heat level? It would make gaining convergence to fire alpha strikes over and over a sacrifice in mobility and make shots against you easier.

#303 Erata

    Member

  • PipPipPipPipPipPip
  • 285 posts
  • LocationGoro Company Dropship MK1, Long live Lord Shang Tsung.

Posted 13 July 2013 - 04:18 PM

View PostDocBach, on 13 July 2013 - 03:57 PM, said:


a lot of stuff beside the point



Here's the deal as simply as anyone can put it.
Due to the health points a robot has, even with your convergence changes, the damage they can take before being put in a critical state is low. What that means is that by the time they are capable of entering 90m after being sniped at, they'll be badly hurt and being able to not precisely aim doesn't matter, even inside of 90m.

People want internals and armor raised because it causes the heat mechanic to matter. The thing that makes you shut down. Overheating.

Heat-inefficient alpha-boats have to spend a lot of time cooling down. In a direct brawl, in the CURRENT state of the game, the amount of health a robot has is low enough that their heat is not a large issue. They can deal enough damage to kill or disable robots.

If health was increased, the time spent not attacking during their cooling phase means that heat-efficient brawlers have more time deal more damage because they are not overheating, and due to increased health, not getting parts disabled or outright killed in one to three alpha strikes. The result is that a brawling robot begins to deal more damage than an alpha boat because it is able to survive being alpha'd more than a few times, and it is constantly attacking instead of waiting to cool off.

No, this doesn't affect ballistic alphas as much, but ballistics pay the price in their tonnage and minor risk of ammo explosion.

Does that make sense?

tl;dr -> Alpha Energy boats spending time cooling off during direct, uninterrupted combat with a heat-efficient brawler after Hit Point Buffs results in the brawler dealing more damage, vs current state of game: Heat-inefficient snipers don't feel impact because overall health to kill a robot isn't high enough to impact their rate of fire.

I'm terribly sorry if I got some points wrong. It's how I understand it in my layman way.

edit: I mean, I don't wanna knock your idea down. It could be interesting down the line and if PGI decides to implement support modules that nerf/buff the convergence effect... but that's a kind of mechanic that is literally a year away at least.... and ultimately it won't solve the problem of high-alpha builds in the current state of the game, given how much damage can be dealt to destroy or disable robots.

Edited by Erata, 13 July 2013 - 04:30 PM.


#304 DocBach

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 4,828 posts
  • LocationSouthern Oregon

Posted 13 July 2013 - 04:29 PM

tl;dr version of my suggestion:

convergence would be gained sort of like missile locks, moving slowly makes enemies converge on you quicker, running hot makes your convergence slower, shooting at faster 'Mechs makes your convergence slower. chain fired weapons are not penalized, large groups of weapons fire have to sacrifice mobility and security to line up that perfect shot. Hill humping and poptarting alpha strikes would be impossible as you would have to reestablish several seconds worth of convergence each time.

'Mechs survive longer because they can no longer effectively shoot everything into a single location every time without sacrificing themselves to return fire. The alpha strike would no longer be the standard attack, but something that would require someone take a lot of time focusing their reticle on a target to achieve.

I wouldn't be opposed to longer battles with any solution, however. I don't care if its from a heat system, additional hit points, or making the alpha strike less easier to use. I'm not saying this is the only solution or crapping on any other solutions, this is just my take on how convergence could be handled.

Edited by DocBach, 13 July 2013 - 04:30 PM.


#305 Tegiminis

    Member

  • PipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 123 posts
  • LocationNot In MWO

Posted 13 July 2013 - 04:33 PM

Changing convergence is a way to punish people who are better shots than you, and is thus an incredibly idiotic idea.

#306 DocBach

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 4,828 posts
  • LocationSouthern Oregon

Posted 13 July 2013 - 04:34 PM

View PostTegiminis, on 13 July 2013 - 04:33 PM, said:

Changing convergence is a way to punish people who are better shots than you, and is thus an incredibly idiotic idea.


No, it requires people of all skill levels to get a lock to fire alpha strikes, though it would reward players who can manage their heat better with quicker convergence. Again, this proposal keeps points of convergence consistent to weapons locations, meaning there is no randomness and players would be informed where their weapons will be fired at if they fire before convergence is complete -- and since the players actions can determine how quickly they gain convergence, players have control over how quickly their weapons converge and when they choose to fire them in a group -- or they can choose to shoot single weapons for immediate point of aim, point of impact accuracy.

Edited by DocBach, 13 July 2013 - 04:36 PM.


#307 DocBach

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 4,828 posts
  • LocationSouthern Oregon

Posted 13 July 2013 - 04:38 PM

I'm a leaderboard ranked player in multiple events, I'm a pretty good shot, actually.

#308 DocBach

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 4,828 posts
  • LocationSouthern Oregon

Posted 13 July 2013 - 04:48 PM

Just pointing out your logical fallacy of assuming that I can't play is all.

I'm not catching where I called anybody a tryhard, perhaps I'm misinterpreting your post.

#309 Otto Cannon

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 2,689 posts
  • LocationUK

Posted 13 July 2013 - 04:52 PM

View PostTegiminis, on 13 July 2013 - 04:33 PM, said:

Changing convergence is a way to punish people who are better shots than you, and is thus an incredibly idiotic idea.


Awww! Such a cute ikkle troll!

Removing the scrubcrutch of instant pinpoint rewards skill.

#310 Team Leader

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,222 posts
  • LocationUrbanmech and Machine Gun Advocate

Posted 13 July 2013 - 05:03 PM

While this is infinitely better than that horrible "fix" that the devs have cooked up, it would make hitting fast traveling mechs near impossible.. So I will respectfully have to give this a thumbs down. If I had to choose between this and the dev system, I would pick this one 10 times out of 10. However if only one system of balancing is to be put in, I would have to side with Bill.

The Bach system relies too much on sitting still and never moving your reticule off target, while the Bill system is more about encouraging diverse weapon builds and is much more user friendly to understand. If you were a casual player, just moving your mouse a little behind a target would screw up your aiming, your sights pointed off to infinity, and you would miss for reasons unknown. Heck this would even tick me off and probably happen to everybody. Just making people miss shots to extend battles isnt the solution, but having them pace their shots for maximum accuracy is.

I appreciate the time you put into this and I like that you have tried to fit it to the lore, but Im afraid the other solution would be more optimal for a wide audience while being pretty easy to adapt to. Also there would be no way to differentiate between pulse lasers and regular weapons when you're using multiple types of weapon in one location.

#311 Team Leader

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,222 posts
  • LocationUrbanmech and Machine Gun Advocate

Posted 13 July 2013 - 05:06 PM

I guess its actually a pretty good system and would make it much more enjoyable but the fact that it would be an act of God to hit a light at medium range kind of kills this for me. I like it but I just dont feel its the best possible solution.

#312 soarra

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • 1,312 posts
  • Locationny

Posted 13 July 2013 - 05:36 PM

View PostTeam Leader, on 13 July 2013 - 05:06 PM, said:

I guess its actually a pretty good system and would make it much more enjoyable but the fact that it would be an act of God to hit a light at medium range kind of kills this for me. I like it but I just dont feel its the best possible solution.

your shots will still go where your aiming. just not all weapons will hit 1 spot

#313 Prezimonto

    Member

  • PipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 2,017 posts
  • LocationKufstein FRR

Posted 13 July 2013 - 07:51 PM

[redacted]


@ Team Leader: Don't discount that singly fired weapons won't have to deal with convergence times. Single Gauss, AC20, laser there would be no penalty. This is already what I do in most builds to layer damage on fast mechs, as I don't want to miss and have 4 second cool downs before I try to hit them again. I doubt this type of play will affect fast mechs much, other than to help mitigate their ability to focus fire with multiple weapons on single hard points while on the run .

Edited by miSs, 13 July 2013 - 08:04 PM.
quote clean up


#314 Ken Fury

    Member

  • PipPipPipPipPipPipPipPip
  • 1,016 posts
  • LocationGermany

Posted 14 July 2013 - 12:59 AM

View PostDocBach, on 13 July 2013 - 03:57 PM, said:


Glad to see the Goons providing well thought out and delivered criticism. Thank you for taking your time to grace your presence in this thread.

Medium lasers convergence speed would be quicker up close, especially within 90m. I don't see, however, how that would put them at any more of a disadvantage against PPC's as they currently are, since they aren't widely known for their countersniping ability right now, either. In fact, up close, they would receive a bonus against PPC's as weapons with minimum ranges would be penalized with lowered convergence speed when firing inside those ranges.

Being static means grouped fire converges much quicker than you can converge upon a moving 'Mech; it's an option, but only if you accept having accurate fire placed against you while you wait for your perfect shot, which could be quite a while based on conditions such as heat, ect.

If big guns are used to single fire at the center torso, then the problem I'm trying to solve is in fact mitigated. Single AC/20's or even chain fired PPC's aren't as large as a problem as say, one strike being able to kill half the 'Mechs in the game like we will have when Clan tech is introduced. The intended idea is to reward players with accuracy for single fired weapons, or they can take time to get perfect convergence on a target, which exposes them to enemy fire; without target data, the sniper will not have convergence; he has to have a target locked, and tracked for whatever period of time the modifiers he is penalized with to gain complete convergence locks. Sniping by firing all of his weapons at once instantly would not be possible. He could fire individual weapons accurately at long range, which again would solve the problem of one hit kills.


So the only problem your trying to solve is dying from one shot? You are okay of dying in 1.5 to 2 seconds ? I'm just trying to point out that having the old firing mechanism for single fire will make this system not working on people who can aim and adapt. Please understand that I too don't want Mechs to get instantly destroyed. But the lever needs to be Firepower / Armor.

View Postsoarra, on 13 July 2013 - 03:44 PM, said:

first of all liking your own post negates anything you have to say.
You are wrong about more armor negating alpha builds, More armor and you will see even more alpha builds..


Sure clicking a button is making a whole lot of difference. If you posts are bad enough that you can't like them thats okay. And more Armor means that DPS builds have the chance to play catch up with Alpha builds. It's a simple gameplay mechanic. Sorry that you don't get it.

View PostDocBach, on 13 July 2013 - 04:38 PM, said:

I'm a leaderboard ranked player in multiple events, I'm a pretty good shot, actually.


To get up in the leaderboards you only need to play a lot of matches, hang back in the game and be capable of actually hitting the enemy, also poopsocking. Playing the 8 queue is where skill is proven not poopsocking events. Though thanks for your contribution.

View PostDocBach, on 13 July 2013 - 04:34 PM, said:

No, it requires people of all skill levels to get a lock to fire alpha strikes, though it would reward players who can manage their heat better with quicker convergence. Again, this proposal keeps points of convergence consistent to weapons locations, meaning there is no randomness and players would be informed where their weapons will be fired at if they fire before convergence is complete -- and since the players actions can determine how quickly they gain convergence, players have control over how quickly their weapons converge and when they choose to fire them in a group -- or they can choose to shoot single weapons for immediate point of aim, point of impact accuracy.


1) This requires a rework of ECM mechanics and the Sensor lock mechanic.
2) This solution makes firing at beyond 800m a bit hard, I regularily fight at 1.5 to 2km on Alpine. If you can't converge without a targeting lock or have to keep the reticule on the Enemy this will not easily work at that distance. Because you have to lead your shots.
3) Also Long range+ 4, Target moving 150 kph +5 = 4+ 9/3 = 7 seconds for "max" convergence.... Yeah this sounds like a good idea.
4) Oh single shot still makes this system obsolete...
5) How does this system deal with Zoom levels? Seriously this is an important part. I guess the Reticule expands and shrinks on each level? If you make such a complex suggestion you have to deal with all issues that crop up.
6) Yes tell the enemy to move slower so you gain convergence faster.


View Postsoarra, on 13 July 2013 - 05:36 PM, said:

your shots will still go where your aiming. just not all weapons will hit 1 spot


Yeah unless you shoot single fire, then you can still put everything in one spot... So depending how I fire I use different aim points. Don't you think this is a bit weird?

View PostPrezimonto, on 13 July 2013 - 07:51 PM, said:

[redacted]
@ Team Leader: Don't discount that singly fired weapons won't have to deal with convergence times. Single Gauss, AC20, laser there would be no penalty. This is already what I do in most builds to layer damage on fast mechs, as I don't want to miss and have 4 second cool downs before I try to hit them again. I doubt this type of play will affect fast mechs much, other than to help mitigate their ability to focus fire with multiple weapons on single hard points while on the run .


I don't get what you want to say here? Do you think Light Mechs are too strong? And no penalty to single fired weapons will not affect time to kill in a significant manner. Do you really want such a complicated system do change instant kills into very fast kills. This is an awful lot of stuff to implement and test.

Also if Pulse Laser converge differently than normal weapons, don't you need different reticules for that too?

#315 Tegiminis

    Member

  • PipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 123 posts
  • LocationNot In MWO

Posted 14 July 2013 - 01:01 AM

View PostDocBach, on 13 July 2013 - 04:48 PM, said:

Just pointing out your logical fallacy of assuming that I can't play is all.

I'm not catching where I called anybody a tryhard, perhaps I'm misinterpreting your post.


Good job deleting your post where you called TBK a "goon tryhard."

If you're going to say something, at least stand by it.

#316 Budor

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,565 posts

Posted 14 July 2013 - 01:09 AM

Pinpoint needs to stay, everything else is either random, favors campers or is to complicated to implement.

Raising survivability and installing heat penalties is the far better solution esp. for people that actually pilot a mech in this game instead of re-positioning a turret.

#317 M0rpHeu5

    Member

  • PipPipPipPipPipPipPip
  • Survivor
  • 956 posts
  • LocationGreece

Posted 14 July 2013 - 01:29 AM

I voted no couse i belave that the this would make the game slow and boring even with perfect hit detection.

#318 Ken Fury

    Member

  • PipPipPipPipPipPipPipPip
  • 1,016 posts
  • LocationGermany

Posted 14 July 2013 - 01:36 AM

View PostTegiminis, on 14 July 2013 - 01:01 AM, said:


Good job deleting your post where you called TBK a "goon tryhard."

If you're going to say something, at least stand by it.


Hey I'm not even good enough for the [OPSEC] Battalion of the Word of Lowtax. Hint: This is where we keep our Tryhards under lock and guard.

#319 Takony

    Member

  • PipPipPipPipPipPip
  • Knight Errant
  • 265 posts
  • LocationHungary

Posted 14 July 2013 - 02:25 AM

Convergence nerf with the already wonky hitboxed lights and hit registration bugged? No way in hell thank you.

#320 malice

    Member

  • Pip
  • Legendary Founder
  • Legendary Founder
  • 17 posts

Posted 14 July 2013 - 05:06 AM

I agree this is an issue but the proposed changes are way too expensive to be budgeted properly because would require both programming and UI design.

I think a better solution would be a discrete server side application of something very similar, however. I think the OP could easily be drawn from for something effectively the same ideas.





6 user(s) are reading this topic

0 members, 6 guests, 0 anonymous users