88 replies to this topic

[MustrumRidcully]

I created a new evaluation of the weapon balance of this game, using the current build, including its unique heat sink mechanics.

TL;DR:
Balance still isn't there. Double Heat Sinks are definitely an upgrade and make high heat weapons more competitive. Ballistics tend to get more efficient the higher their range, Energy Weapons it's the other way around - which is what you would expect for all weapons. Range is an advantage, and so you need to pay for it in some manner - for example by dealing a bit less damage for the same weight.

Methodology

I went with a slightly different approach this time than for the first time, to better account for the unique aspects of the heat sink implementation. By which I primarily mean the way engine heat sinks work (both general, as well as specifically they work now for engines).

The below chart is based on the following constraints:
A Mech with a 250 Engine Rating is used (this gives it 10 "free" heat sinks)
The Mech has no defined tonnage limit at this point, but it is equipped with enough instances of each weapon in the chart so that it achieves the following goals:

1) It manages to deal 160 damage within 20 seconds.
2) It does so without overheating within that time (it might mean it overheats at 21 seconds)
3) It has enough ammunition to repeat this 6 times. (So that's 120 seconds or 960 damage worth of ammo ).

I chose the values for damage and time based on the general experience of this game - I believe most "intense" engagements are within the 20 second time limit - by then, the combatants either seperate again, or one of them is defeated.
The damage value is based on the idea that this might about the damage you will deal to core a single 100 ton Assault mech with maxed out armour, with enough leeway to account for missing the center torso.

After all this is done and calculated, I compare the weight of the weapon system including ammo and heat sinks vs. the damage it can inflict. (E.g. Damage / Weight). This is the "Efficiency" of the weapons

For the weight, I did not calculate the engine weight or engine heat sinks, of course. (That means that many ballistic weapons may have more heat sinks than they'd need.

The Spreadsheet
The excel spreadsheet on which this chart is based on can be found on Google Drive. (now converted to Google Docs - it is still as "view only", so if you want to edit something, you need to download it or copy it to your own document.)

The Chart



The vertical axis lists the efficiency (damage dealt in the targeted time divided by tonnage to run it), the horiziontal axis obviously lists the weapons (within the categories of ballistic, energy and missiles, ordered by range):

Alternate Sorting (Ballistics and Energy weapons sorted by Range)

Spoiler

Ballistics
As we see on this table, the only ballistic weapon that notably gains with Double Heat Sinks is the AC/2. The other ballistics are "carried" by their engine heat sinks alone.

Note that the Ultra AC/5 Double Shot stats do not account yet for the jamming effect. I think what is important to notice here though is that the Ultra AC/5 would be great already without the double shot - and I wouldn't be surprised if the new Jamming probability and jam duration would make not using the double shot feature more useful.

The LBX-10 AC is overrated on this chart, since I didn't assign it an efficiency yet - which it should have, since it has the same drawbacks as the SRMs and LRMs - it spreads it damage around. The efficiency of the LBX-10 AC in this chart would represent you basically standing within 100m or less of the enemy, which isn't really utilizing its range, but would mean using it to brawl like with an AC/20.

Energy Weapons
The Energy Weapons notable benefit from Double Heat Sinks. Without them, even the more efficient ones are underpowered against most Auto-CAnnons and the Gauss, but with DHS, some of them pull ahead.

Very noticeable is how Medium Lasers and Small Lasers are more efficient than most weapons, with the small laser pulling ahead considerably.

The Pulse Lasers are extremely weak. The ER Large Laser is actually less efficient than the PPC.

Direct Fire Weapons
If you compare weapons with similar range, it still seems that other than the AC/20 and AC/10, most ballistics excel their non-ballistic counterpart, some by a notable margin.
While the ER PPC has the largest range of all, it seems so much less efficient than any other weapon.

Missiles
Streaks may seem low on this chart, but I am currently not applying any to-hit based modifiers on the weapon efficiency. SSRMs are less "efficient" only if you ignore that they never miss.


Alternative Scenarios
I also greated alternative TET scenarios. This is a series of graphs for those cases.

Spoiler

The first two charts are based on "low" damage targets. That means that for example you often don't need more than 2 of the heavier ballistics to achieve the goal. The damage figure I would consider kinda realistic, the drawback is that the ballistic builds here usually have a lot of "waste" dissipation or heat capacity left over, meaning if these were mech builds, you could easily upgrade their damage output by adding a lighht energy weapon or maybe even a SRM.

For the number of heat sinks, it matters mostly how many weapons you have, and how long they are supposed to fire. The longer, the more sinks you tend to need. How many weapons you fire depends on what I call "damage volume"), basically - how much damage do I want the setup to deal at a minimum in the targeted time frame.
For ammo consumption, I need not just to consider how much ammo a weapon needs in one targeted engagement time frame, but also how many I expect to have. I describe this as the frequency. I tend to assume that if you have a high damage volume, you don't need quite such a high frequency - there are only 8 enemy mechs to kill, after all.



Chart 1 - Single Heat Sinks, Low Damage Volume, High Frequency
The first chart describes damage to tonnage efficiency based on single heat sinks (with 10 "free" engine heat sinks).
http://i883.photobuc...Frequency01.png


Chart 2 - Double Heat Sinks, Low Damage Volume, High Frequency
The second chart describes damage to tonnage efficiency based on double heat sinks (again with 10 "free" engine double heat sinks.)
http://i883.photobuc...Frequency01.png


Chart 3 - Single Heat Sinks, High Damage Volume, Low Frequency
The third chart describes damage to tonnage efficiency based on single heat sinks (with 10 "free" engine heat sinks), but this time, the damage requirements have been doubled. (Which means that at least starting at the 20 second duration mark, even most ballistics need to add a few heat sinks and it would no longer be feasible to add a medium laser or small laser to the build without negatively affecting its endurance.)
http://i883.photobuc...Frequency01.png


Chart 4 - Double Sinks, High Damage Volume, Low Frequency
The third chart describes damage to tonnage efficiency based on double heat sinks (with 10 "free" engine heat sinks), but this time, the damage requirements have been doubled. (Which means that at least starting at the 20 second duration mark, even most ballistics need to add a few heat sinks and it would no longer be feasible to add a medium laser or small laser to the build without negatively affecting its endurance.)
http://i883.photobuc...Frequency01.png





In this graph you'll notice that the PPC for example will pull ahead of the Gauss RIfle with DHS.

Raw Data
This can be found in the Excel sheets, but here is one chart for the 20 second / 160 damage / 6 engagements chart.

Spoiler

You may notice that you don't need many - or any - heat sinks to run most ballistics for this duration and damage goal.

Single vs Double Heat Sinks
There seems to be very little reason to not use DHS. Unless you run out of Crit Space. WHich is quite possible.


How do I use this information?
If you're a developer or like to pretend you are one, you could use the underlying sheet to tweak weapon values or heat sink mechanics.

If you are a min/maxer, you could use this to find the ideal setup for your mech. If you even have a better idea than I do how well you can work with different weapons, you could plug in different weapon efficiency stats.

In either case, you can use this as evidence that I invest too much time in analyzing this game rather than playing it. But I can assure you, despite all the number crunching, no digits were harmed.

Edited by Niko Snow, 16 April 2013 - 04:45 PM.

[Stingz]

It's pretty obvious to anyone that running DHS is always better, since the heat scale is too much for SHS. Triple fire rate and standard heat dissipation doesn't work too well.

These graphs show how badly DHS outclasses SHS.

Edited by Stingz, 25 November 2012 - 07:23 AM.

[MustrumRidcully]

Stingz, on 25 November 2012 - 07:23 AM, said:

It's pretty obvious to anyone that running DHS is always better, since the heat scale is too much for SHS. Triple fire rate and standard heat dissipation doesn't work too well.

These graphs show how badly DHS outclasses SHS.

The hidden secret of this chart is the "crit efficiency"; of course.

Just as an example: To have the Medium Lasers achieve the stated goal (200 damage in 20 seconds withoutoverheating), you need 6 medium lasers and 34 out-of-engine heat sinks... That's a lot. With DHS, it improves down to 14 DHS - which means 42 Crits. Finding 42 free crits for the DHS will not be an easy task. Better hope you got weapon hard points in your legs and head.

[Asatruer]

I can fit a grand total of 9 external DHS in my HBK-4P (5 DHS shy of 14 with a 250 engine) to get more DHS into that clunker I would have to drop 2 tons of armor to drop the endo steel, then 1 additional ton of armor per DHS I wanted to add in the newly opened critical spaced. Makes me wish I could sacrifice various actuators in the arms for the cause.

[Kobura]

You might like my slightly different comparative "is this weapon useful for what I want" chart, MustrumRidcully:

https://docs.google....5QRHd5Wnc#gid=0

It focuses more on using weapons in roles, and what weapons are being used outside their roles and why (DamageVsClosing for instance, displaying the damage that can be inflicted on a charging enemy moving tabletop 5/8 from weapon maximum effective range)

It also shows that the gauss is almost, almost balanced, the only painful item is the extreme long-range snapshot-damage compared to its' competitors, the PPC and ERPPC.

My table doesn't take into account damage over time, heat, or ammo though, it was more a straight comparison for helping keep certain guns doing what they were meant to be doing (this quad-AC2 infighter stuff makes my CBT-lover-at-heart twitch, for instance)

[Indoorsman]

I think LRM are still too good, and aren't they getting a .1 dmg per missile buff next patch? They don't SEEM too good on that chart. Of course your math isn't really about weapon balance but more about heat balance? Maybe it's just that I use an AWS-8R which can alpha strike 60 missiles at a time somehow, that probably needs looking into. It one shots commando and can 3 shot stationary Atlas. And I can get 3 shots off w/o heat being an issue... I'll post a video later when it finishes uploading.

A video is worth more than a thousand maths :-p

[xXBagheeraXx]

Stingz, on 25 November 2012 - 07:23 AM, said:

It's pretty obvious to anyone that running DHS is always better, since the heat scale is too much for SHS. Triple fire rate and standard heat dissipation doesn't work too well.

These graphs show how badly DHS outclasses SHS.

As they SHOULD, in cannon most mechs did away with singles as they became available...think of DHS as clan tech in that regard. Its just BETTER.

[MustrumRidcully]

Indoorsman, on 25 November 2012 - 10:59 AM, said:

I think LRM are still too good, and aren't they getting a .1 dmg per missile buff next patch? They don't SEEM too good on that chart. Of course your math isn't really about weapon balance but more about heat balance? Maybe it's just that I use an AWS-8R which can alpha strike 60 missiles at a time somehow, that probably needs looking into. It one shots commando and can 3 shot stationary Atlas. And I can get 3 shots off w/o heat being an issue... I'll post a video later when it finishes uploading.

A video is worth more than a thousand maths :-p

As I said - I was experimenting a bit with giving weapons an efficiency modifier - this could be something that accounts for difficulty to hit and the like. For missiles, I made this an efficiency of 50 % - LRMs spread their damage around a lot. I gave ballistics and energy weapons the same modifier (1), however. If there wasn't the efficiency modifier for LRMs and SRMs, they'd be far higher (I believe it may not exactly be 2 times, but close enough).

Basically, if someone thinks that LRMs need a buff, he must be assuming that their damage is wasted twice as often as the damage from energy and ballistic weapons...

If you wish to play around with the chart yourself, you can download it via Google Drive: https://docs.google....a18wVG9TUTBjaHM

Edited by MustrumRidcully, 25 November 2012 - 11:53 AM.

[Indoorsman]

Here's LRMs having much less than 50% spread, I deal 1,068 dmg more than the next highest dmg :-o



EDIT:
To be fair, I have Artemis for all the kills in that video and for some of the kills I was able to tag too. But you kinda have to account for best/worst case scenarios anyways.

Edited by Indoorsman, 25 November 2012 - 03:58 PM.

[Stingz]

Indoorsman, on 25 November 2012 - 10:59 AM, said:

Maybe it's just that I use an AWS-8R which can alpha strike 60 missiles at a time somehow, that probably needs looking into. It one shots commando and can 3 shot stationary Atlas. And I can get 3 shots off w/o heat being an issue... I'll post a video later when it finishes uploading.

If a Light pilot is getting pounded by 60 missiles, they messed something up. Like not using that speed to get behind/under cover, or avoid getting spotted.



Add:
After seeing your loadout OFC It will get killed, x4 ALRM 15s + TAG is insane setup. Artemis + TAG means about 80-100% hit rate with those LRMs. (WTH, thing must cost entire Atlas mechs to run.)

Counters:
- The Awesome melts under LRM fire, brick wall target.
- Enemies will spot your setup, and try to sneak in under 180m (though I don't think that happens often).
- About 36% heat per salvo isn't very practical, neither is the ammo burn and cost.
- Speaking of ammo, that Awesome probably has horrific ammo explosions.
- ECM will also alleviate the problem of LRMs melting everything, when it EVENTUALY comes out. (*Raven-3L cries in corner*)

Edited by Stingz, 25 November 2012 - 07:28 PM.

[Koningswulf]

What one can see in this charts and that of Kobura is that small and medium laser are superior to all other weapons as long as you have the hardpoints and ton for them. The energy pulse weapons and PPC weapons are extremly bad weapons of choice
in both DPS and HPS. The ballistic weapons are fine except the AC/2 with a way to high DPS and also has shake.
I strongly hope the Devs will nerf the small and medium lasers and the AC2.
buff all the Pulse weapons and the PPC.
Also please lower the heat for the Energy Extended weapons, its ridicoulous high now. The extra range os not a enough of a good bonus for that increase in heat.

[Kobura]

Or, just make DHS do their whole job, and not just 75% of it...

[Indoorsman]

Stingz, on 25 November 2012 - 07:08 PM, said:

WTH, thing must cost entire Atlas mechs to run

It is break even after a win, lost money on a loss... if I spent much ammo

[MustrumRidcully]

Koningswulf, on 25 November 2012 - 08:31 PM, said:

What one can see in this charts and that of Kobura is that small and medium laser are superior to all other weapons as long as you have the hardpoints and ton for them. The energy pulse weapons and PPC weapons are extremly bad weapons of choice
in both DPS and HPS. The ballistic weapons are fine except the AC/2 with a way to high DPS and also has shake.
I strongly hope the Devs will nerf the small and medium lasers and the AC2.
buff all the Pulse weapons and the PPC.
Also please lower the heat for the Energy Extended weapons, its ridicoulous high now. The extra range os not a enough of a good bonus for that increase in heat.

The ballistic weapons are not fine, I think. The graph shows that longer range weapons tend to be more efficient than short range weapons - it should be the other way around.

I think part of this is inherited from the table top - in there, 4 medium lasers might technically have the same DPS as an AC20, but the 4 Mediums will not hit the same spot. So the table top rules weighted against high damage weapons, because they could be extremely effective in dealing kill shots.
But with convergence, that distinction is gone.

[Asatruer]

MustrumRidcully, on 25 November 2012 - 11:12 PM, said:

I think part of this is inherited from the table top - in there, 4 medium lasers might technically have the same DPS as an AC20, but the 4 Mediums will not hit the same spot. So the table top rules weighted against high damage weapons, because they could be extremely effective in dealing kill shots.
But with convergence, that distinction is gone.

Pretty much yeah, and it is also possible for not all four medium lasers to hit. Over the course of multiple turns, if given the unlikely stable 50% hit chance, both the 4MLs and one AC/20 will do the same DPS, but the 4MLs will more likely be able to do at least some damage every turn. While not a very strong benefit, it probably is useful in some cases. The 4MLs also gets the more noticable benefit of being more damage resistant, as they have to be destroyed independently rather than all at once, still not a huge perk, but it is one.

While convergence plays a larger part in boosting the performance of the low end light weapons, I think that being able to group fire has a pretty under recognized effect as well, in that it is what is effectively removing TT's to-hit chance on a per weapon basis. The two things are mostly as inseparable as two sides of a coin, but without group fire, the fact that all the weapons on different points of the mech all aim at the same spot the retical is pointing at would not make as large of an impact on the smaller weapon vs larger weapon balance. One place where its effect can be seen, and were convergence is a non-factor, is the alpha style firing of multiple guided weapons (SSRMs and LRMs) or to a lesser degree the unguided scatter weapons (SRMs and LBX).

[MustrumRidcully]

I believe I have an example calculation for TT somewhere:

Quote

To hit a stationary Atlas Center Torso at 270m with 2 Medium Lasers, you need to roll
2d6 vs 4 (gunnery skill) +2 (long range) = 6. That's roughly a 50 % chance to hit with one medium laser, to hit with both, it becomes a 25 % chance.

Now, the chance to hit Center Torso after a succesful attack roll, is about 14 %. To do it twice in a row, it would be about 2 % chance.

So you overall have a 0.5 % chance to hit center torso with 2 medium lasers on any given turn.

A PPC would roughly have a 85 % chance at that distance to hit, and also the 14 % chance for a CT hit. That's giving us a ~12 % chance to hit Center Torso.

Edited by MustrumRidcully, 26 November 2012 - 01:43 AM.

[Vapor Trail]

One thing occurs to me. Is this utilizing "True Doubles" (0.2 heat/sec) all the way through, or our current mish-mash of 0.2 and 0.14
Nevermind. I downloaded the spreadsheet and opened it up.
Something else.

UAC5.
How does the jam chance affect the weapon over long periods of time?Shouldn't the formula be something like this:

average cooldown = (singleshotCD/2) + ((mean value of jam time) * (percentage jam chance))/2

Basically what this says is that you can fire the weapon twice in the span of a single cooldown of singlefire mode.

But when it jams (25% of the time) the average jam value, divided by two, is added to each of the cooldowns of the weapon... effectively tacking on the entire jam time to the second shot, and averaging the two.

The math basically winds up with the double shot's average cooldown (and thus damage output over time, and thus damage efficiency) being worse than that of the single by a small margin.
Which is how I think it should be.

Formulae:

= (singleshotCD/2) + ((mean value of jam time) * (percentage jam chance))/2

= (1.1 / 2) + ((3+8)/2) * (.25)/2)

= 1.24

Formula to be entered into WeaponSheet cell E8:
=(E7/2)+ (((3+8)/2) * (0.25))/2


Not sure on the max/min jam times for the UAC 5 though.

Edited by Vapor Trail, 26 November 2012 - 04:13 AM.

[Slanski]

Some of this is expected:
The MLaser and SLaser always were a low cost converter of heat sink to damage power. The unseen Warhammer and Marauder spending 2-3 tons on them to fire instead of the PPC at under 90m.

What I find underused in your otherwise very well constructed chart (though please name the Axis next time!) is the opportunity of range. I think most of the minmaxing and mismatching of weapon cost (DPS/ton, Alpha/ton, HPS/ton) results from skewed range. In a word: MLasers fire too far, so do LLasers. They develop full DPS to their max range and fall off for twice beyond. This combined with map design and small maps means you purchase a range advantage on the PPC/ER PPC/ER LLaser that is hardly usable in engagements.

Suggestion: Half current Laser fall off values or even cut them to 1/3rd (Lasers, not energy weapons, the PPC needs that range), so that damage degrades rapidly after 270 on an MLaser. Combine the low cost efficiency of these systems with an opportunity cost at range. Then improve out of engine DHS efficiency to 2.0 or 1.7 (plus skills), to permit the PPC mechs right of existence.

Design problem: How to make the PPC viable without letting Mlasers and LLasers take over and destroy the game.

Edit:
@Asatruer:
Very well observed. I think a possible solution would be that if you group more than 2 Lasers/weapons only the first 2 will converge exactly on reticle, but the rest of the firing group shows a certain "spread" on fire and would somewhat diverge from cross hair. This would require multiples of low end weapons to either spread their DPS a little like SRMs over the target or have the pilot acurately chain fire them into target, while being required to be on target and face the opponent, exposing their CT. The Alpha advantage of big guns would be reinstated (and the resulting cost in tons for them).

Edited by Slanski, 26 November 2012 - 04:00 AM.

[Vapor Trail]

There needs to be a third axis to the graph to get Range data on there. The Gauss will hold it's efficiency steady for longer than the PPC's optimum range... and then hold the efficiency better as the range increases past the Gauss' optimum. The definite integral of the curve efficiency(range) curve (Calculus speak, meaning the area under the line generated by the efficiency as a function of range curve) is going to be much greater for the Gauss than it is for the PPC.



Possible partial solution to design problem: extend laser burn time. Smaller lasers have longer base burn times.
Eg
Small Laser, 100% of current minimum duty cycle.
Medium Laser, 66.67% of current minimum duty cycle.
Large Laser, 33.33% of current minimum duty cycle.

Or burn times of
SL: 3 sec
ML: 2.66 sec
LL: 1.4 sec

Actual duty cycle times are unchanged. So a small laser basically can fire a continuous beam... but it doesn't output any more or less DPS than it does now.

Those are just arbitrary figures just to illustrate the mechanic, not serious estimates of where they need to be.

The point behind this is that the longer the burn time on the smaller lasers, the less chance there is of them delivering their entire energy to the target location, either by the gunner missing the mark, or by the target actively denying the shot.

Math really won't help much here I think. Trial and error might be better at setting burn times.

The other half of that would be to up heat sink base dissipation rate. Not much, perhaps .01 or .02, but enough that the high heat weapons (ERPPC, PPC, etc) benefit enough that they become viable again, and the smaller energy weapons (which are weaker due to their increased burn times) come back up. The Large Lasers might still be a problem, so looking at their burn times is a must.

The increasing of base burn times would have the side effect of improving the relevance of the Pulse weaponry as well.

Edited by Vapor Trail, 26 November 2012 - 04:22 AM.

[MustrumRidcully]

Slanski, on 26 November 2012 - 03:54 AM, said:

What I find underused in your otherwise very well constructed chart (though please name the Axis next time!) is the opportunity of range. I think most of the minmaxing and mismatching of weapon cost (DPS/ton, Alpha/ton, HPS/ton) results from skewed range. In a word: MLasers fire too far, so do LLasers.

If I was the systems guy of Piranha Games, this is what I would do with this chart:
Ignore it. No Player has a clue how this works, no point looking at it
I would try to find values for the weapons where the efficiency of all weapons drop off with their range. The precise decline by range is something that may need some experimentation or a deeper mathematical analysis.

If you know what the decline should look like, you could of course put this into the efficiency calculation and instead try to go for a straight line.