Jump to content

Heat As Main Factor For The Balance.


13 replies to this topic

#1 DanielZX

    Member

  • PipPip
  • 42 posts

Posted 10 March 2013 - 09:59 PM

Greetings mechwariors,

As far as I can see nowadays MWO gameplay suffers from several problems which were already discussed at forums. Everything below is just IMHO. This post is compilation of my several previous posts
Here are some of them:
1) Gameplay pushes players to use boat configs: 4xppc+, 6xsrms6 and so on due to ability to fire all-weapons-at-one-single-point and ineffectiveness of single weapon configs.
2) Lights are a little bit overpowered ( due to synergy of high mobility, potent ssrm and ecm ability + lag shield and errr... mb broken damage models)
3) XL engines in case of facing boat configes make mech too vulnerable because of ability to focus fire in one single point. And their bonus from less weight is too low to compensate it.

I need to mention that imo, right now balance between weapons and armor is fine,we don't need another armor increasment, we don't need to make random hit mechanics as in TT - it will eliminate competitive part of gameplay and difference between ppl who are trying to shoot directly or just firing in enemies direction, also random hit mechanics will give chance of "golden bullet" shots.

My thoughts about this situation:

1) The problem is not only in convergence but in precision aiming AND ability to keep weapon firing on boat mechs several times.So I suggest follwing changes: we need adjust heat containment level for every mech (decrease it drasticly,), increase passive heat generation while running full speed to 15-20%+, and increase cooldown effect from heatsinks.What it will give us: 4ppcs+ or 6srms6 builds would be able to fire just SINGLE alpha strike and they will be turned off or should overrun reactor with possibility of destroying it, so there will be no need in focus builds or they will be used much more carefully. Also it will partially solve problem with ping difference - ppl with ping below 100 have HUGE advantage with current game mechanics vs ppl with ping 200+.If you would be able to make only one single shot from 4xppc before turning off you will you will think twice. Increased effectivness of heatsinks: you won't be able to fire 3+ppc simultaneously, but you would be able to chain-fire longer, it will increase survivability under fire and you will need more skill for sniping.
Optional idea: (but it needs changes in game mechanics) Other way to solve problem of boat-builds is to add level of power output from the reactor, just for example 80 tonns awesome's reactor would be able to ignite instantly only 2 ppc, imagine -lets make power output from reactor 20 and to fire 1 ppc you will need 10 power, so you can use only 2 of them in 1 moment, for sure you can mount 3 or 4 or even more but in one time you will be able to fire just from 2ppcs, Standard engines will have less power output and XL as far more advanced will have higher output. For simplicity - standard engine 200 will provide 18 power output and XL engine 200 will give us 20 power output.

2) Nowadays lights can run fast, hit hard and can carry ecm ( some of them - obviously most popular), movement mechanics is pathetic - on open area light mechs can run in circles with speed 100+ kph and fire streaks and mlas almost non-stop. High mobility with lag shield and difficultes of hitting at their center torso or leg direclty gives them very high survivability. I suggest to HIGHLY increase passive heat generation while running full speed to 15-20% or even more. Light mechs should choose between using speed to get advantage at positioning or to move slower but fire more often, right now they can just run in circles and fire non-stop at heavy/assault back.

3) XL engines:with ability to fire directly to torso sides, weight bonus from XL engine is too low and mehcs durability penalty is too high. I suggest when the right/left torso sides are destroyed it won't cause instant mech destruction but the effectivness of engine will be decreased also passive heat generation should be increased. Example: Pretty baby with XL340, mech fully operational, speed 69kph, passive heat generation at max speed 4%, and if right/left torso will be destroyed max speed should drop down to 45-50kph and passive heat generation should be 15-20%

P.S. hope anyone from dev team will read this and answer :) .

Edited by Zin, 11 March 2013 - 05:29 AM.


#2 Yiazmat

    Member

  • PipPipPipPipPipPipPip
  • The Spear
  • The Spear
  • 531 posts
  • LocationCentral CA

Posted 10 March 2013 - 10:45 PM

#3 I like. a lot actually.give it a 33 percent speed drop and 33 percent passive heat increase per side torso lost. we'll never see this, but it's fun to theorize

#3 Deathlike

    Member

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

Posted 10 March 2013 - 10:48 PM

It depends on how you tweak the speed. Heavier mechs don't care for it, but too low a change would not really hurt a light mech.

#4 Neolisk

    Member

  • PipPipPipPipPipPipPip
  • 547 posts
  • LocationMississauga, ON

Posted 11 March 2013 - 05:14 AM

+1 for introducing engine heat, makes perfect sense after all. This would make you want to stay still to maximize firepower. But in this case you also become an easy target. Obviously for heavy mechs, it could be, for example +10-20% of their heat consumption at max speed, and more like 50-70% for lighter mechs.

Also +1 for 33 percent speed drop and 33 percent passive heat increase per side torso lost for XL engines.

#5 Corvus Antaka

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Knight Errant
  • Knight Errant
  • 8,310 posts
  • Twitch: Link
  • LocationInner Sphere

Posted 11 March 2013 - 06:25 AM

the heatcap is too high.

SHS suck!

DHS outside the engine suck!

these are primary concerns with heat, heat can well be used to balance alpha strike heavy mechs. 4 ERPPC should be the max a mech can alpha strike, that should put a mech to 95% heat.

#6 Zyllos

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,818 posts

Posted 11 March 2013 - 06:47 AM

View PostZin, on 10 March 2013 - 09:59 PM, said:

...
1) The problem is not only in convergence but in precision aiming AND ability to keep weapon firing on boat mechs several times.So I suggest follwing changes: we need adjust heat containment level for every mech (decrease it drasticly,), increase passive heat generation while running full speed to 15-20%+, and increase cooldown effect from heatsinks.What it will give us: 4ppcs+ or 6srms6 builds would be able to fire just SINGLE alpha strike and they will be turned off or should overrun reactor with possibility of destroying it, so there will be no need in focus builds or they will be used much more carefully. Also it will partially solve problem with ping difference - ppl with ping below 100 have HUGE advantage with current game mechanics vs ppl with ping 200+.If you would be able to make only one single shot from 4xppc before turning off you will you will think twice. Increased effectivness of heatsinks: you won't be able to fire 3+ppc simultaneously, but you would be able to chain-fire longer, it will increase survivability under fire and you will need more skill for sniping.
Optional idea: (but it needs changes in game mechanics) Other way to solve problem of boat-builds is to add level of power output from the reactor, just for example 80 tonns awesome's reactor would be able to ignite instantly only 2 ppc, imagine -lets make power output from reactor 20 and to fire 1 ppc you will need 10 power, so you can use only 2 of them in 1 moment, for sure you can mount 3 or 4 or even more but in one time you will be able to fire just from 2ppcs, Standard engines will have less power output and XL as far more advanced will have higher output. For simplicity - standard engine 200 will provide 18 power output and XL engine 200 will give us 20 power output.

...


While I like the idea that heat generation for moving should be higher and the heat capacity should be lower so that if you want to fire many weapons, you have to chain fire them instead of 100% alphas. This is so that you have to keep your aim against a target.

But there is a flaw with this argument.

RoF of weapons. With high RoF, and with low heat capacity to keep people from alphing all weapons into a single point without shutting down, is that then it is better to remove one of the weapons in the alphing and add more heatsinks. Why chain fire 3 shots from PPCs (each being approximately 1.0s apart in time) when instead you could just remove a PPC and alpha strike them? High RoF is invalidating reasons to wield multiples of the same weapon if heat is too high to alpha strike them.

High RoF coupled with weapon convergence is a severe issue to this game. High heat capacity is a side effect of wanting to allow mechs to fire several times before shutting down.

Based on that, I really suggest removing weapon convergence for specific locations. Namely for torso mounts. I suggested this a while ago and got a few likes from the community. Basically, individual weapons never converged but instead shoot straight (torso mounted weapons are pointed inwards around the Torso crosshair but will not converge).

The only convergence that happens are that the arms will both converge onto the Arm crosshair. The individual weapons on the arms themselves will still fire spaced relative to their placement on the arms but be converged onto the Arm crosshair.

The torso mounted weapons will always be pointed toward the Torso crosshair relative to the cockpit. Thus, most weaponry will land below the Torso crosshair. The weapons will be ranged in to the distance the current target selected. If no target is selected, then it will range to the longest weapon in the torso weapon group. Here is an example of what I am saying:

Posted Image

If you can see from above, torso weapons will aim relative to their placement on the mech in accordance to the distance to the target. The convergence "square" will be of the same size, no matter where it is pointed at, when a target is selected. When no target is selected, then it will converge at the furthest weapon in the torso group (in this case, at 270m for the Medium Laser).

When the target gets closer, this square will enlarge. Thus, for those hunch lasers to land where you want them, you have to physically look down for those lasers to land where you want them. Thus being up close makes it kinda harder to converge weapon fire while being at optimum distance of the weapon makes it easier.

Arms also do this, but converges their square based on where the Arm crosshair is pointed (basically how it works now except the individual weapons will not all converge onto a single point but instead land relative to their placement).

What this does is implement a way to remove a good amount of convergent fire from mechs while not adding a random element like cone-of-fire. It is done in such a way that a skilled MechWarrior can get most of their weapons to fire and hit a single location, but will take additional time to aim, but in the heat of battle, it will most likely be better to attempt to just get all the shots to even land on the mech, thus removing a lot of the converging fire.

It will also place a huge emphasis on arms because they can converge a lot of weaponry easily onto single points. So it will be more important to aim for arms on mechs that have lots of firepower in their arms.

And this also adds another dimension to the differences between varients. Weapon placement becomes a big deal. The HBK-4P is a good example of this. Many of the lasers will be good at focusing weaponry into a single point, but the draw back is that it is a big target. But sense convergence is not 100%, it might be as simple as it is now to knock out that hunch. Even in an Atlas. And if that is the case, think of how tough that Atlas might now be?

Edited by Zyllos, 11 March 2013 - 07:01 AM.


#7 DanielZX

    Member

  • PipPip
  • 42 posts

Posted 11 March 2013 - 07:02 AM

We can avoid convergence problem if we will be able to set focus point distance like in flight simulators. For snipe configs you should set distance at optimal your weapon allows, for close combat you will be able to set mb closer then 200m and etc. But then devs should add two options to set distance separatly for arms mounted weapons and for torso mounted.This is how it works in real life when different weapons are placed in different points of airplane as an example.

Edited by Zin, 11 March 2013 - 07:04 AM.


#8 Zyllos

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,818 posts

Posted 11 March 2013 - 07:08 AM

View PostZin, on 11 March 2013 - 07:02 AM, said:

We can avoid convergence problem if we will be able to set focus point distance like in flight simulators. For snipe configs you should set distance at optimal your weapon allows, for close combat you will be able to set mb closer then 200m and etc. But then devs should add two options to set distance separatly for arms mounted weapons and for torso mounted.This is how it works in real life when different weapons are placed in different points of airplane as an example.


The only problem with this is that it is adding complication to the game. I want the game to automate the process.

Edited by Zyllos, 11 March 2013 - 07:08 AM.


#9 MaddMaxx

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 5,911 posts
  • LocationNova Scotia, Canada

Posted 11 March 2013 - 09:39 AM

For a Single instance of a weapon, no worries right, it shoots where you point. Torso or arm directed.

What if when you add a second (and or subsequent units) the convergence of that weapon set is auto calculated (and locked) to its rated Maximum Range.

So 1 ML is pinpoint at all ranges out to Maximum, where as 2 ML are only pinpoint at exactly convergerange = "540m"

If every weapon has its preset maximum convergerange = "540m" equal to its actual max range then weapon groups fired from a group set across a mech chassis would only ever meet at that max range.

Building high alpha weapons groups would require a very exacting set of weapons be grouped and the use of those weapons would be confined, to a certain extent, to those pre-set ranges.

Yes you could rush into half of your convergerange = but then the spread would be preset already and hitting would be a matter of sweeping the Beams, or shooting with an offset knowing that likely only 50% of your ordinances will hit.

Given this setup, a Mech could be designed to have a stepped convergerange = config thus that as a Mech moved forwards into the fray, a new set of weapons would become the pinpoint primaries. Of course, the Reticule, or weapons console itself (turn green for the current pinpoint set) would give the pilot this info in real time due to the frantic nature of the game play.

Edited by MaddMaxx, 11 March 2013 - 09:40 AM.


#10 Corvus Antaka

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Knight Errant
  • Knight Errant
  • 8,310 posts
  • Twitch: Link
  • LocationInner Sphere

Posted 11 March 2013 - 11:26 AM

No to the XL engine. You get a lot of speed benefit from an XL engine, and the whole point is the weakness it gives you via sidetorsos.

#11 DanielZX

    Member

  • PipPip
  • 42 posts

Posted 11 March 2013 - 12:32 PM

In TT there is no focus fire at single part of the mechs, so XL engine really don't have any penalties at all. Speed benefit is actual only for lights and some meds, but for lights chance of beeing killed by hitting to the torso sides is negated by small hitbox and high movement speed (same chance to hit center torso or sides), other mechs suffer greatly most of the time.

#12 Yiazmat

    Member

  • PipPipPipPipPipPipPip
  • The Spear
  • The Spear
  • 531 posts
  • LocationCentral CA

Posted 11 March 2013 - 12:55 PM

View PostColonel Pada Vinson, on 11 March 2013 - 11:26 AM, said:

No to the XL engine. You get a lot of speed benefit from an XL engine, and the whole point is the weakness it gives you via sidetorsos.


Why? You'd rather DIE instantly than be able to keep fighting? Foolishness, Panda, foolishness. So what if you slow down, you're still able to lay down some fire before you get CT'ed. There's no reason NOT to like this.

#13 Yiazmat

    Member

  • PipPipPipPipPipPipPip
  • The Spear
  • The Spear
  • 531 posts
  • LocationCentral CA

Posted 11 March 2013 - 01:02 PM

View PostZyllos, on 11 March 2013 - 06:47 AM, said:


While I like the idea that heat generation for moving should be higher and the heat capacity should be lower so that if you want to fire many weapons, you have to chain fire them instead of 100% alphas. This is so that you have to keep your aim against a target.

But there is a flaw with this argument.

RoF of weapons. With high RoF, and with low heat capacity to keep people from alphing all weapons into a single point without shutting down, is that then it is better to remove one of the weapons in the alphing and add more heatsinks. Why chain fire 3 shots from PPCs (each being approximately 1.0s apart in time) when instead you could just remove a PPC and alpha strike them? High RoF is invalidating reasons to wield multiples of the same weapon if heat is too high to alpha strike them.

High RoF coupled with weapon convergence is a severe issue to this game. High heat capacity is a side effect of wanting to allow mechs to fire several times before shutting down.

Based on that, I really suggest removing weapon convergence for specific locations. Namely for torso mounts. I suggested this a while ago and got a few likes from the community. Basically, individual weapons never converged but instead shoot straight (torso mounted weapons are pointed inwards around the Torso crosshair but will not converge).

The only convergence that happens are that the arms will both converge onto the Arm crosshair. The individual weapons on the arms themselves will still fire spaced relative to their placement on the arms but be converged onto the Arm crosshair.

The torso mounted weapons will always be pointed toward the Torso crosshair relative to the cockpit. Thus, most weaponry will land below the Torso crosshair. The weapons will be ranged in to the distance the current target selected. If no target is selected, then it will range to the longest weapon in the torso weapon group. Here is an example of what I am saying:


If you can see from above, torso weapons will aim relative to their placement on the mech in accordance to the distance to the target. The convergence "square" will be of the same size, no matter where it is pointed at, when a target is selected. When no target is selected, then it will converge at the furthest weapon in the torso group (in this case, at 270m for the Medium Laser).

When the target gets closer, this square will enlarge. Thus, for those hunch lasers to land where you want them, you have to physically look down for those lasers to land where you want them. Thus being up close makes it kinda harder to converge weapon fire while being at optimum distance of the weapon makes it easier.

Arms also do this, but converges their square based on where the Arm crosshair is pointed (basically how it works now except the individual weapons will not all converge onto a single point but instead land relative to their placement).

What this does is implement a way to remove a good amount of convergent fire from mechs while not adding a random element like cone-of-fire. It is done in such a way that a skilled MechWarrior can get most of their weapons to fire and hit a single location, but will take additional time to aim, but in the heat of battle, it will most likely be better to attempt to just get all the shots to even land on the mech, thus removing a lot of the converging fire.

It will also place a huge emphasis on arms because they can converge a lot of weaponry easily onto single points. So it will be more important to aim for arms on mechs that have lots of firepower in their arms.

And this also adds another dimension to the differences between varients. Weapon placement becomes a big deal. The HBK-4P is a good example of this. Many of the lasers will be good at focusing weaponry into a single point, but the draw back is that it is a big target. But sense convergence is not 100%, it might be as simple as it is now to knock out that hunch. Even in an Atlas. And if that is the case, think of how tough that Atlas might now be?


So, you're comparing 1940's era fighter planes and .50 cal. machine guns to 3050 walking tanks with laser cannons? Who's to say they haven't invented focusing lenses to angle the lasers at the same point on the fly? I'm to say.. Because that's probably how it works, because it's already in game.

#14 Corvus Antaka

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Knight Errant
  • Knight Errant
  • 8,310 posts
  • Twitch: Link
  • LocationInner Sphere

Posted 11 March 2013 - 01:32 PM

View PostYiazmat, on 11 March 2013 - 12:55 PM, said:


Why? You'd rather DIE instantly than be able to keep fighting? Foolishness, Panda, foolishness. So what if you slow down, you're still able to lay down some fire before you get CT'ed. There's no reason NOT to like this.


As an XL user, theres a lot to like.

For game balance, no. It'd be too OP, and standard engines wouldn't be as inticing as they are now.





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users