Fractal Softworks Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Starsector 0.9.1a is out! (05/10/19); Blog post: Raiding for Fun and Profit (11/27/19)

Pages: 1 ... 5 6 [7] 8

Author Topic: The Balance Beam  (Read 21467 times)

Ghoti

  • Captain
  • ****
  • Posts: 283
    • View Profile
    • Email
Re: The Balance Beam
« Reply #90 on: May 01, 2017, 12:03:44 PM »

Personally I like where this thread is coming from.

That said the objective of making all weapons equally valuable for cost, and aligning them all rigidly to a formula is definitely a great idea. It's an interesting experiment. Props xenoargh. Maybe not the kind of thing I want to see fully implemented in vanilla, but I like the idea for a mod. Or maybe even a meta mod  ;)
Logged

DeltaV_11.2

  • Lieutenant
  • **
  • Posts: 56
    • View Profile
Re: The Balance Beam
« Reply #91 on: May 01, 2017, 01:01:28 PM »

This ain't "armchair"; it's math; it doesn't care about what you want it to say.  You wanna argue about weighting, fine, let's have a grownup discussion.  Otherwise, go away, the adults are working here.
Spoken like somebody who has never been within 20 feet of a course on statistics and data analysis. Bad math is worse than no math, and how you analyze a set of data can have immense implications on the validity of your results. The math can't lie, but how you interpret and choose to apply it can, because your interpretations and applications are fundamentally subjective.

To start with, your presumptions totally ignore the value of alpha strike weapons, and single-shot damage in general. A gun that deals 2000 damage every 1 seconds is better than one that deals 250 damage every 0.125 second, because the first will both have a higher damage rate in any real situation except the limiting case(consider a firing period of 2.5 seconds, the first fires 3 times the second 21, for 6000 vs 5250 damage dealt), and is easier to employ as it doesn't require constant aiming and is forgiving of short periods of no shot opportunities.

In Starsector, single-shot damage is also very important because of how it impacts armor penetration. The thread you linked actually goes into a lot of depth about what this means for many weapons.

Your hit calculation is doubly flawed because it ignores the huge effects of ITU/DTC which appears on many ships, and it presumes that combat always takes place at maximum weapon range. Trying to arbitrary categorize how important weapon range is in order to apply fixed factors is also not going to go anywhere useful. The utility of range is fundamentally not disentangleable from the weapon's performance both relative to other weapons(having the most range in a category has a lot more value than second-best), and relative to itself- a weapon with a lot of alpha threat is much more dangerous with long range.

There's no valuation mechanism for turret track rate or projectile velocity, which are in practice more important than weapon arcs in a lot of scenarios. Energy weapons get a significant edge out of this because they come with higher velocities than all but a few ballistic weapons, making them easier to hit with. This also ties in with how good a weapon is against different types of target- it matters less on bigger weapons or on ships that are roled around fighting capitals and cruisers.

OP value is not linear and should not be. Slots are a finite balancing resource on a ship, and therefore OP must have diminishing returns to scale for the same mounts in order to make slot restriction meaningful. Having things otherwise results in optimums being single high-OP weapons, which is problematic in balance terms. From an abstract perspective, there's also the reality that some weapons should simply be better than others. All weapons being the same flavor is boring gameplay, and if there aren't any meaningful decisions to be made there's no play to be had.

In general trying to separate weapon performance from their platforms is not going to work. Weapons don't exist in a vacuum and trying to balance them as if they do doesn't work. Yes, that precludes analyzing things with a simple spreadsheet. Life is hard. Using math to analyze complex systems is harder. Deal.

Looking at your sheet here's the major balance issues I see in it and why they exist. Gauss Cannon- high alpha, best range. Gauss is a very dangerous weapon and giving it decent relative flux stats is insane. With 800 alpha it armor punches like a 200 damage HE weapon, so not actually half bad. Bad DPS does not fix the issue that a Gauss that doesn't have serious flux costs can eventually pick any ship apart at range with little risk of retaliation even if it slips up. Advancing its timing cycle makes it even better, so no. Just no.

Hellbore at 1600 HE alpha is nuts. It will take out nearly any armor in a single hit(only thing that resists is HA Onslaught, and even that loses most of its armor). 800 shield alpha is also not inconsiderable for any weapon. Sure, you pay in flux and DPS. Doesn't matter in practice because the threat of it will force raised shields even if not fired, and the damage dealt over a long time is less important than the big punch right now(and the HAG can't beat the Hellbore spikes until 15-20 seconds or so).

800 range on energy weapons. Not a good idea. High-tech ships are faster and have better flux stats. An Aurora could nose into HB range, fire off a shot to cause hard flux, and back of to let it vent. The other ship can't safely drop shields because the Aurora can sprint back in with it's high velocity and excellent against armor HB to punish them. And lower efficiency means that the Aurora is penalized if it does anything other than play in an extremely boring way or for a player fighting it, a frustrating one.

All of the medium kinetics except for Heavy Needler: Worthless. Flux efficiency is great. It doesn't compensate for not dealing credible amounts of damage. Soft flux can be dumped with shields up, hard flux from hitting the enemy can't be. Also the tradeoff for 100-200 extra range over LDAC is not worth it for a medium slot and much more OP.
Logged

SafariJohn

  • Admiral
  • *****
  • Posts: 1440
    • View Profile
Re: The Balance Beam
« Reply #92 on: May 01, 2017, 01:25:25 PM »

I want to apologize for my post last night, xenoargh. I was rude and antagonistic. That’s no way for anyone to act.

I think you and I can agree we want Starsector to be balanced, right?

I think we can also agree it’s unproductive to straight-up tell someone, “You’re wrong.” Saying your version of something, weapon stats in this case, is better than someone else’s is a way of saying, “You’re wrong,” isn’t it?


But most of the people around here are pretty reasonable; if you show them one of their things is bad they’ll take a look themselves, accept your proof, and fix it. So that’s probably not why so many people got upset.

I suspect the reason the hornet nest got stirred up is because of a more personal way you told everyone they are wrong. One of the fundamental assumptions of your project here, give or take, is that weapons can be balanced through math alone. The other meaning of that is, “Everyone who thinks you need testing to find weapon balance is wrong.”

After all the work people have put into testing vanilla (and their mods!) to find a good balance, that sounds to me like a mighty big insult. :( I know you don’t intend it that way, but that’s how it comes across.


The work you’ve put into this and your other projects demonstrates your intelligence and dedication, and I’m sure you’ve held your tongue on many of the less-than-nice things you’ve wanted to reply with, particularly last night, so you have pretty admirable patience, too. Smart? Dedicated? Patient? Sounds like a good person to me.

I believe the Balance Beam could be a useful tool in our arsenal, but I don’t think it can be the be-all, end-all of balancing weapons. Would you please give these other smart people’s point of view a second chance?


Note: Typed this up before DeltaV's heck of a post appeared.
Logged

intrinsic_parity

  • Admiral
  • *****
  • Posts: 990
    • View Profile
Re: The Balance Beam
« Reply #93 on: May 01, 2017, 02:27:37 PM »

I think it's worth thinking about the idea behind mathematical modeling for a second. Mathematical models reflect reality, they do not define it. If a mathematical model does not produce results that reflect reality, it is worthless. This how we discover things scientifically. We propose a mathematical model and then check if it's predictions match what actually happens. Then refine the model until it reflects that. In this case, reality is gameplay. The mathematical model is supposed to reflect actual balance in gameplay. The only empirical evidence we have is player experience. That is what gameplay should be (and is) balanced around. If the mathematical model does not produce gameplay that people find to be balanced, then it is not a good enough model. If you propose some equations to define gravity, but when you actually drop an object, it behaves differently than the math would suggest, then the equations are at fault (assuming of course that your experiment is accurate).

I think this process has been slowly happening in this thread. People say 'this model doesn't reflect x well' and then you modify it until it does. But it's worth remembering that player experience is what you are balancing against, and math without context isn't useful (at least from the standpoint of trying to model reality, abstract math is whole different ballgame).

That being said, I agree with the philosophy of this thread; I had a lot of fun in the thread discussing how alpha damage affects armor penetration, however, I think that trying the abstract weapons outside of the context that they are used in misses a lot of game balance. The performance of a weapon depends hugely on the capabilities of the ship it is mounted on. Long range means more on a fast ship, low OP means more on a small ship, low flux/sec is more valuable on a ship with a limited flux pool etc. so if you really want your model to be accurate, you need to consider that.
« Last Edit: May 01, 2017, 02:29:18 PM by intrinsic_parity »
Logged

Ghoti

  • Captain
  • ****
  • Posts: 283
    • View Profile
    • Email
Re: The Balance Beam
« Reply #94 on: May 02, 2017, 10:01:58 AM »

quiet right intrinsic_parity. Newtons law of gravitation is wrong, but the formulas are still handy. The formulas arn't truth, but they can certainly tell us things.
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 4823
  • naively breaking things!
    • View Profile
Re: The Balance Beam
« Reply #95 on: May 03, 2017, 07:00:24 PM »

Actually, I've missed all the noise, folks.  I appreciate the apologies, however heartfelt, but frankly, I feel you folks are largely missing the point.

1.  Are SS weapons balanced well?  No, they aren't.  If we disagree about this, we're disagreeing about everything, so let's start with that and talk about examples:

A.  Is the Heavy AC worth using, if you have Light Needlers?  No, it's not; while it slightly out-performs Light Needlers on per-shot damage vs. Armor, that's literally the only stat where it out-performs them, and that for a huge OP penalty.  When we compare the Light Needlers with the Arbalest, it's even worse.

B.  Is the Light Mortar worth using vs. the Light Assault Gun?  Generally, no; unless you're starved for OPs, the LAG's performance is hugely superior, largely due to higher hit rates and DPS.

These are just two cases where we're comparing weapons in similar job roles, but there are pretty obvious issues just in these areas.

2.  Are my methods valid, mathematically speaking?

The hit-rate stuff is completely obvious trig, so yes, it's valid. 

The math behind TTK is just running the numbers, so yes, it's valid. 

The math behind Damage / Flux is just like TTK; it's clean.

These things are totally defensible. 



So, why all the flak and griping?  Well, let's look at the stuff where the math is arguably bad, and then we'll talk about the other reasons why there's so much griping.

1.  The weighting of damage types is arguable, and we should argue about it, but I think that it's pretty solidly supported by playtesting at this point.  It's probably not perfect, but perfect will have to wait for more testing, frankly. 

SS feels considerably harder over here, in the sense that a lot of the "junk weapons" actually don't feel like junk now.  As that's pretty much what I wanted to see happen, it's feeling valid.

2.  The single-shot damage question is viable and reasonable, but I think it gets weighted far too heavily in a lot of people's minds.

Look, there are two situations where a high-TTK, high-damage-per-event weapon really shines:

A.  When an opponent can be brought to, or over, their Flux Capacity and enter Overload as a result of the hit.  AM Blasters are the classic example here; they're really slow weapons that pack so much punch that they can bring a Frigate to its knees in one shot.

B.  When a single shot can tear through Armor and leave Armor at the hard-capped 15% value.  Here, we're talking about a continuum; when a Light Assault Gun hits a Frigate, it'll tear up Armor but not Hull, when a Heavy Mauler hits, the Armor will generally fail and some Hull damage will happen, and when a Hellbore hits, a large proportion of the damage will get through.

This is great and all, but frankly, it's less-important than people make it out to be, when we're not doing small-scale fights in the Sim. 

In fleet battles, generally speaking, Flux-locking (raising Hard / Soft Flux to the point where ships cannot maintain both their firepower and shields) is, generally speaking, more important.  It's only when Flux-locking has happened that scenario B really starts coming into play; until then, HE weapons in particular are really bad Flux-traders and are actually losers up until that point. 

This is why HE got weighted like it did in the current balance; I feel very strongly that it was inefficient and my fleet compositions pretty much just use Kinetics and Frag in Ballistics, because a fleet that Flux-locks well is punching way above its weight.  This wasn't just my conclusion; it was supported by a lot of us who minmax play.

Now, that doesn't mean that single-shot damage doesn't matter; it does.  But it's quite situational in nature, and therefore, shouldn't be given a huge weighting, if any, especially if the weapon involved has a really poor TTK.  For example, the AM Blaster's great at knocking down Shields and is OK at cracking Armor, but generally speaking, you use it to start a kill, not finish one; other weapons with better TTK and Flux efficiency are used to finish the job. 

What's that situational use-case worth?  Frankly, it's either extremely important or not important at all; a single AM Blaster is just a terrible Flux-trade against an Apogee with full Capacity available, but it can be the death-knell of an Onslaught if it opens up Armor at a critical moment or leads to an Overload that allows bombers to come in. 

This is part of the Fun in a game like this, where we clever humans can feel awesome when we've judged the moment correctly... but in real terms, is it actually way better than the more-boring Flux-trading alternatives, especially when we're talking about AI ships fighting each other? 

Generally speaking, no; it doesn't lead to shorter engagement times (which is about the only fair way to measure this) and, other than the Heavy Mauler, which is absurdly efficient, it's usually a bad idea to even give the AI ships these weapons; humans simply have better judgment about when to burn Time in order to pick the moment when these kinds of mechanics will work best.  If you're in any doubts about that, I invite you to try arming Medusas with AM Blasters and letting the AI pilot them; you'll quickly see that doesn't generally work out.

In conclusion, while I feel like there are certainly valid arguments about that point and others, the core approach isn't bad and the resulting balance isn't terrible, and it largely fixed the obvious problems in Vanilla, which is pretty much the whole point.

But why this, rather than buff / nerf?  Well, two reasons:

A.  Buff / nerf is largely the same process, but without any solid basis.

B.  Buff / nerf tends to largely create statistical outliers that then need more buff / nerf; it doesn't provide any way to talk about the "why" of outlier cases.  This approach does.



Anyhow, that's basically how I think about all this. 

I know it's controversial, because, my goodness, if I ever release a proper mod for and people largely agreed that it improved balance... well, gosh, modded content would also follow suit, and the game would be better as a result... oh, wait, that's not terrible, is it? 

I mean that quite gently: this isn't something mandatory and it's only going to be popular if I'm actually able to deliver; the market doesn't care what I think about my math, it only cares if the results are Fun.

So, I'm going to continue to work on this until I'm either bored to tears of yet-another fine-tuning or until I get enough useful feedback to correct the remaining problems, etc., or I conclude that I've failed. 

If that incenses you, fine, nobody's forcing you to do anything and you can pretend this doesn't exist; it's not like it's hurting anybody here that I'm playing with some numbers, after all.  But for goodness sakes, stop the personal attacks; I get that people don't like their assumptions challenged, but for heaven's sakes, it's not affecting you in any way if you don't download and install it, lol.  Please take deep breaths and move on to something that actually pleases you :)
Logged
Check out my SS projects :)

Thaago

  • Global Moderator
  • Admiral
  • *****
  • Posts: 4235
  • Quantum Mechanic
    • View Profile
    • Email
Re: The Balance Beam
« Reply #96 on: May 03, 2017, 08:33:10 PM »

I'm not going to pull up my quotes from previous, but you know where I stand on both the hit rate and damage/flux - I think you are using the wrong assumptions for hit rate so have bad values, and that using a ratio for damage/flux is fundamentally wrong. TTK is ok, but does not take into account that shields are a regenerating resource while armor is not. I like to think that my feedback on this is the type of criticism you are looking for, but basically, while I think mathematical analysis in general can be applied to SS balance in better ways than it has in the past, the details of what you are doing are not valid. Not to say you might not improve them and make them valid. :)

In my opinion the balance on vanilla weapons is ok. Not 100% good, but also not bad. Of your two examples, I agree that the LAG is very much superior to the mortar and that the mortar could use significant improvement. However, I disagree on the light needlers vs heavy autocannon. The heavy autocannon is a superior choice against anything that it can hit consistently - cruisers, capitals, and slow destroyers. When in close range, as happens when you give the AI an "eliminate" order, it reliably hits destroyers as well. It uses 1 more OP and has a lot more DPS, against shields, armor, and hull. There are situations when you want to use a light needler instead of heavy autocannon, but that is more that the player wants to mount Flak instead in the medium ballistic slot, while still having kinetic firepower. I'll note that this kind of opportunity cost analysis is missing from the model, and that putting it in will by definition include expanding the model to include ship compositions. Which is a lot of work.

The reason why you've gotten so much flak about this project: you've repeatedly promoted the method you are using as the one true way of balancing, implicitly and explicitly disparaging the balance work that others have done, while deliberately ignoring important factors. This wouldn't be a problem if the resulting numbers were solid, but they haven't been: each iteration has had some or most weapons being extremely out of balance, worse than anything in vanilla. That combined with your attitude has yielded some extreme reactions from people who care about the work that they've done. I think a lot of the heat directed at you was out of line, but I'm not surprised to see it.
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 4823
  • naively breaking things!
    • View Profile
Re: The Balance Beam
« Reply #97 on: May 03, 2017, 09:39:00 PM »

Quote
I think you are using the wrong assumptions for hit rate so have bad values
Well, the hit-rate now reflects things that are Destroyer-sized; I think that's as far as we can push that out without it getting silly, in a game where we have very small targets (missiles) up to huge targets (a Paragon with its shield up).  That bar has to get set somewhere, and I think presuming that hitting smaller targets is important is valid, given that since 0.6, I've never seen a point to buying ships larger than Destroyers (other than the Apogee in 0.72).  Combat, even at the large-fleet stage of SS, is largely about killing faster escorts before destroying larger ships, when we're talking about fleet engagements; if you're engaging in Megas-style gameplay, then your argument makes more sense, but I'm aiming for more fleet-style combat, where I expect both sides to have a lot of small stuff.  0.8's emphasis on fighters reinforces this need, imo.

The basic math's just fine, though, if that's what we're arguing about.  It's taking the average accuracy of the weapon (best : worst) which is about the only way to go about it, because we can't balance against either end all the time without absurdity; the first shots from a lot of inaccurate weapons are 100% accurate, but doesn't remotely describe their performance!  

If we wanted to get super-technical, we could do a series comparision, say 50 continuous shots or whatever number seemed reasonable; that would weigh the accuracy values further into the worst-case scenarios.  I think that's unfair to weapons where their accuracy regenerates pretty fast, though; usually, due to all of the bobbing back and forth that happens in fleet-combat scenarios, we're seeing something like mean accuracy, rather than best / worst cases.

Dunno what else to say about this, frankly; if we adjust the size value of the theoretical target up much more, than we have clear absurdities happening, and accuracy matters a lot more than people think it does; when people play, they're missing all the time, and so is the AI; but we don't pay much attention to it because it's a negative event, rather than a positive one, which is natural.

Quote
TTK is ok, but does not take into account that shields are a regenerating resource while armor is not.
TTK isn't where that's been factored; it's factored in Damage / Flux, where the different damage types get their weighting values applied.  TTK is straight damage-to-hull; Damage / Flux is where I'm approximating the value of different damage.  

Could this be significantly improved by doing two values in TTK- one for straight Hull kill, one for Armor?  Probably, but I suspect the impact's not all that significant and can get misinterpreted pretty easily; just because a Hellbore can gut a Frigate on a successful hit doesn't mean I actually fear one when it's mounted on a Mudskipper Mk. II, unless I'm caught Venting or am Overloaded; it's still not doing much Flux damage, frankly.

I think that people tend to forget that, in general, the few weapons this argument applies to, in Vanilla, in a really significant way (we're talking about how long it takes to make a kill here) are all poor Flux traders.  The best of the bunch in Vanilla is the AM Blaster, but it's a special-purpose gun if there ever was one, and it's still trading Flux very poorly vs. Kinetics.  

The idea that armor-stripping weapons significantly shorten TTK is largely wrong; most of the time, by the time an opponent is Flux-locked, they're going to get killed by whatever's hitting them without a response.  Armor-stripping weapons, for the most part, don't do anything but waste time to get opponents Flux-locked.  I know this seems less than sensible if you're trying to solo / alpha enemies, but that's not typical of fleet-combat conditions... and like I said, that's where I'm tuning for, not Rambo-style engagements.  

So armor-stripping weapons pretty much required better performance to be all that competitive, frankly; I'm unapologetic about it and I don't really mind if that made, say, a strip-and-kill Dominator build a little better; I'm not tuning for that.

Anyhow, I"ll take the Armor-sim vs. one block case and I'll try it out.  I'm not predicting that the results will change my opinion much, but maybe I'm guilty of just not weighting it enough when we cross some threshold, like how the Pulse Laser became very effective in 0.72 once Captain bonuses stacked up, because it was suddenly in the top group of Armor-killers that were efficient.

Quote
I disagree on the light needlers vs heavy autocannon. The heavy autocannon is a superior choice against anything that it can hit consistently - cruisers, capitals, and slow destroyers.
I really didn't find this to be true in Vanilla, largely because of the efficiency of Light Needlers, in terms of Damage / Flux; they're simply better at Flux-locking, and that was what mattered most.  The Heavy AC's inaccuracy is a significant factor as well.  In 0.72, this was more clear-cut than it is in 0.8, to be sure; the rebalanced Hammerhead makes a stronger case for the use of the Heavy AC, because it finally has a stat-line that can use it well, and the Enforcer's no longer able to use three Light Needlers and a pair of Flaks and be one of the best solutions to all problems, because of its speed and lack of fixes through Captain buffs to fix it.  But that's ship balance / skill balance creating a case, which is a bass-ackwards way to balance things.  Start with the simple stuff first, get it right, then move to the more complex cases.

Quote
using a ratio for damage/flux is fundamentally wrong
It's about the only way to reliably deal with Vanilla's basic schema, where these things are close to 1:1 except when they're not, because <reasons>.  I'm totally open to another way to handle that and would welcome another way to deal with it mathematically, but in the end, it's pretty much why it's there in the final equation.

Pretty much everything in Energy that deals straight Hard Flux works that way, the Light ACs were set that way and the combat Beams worked that way until recently (changing the Gravitons to Kinetic, for example, was a big sideways buff).  All of the initial complaining was because these things weren't true, too, which I took to heart and made a priority; it's clear to me that that relationship is a major part of the feel, to most players.

We're simply not going to agree on the idea that OPs should represent Vents not taken; this is less-true now, with the expansion of OPs for ships to allow for more Hull Mods, than it ever has been.  Back in the early balance of SS, I think your point was considerably more valid, but the game's changed a lot; OPs can get used in a lot of ways that all can be sensible.  

Part of this disagreement is that, at the end of the day, we need a yardstick.  Either we count Flux as the "currency" of weapons or we count OPs; in game-design terms, those are the two counters players "spend" to achieve a result.  Both need to be pretty important factors, but which should be king?  I think OPs should be, because choosing one weapon over another, especially in cases like the Heavy AC vs. Light Needlers, is a choice of, "one of these things is not only really efficient, is relatively accurate, is in the same range band, but I can use the OP difference to buy more Vents or Hull Mods"- which is why it's a prime balance case and why I feel the Heavy ACs are a bad deal in most situations.

Quote
This wouldn't be a problem if the resulting numbers were solid, but they haven't been: each iteration has had some or most weapons being extremely out of balance, worse than anything in vanilla.
Can you provide a good example of that from the current numbers?  I'm generally feeling that this has been pretty close to the mark, when playing over here.

Quote
you've repeatedly promoted the method you are using as the one true way of balancing, implicitly and explicitly disparaging the balance work that others have done
I'm not backing away from that.  Playing modded SS is like watching a train-wreck of bad game balance; modders tend to either over-nerf their stuff or have balance that's both over-powered and under-powered in the same mod.  Yet everybody acts like this is some mystical process, when it's not; they're simply responding to feedback, like Alex is, just with a smaller sample size.  None of this looks like serious attempts to grapple with the issues; weapons aren't all that complicated in this game and we shouldn't just say, "it can't be done".  

But lo, somebody tries to get it done and gets flamed constantly rather than getting playtesting results, even though, for those of you who haven't been here very long, we used to have a ship-balancing tool, even, that people used to use to get to rough balance (until it got broken by CSV changes and no maintenance) which wasn't even controversial at the time.  But one of the people who's been playing SS forever, has built more modded content than practically anybody, wrote the only general-purpose AI, etc., etc., etc. does this, and boom, fires and lamentations.    

It's a little disappointing and frustrating, but oh well, I'll eventually get there regardless, if I don't get bored first.
« Last Edit: May 03, 2017, 10:04:44 PM by xenoargh »
Logged
Check out my SS projects :)

Dark.Revenant

  • Admiral
  • *****
  • Posts: 2539
    • View Profile
    • Sc2Mafia
    • Email
Re: The Balance Beam
« Reply #98 on: May 04, 2017, 11:26:38 AM »

The Heavy AC actually does significantly more damage to hull than a Light Needler due to the armor change; if you were talking about 0.7.2a I'd agree with you about that weapon though.

Many mods have sketchy balance decisions; Scy and Diable weapons generally trend to the weak while BRDY and Templars generally trend strong; this is clear with plain observation.  THI and II seem about on par with vanilla to me.  SWP is more complicated because several SWP weapons are actually intended to be junk weapons, while others are intended to fill a niche (and not be terribly powerful outside of that niche).

I use math when coming up with first draft weapon balance.  Hell, I use math for many things including weapon/ship price.  But it's only a first draft; too many unmodelable issues prevent a spreadsheet from being the end-all-be-all.  If a weapon deviates from "the math", that merely means the creator should convincingly justify the deviation, not that the weapon must change.  It's like the MISRA standards; blindingly following them results in extreme tedium; true understanding lets you deviate at the right time to make something better.
Logged

Wyvern

  • Admiral
  • *****
  • Posts: 2412
    • View Profile
Re: The Balance Beam
« Reply #99 on: May 04, 2017, 12:02:44 PM »

On armor and time-to-kill:

Yes, you're right, that flux-locking low-per-hit-damage high-dps kinetic weapons are generally superior to high-per-hit low-dps HE weapons, and that if you have to choose to field solely one or the other it would be better to go with the former.

Unfortunately for your calculations, that's not the choice a player faces.  And if, say, you've installed two heavy needlers on an Eagle, that drastically changes the value of choices for that Eagle's third medium ballistic slot; adding a third heavy needler won't improve the kill time much, and will allow your targets to safely drop shields and absorb hits on armor (giving them a major flux advantage for hitting you back).  Adding a mauler, by contrast, will dramatically improve kill times, as well as making it much more likely that your needlers will actually win the flux war as intended.

In other words, weapons need to be examined in context, not just on a one-for-one comparison.  Having a mostly-kinetic-damage loadout with one high-per-hit weapon for stripping armor is a vastly more potent configuration than going for exclusively kinetic damage, and the high-per-hit weapons need to be valued on that basis.
« Last Edit: May 05, 2017, 10:08:45 AM by Wyvern »
Logged
Wyvern is 100% correct about the math.

TaLaR

  • Admiral
  • *****
  • Posts: 1909
    • View Profile
Re: The Balance Beam
« Reply #100 on: May 05, 2017, 01:25:57 AM »

Unfortunately for your calculations, that's not the choice a player faces.  And if, say, you've installed two heavy needlers on on Eagle, that drastically changes the value of choices for that Eagle's third medium ballistic slot; adding a third heavy needler won't improve the kill time much, and will allow your targets to safely drop shields and absorb hits on armor (giving them a major flux advantage for hitting you back).  Adding a mauler, by contrast, will dramatically improve kill times, as well as making it much more likely that your needlers will actually win the flux war as intended.

Going further, replacing one of Heavy Needlers with Heavy AC can be useful too.
Needlers fire in synchronized bursts and even when Mauler is involved can be shield flickered/phase skimmed/fortress shielded somewhat efficiently (by player, AI is not that good most of the time). Having to deal with 3 different attack patterns makes mitigating them even more difficult.
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 4823
  • naively breaking things!
    • View Profile
Re: The Balance Beam
« Reply #101 on: May 05, 2017, 04:15:51 PM »

All right, a new build's up, factoring in high-alpha / Armor-TTK weapons.

After looking at these issues for a bit, I finally figured out a formula that I felt punished this stuff fairly.  It's a tricky issue, frankly; some weapons excel in this area but suck in all others, etc., etc., so coming up with something that didn't push the numbers wildly out of whack was a bit harder than expected. 



Essentially, though, the basic formula for Armor TTK is just a simple solution vs. one block of Armor, where we're measuring shots-to-kill or time-to-kill for Beams.  So the basic formula there is:

=10000/(base_ttk*if(P7="KINETIC",0.5,1)*if(P7="HIGH_EXPLOSIVE",2,1)*if(P7="FRAGMENTATION",0.25,1)*if(P7="ENERGY",1,1))

Base_ttk is computed elsewhere, and includes burst behaviors, etc., etc.

The tricky bit was deciding how to apply the weighting to the final numbers.  I decided that, in the end, it works best as an additive:

+max((max(300/armor_ttk,1)*0.002),0)

What's the 300?  The baseline Armor TTK, in seconds, for a LMG vs. 10K Armor, ideal, one-block case.  Just in case you're wondering.

Net results were mainly pretty modest, as this mainly hits a few egregious cases, but this nerfed:

Hellbore, AM Blaster, Mining Blaster, Mjolnir, most of HE to some extent, the Arbalest.

Not sure about these nerfs, to be really honest.  I feel pretty strongly that Armor-kill is pretty unimportant, frankly, because Flux-lock means your opponent dies eventually no matter what you're hitting with, and Armor-killers just save time at the very end of the kill.

I rarely see anything less than a Capital survive lengthy Flux-locking without having to get out of range of opponents, so it's kind of moot; if you're Flux-locked and can't leave the combat, you're dying slowly.  Players tend to handle this by ducking out of the fight and Venting, if possible... the AI tends to handle this by dropping shields a lot, which sometimes works, sometimes means it's just getting chipped away.
Logged
Check out my SS projects :)

Axisoflint

  • Ensign
  • *
  • Posts: 12
    • View Profile
    • Email
Re: The Balance Beam
« Reply #102 on: May 06, 2017, 03:43:02 PM »

On armor and time-to-kill:

Yes, you're right, that flux-locking low-per-hit-damage high-dps kinetic weapons are generally superior to high-per-hit low-dps HE weapons, and that if you have to choose to field solely one or the other it would be better to go with the former.

Unfortunately for your calculations, that's not the choice a player faces.  And if, say, you've installed two heavy needlers on an Eagle, that drastically changes the value of choices for that Eagle's third medium ballistic slot; adding a third heavy needler won't improve the kill time much, and will allow your targets to safely drop shields and absorb hits on armor (giving them a major flux advantage for hitting you back).  Adding a mauler, by contrast, will dramatically improve kill times, as well as making it much more likely that your needlers will actually win the flux war as intended.

In other words, weapons need to be examined in context, not just on a one-for-one comparison.  Having a mostly-kinetic-damage loadout with one high-per-hit weapon for stripping armor is a vastly more potent configuration than going for exclusively kinetic damage, and the high-per-hit weapons need to be valued on that basis.

I'm not a modder or in any way someone who looks at the mathematics of SS seriously. I just play the game. I'd just like to chip in that this comment however sums up how I feel about this thread.

Balancing is really hard when the variables are so extensive. The sheer fact alone that there are multiple weapon slots on (nearly?) every ship means that you essentially can't balance one weapon against another, because what happens when you put one of each on a ship? Or three different weapons? Or two ballistic and two missile?

I don't personally see how you can math out the balance between weapons when you need to resolve putting them on different ships with varying armour values, different special abilities and different weapon loadouts. I saw earlier that you said that you should balance ships afterwards, but what will that involve? Removing all but one hardpoint from every ship?

I think questioning things is a great idea and I'm pretty sure nearly every single person on this forum (esp the modding subsection) is interested in balancing things well and making sure it's a challenging yet rewarding game but I don't see how that's possible without taking real actual in game data and using it in the process. There's a reason that mod developers release updates to their mods with adjustments to weapon/ship values - when lots of people play them, they find the flaws/exploitable quirks of the mod.

It's an ongoing process and one I can't see fixed with a spreadsheet.

Ymmv. It's been an interesting discussion whatever happens :)
Logged

Death_Silence_66

  • Commander
  • ***
  • Posts: 241
    • View Profile
Re: The Balance Beam
« Reply #103 on: May 25, 2017, 03:17:32 AM »

Bit late to the game here, but I have a bit of criticism.

The Heavy autocannon seems anemic. It is only marginally superior to the arbalest autocannon in dps and slightly superior in range while being nearly useless against armor.
The Thumpers don't fire, which seems to be a bug.

Also, glad to see the storm needler has a ROF worthy of its name.
« Last Edit: May 25, 2017, 03:23:19 AM by Death_Silence_66 »
Logged

xenoargh

  • Admiral
  • *****
  • Posts: 4823
  • naively breaking things!
    • View Profile
Re: The Balance Beam
« Reply #104 on: May 25, 2017, 01:42:22 PM »

I'll take a look at the Heavy AC; it's roughly the same DPS it was in Vanilla, but I can re-jigger a bit.  Not sure what's up with the Thumper; it works all right here, but I may be guilty of not updating the csv.  I'll update it shortly.
Logged
Check out my SS projects :)
Pages: 1 ... 5 6 [7] 8