65 replies to this topic

[MustrumRidcully]

Sprouticus, on 07 November 2012 - 09:23 AM, said:

so you did math but didnt actually test the game?

Do you need to test the game to show that if you have 1 Jenner, and another Jenner appears, there are now 2 Jenners?
Do you need to test the game to show that if you have two weapons that deal 5 damage each, that them together will deal 10 damage?

Super Mono, on 07 November 2012 - 09:26 AM, said:

What do you know, when you directly implement only half of a game's rules and systems it throws the whole thing out of whack..

I may have predicted some such, maybe even without playing the game and just using math. But I am not sure my methods would be trusted by anyone.

[Vapor Trail]

Sprouticus, on 07 November 2012 - 09:23 AM, said:

so you did math but didnt actually test the game?

Vapor Trail, on 07 November 2012 - 06:23 AM, said:

First off, this is going to be highly theoretical using the weapons numbers from BEFORE yesterday's patch. It may not be an accurate picture of where the weapons are currently.

So you posted in the thread but didn't actually read it?

[Super Mono]

Sprouticus, on 07 November 2012 - 09:13 AM, said:

2) PGI has to balance more than just right now. They have new mechs coming out, and those new mechs will allow configs we cant look at right now.

They're directly implementing Mechs and variants from the TT into the game. If they implement anything that was designed around 2.0 DHS it will most likely be broken and useless as their loadout was designed around increased heat dissipation.

[Sprouticus]

MustrumRidcully, on 07 November 2012 - 09:29 AM, said:

Do you need to test the game to show that if you have 1 Jenner, and another Jenner appears, there are now 2 Jenners?
Do you need to test the game to show that if you have two weapons that deal 5 damage each, that them together will deal 10 damage?


I may have predicted some such, maybe even without playing the game and just using math. But I am not sure my methods would be trusted by anyone.

When you are ONLY using math to evaluate the effect of a change on a game, you may be missing something. As a point in fact he is missing stuff because his assumptions are wrong. But Im not going to go into that argument, I explained it in other posts last week when you were all slitting your wrists.

Heat is FINE with 1.4. Really, it is. I'd be ok raising it to 1.5 or even 1.6, but with the expert XP bonus no higher.


Unlike the canon TT lovers, i am willing to accept changes from TT if it serves the game better and is up in the air due to conversion from TT to live actin video. And part of that is messing with the heat/dmg/RoF system. It is simply required.

PGI wants SHS to have a place. The only way to do that is make a niche for them. That is what they are doing, like it or not.

Edited by Sprouticus, 07 November 2012 - 09:44 AM.

[Super Mono]

Sprouticus, on 07 November 2012 - 09:43 AM, said:

When you are ONLY using math to evaluate the effect of a change on a game, you may be missing something. As a point in fact he is missing stuff because his assumptions are wrong. But Im not going to go into that argument, I explained it in other posts last week when you were all slitting your wrists.

Heat is FINE with 1.4. Really, it is. I'd be ok raising it to 1.5 or even 1.6, but with the expert XP bonus no higher.


Unlike the canon TT lovers, i am willing to accept changes from TT if it serves the game better and is up in the air due to conversion from TT to live actin video. And part of that is messing with the heat/dmg/RoF system. It is simply required.

PGI wants SHS to have a place. The only way to do that is make a niche for them. That is what they are doing, like it or not.

It's not creating a niche, it's creating DHS that only benefit small mechs boating S/M lasers.

[Targetloc]

Every balance move PGI has made reminds me of 12 years ago when I was helping to balance a MUX combat system (not Btech unfortunately).

I spent months tweaking numbers and formulas for stats and attacks, doing hundreds of iterations of 'test battles' using test copies of real characters. Things were slowly getting better, but every couple months a player would stumble on a new balance problem that hadn't come up before and I'd have to rework stuff.


And then, I took a stats-calculus class and one of the first lessons was about dice probability distributions. And the light went on. And then modeling probabilities. And finding intersections in linear equations. Means, standard deviations and upper and lower limits. Ding. Ding. Ding.


Going back to the game I realized every balance change I'd made up to that point was technically in the right direction, was supported by data, and occasionally right on the mark... but was essentially throwing darts blindfolded. The data and play-testing was at least facing me in the right direction and telling me 'warmer' or 'colder' on the hits, but the odds of actually hitting the bullseye without a few hundred iterations was astronomical.


Suddenly, I was able to find the exact swing point where one stat would be situationally better than the other. I could find the average case, and the absolute performance bounds using the best case and the worst case (hello min/maxers).

And I also found that some of my tweaks, while making things a little better, would have NEVER fixed the problem, because the core formulas gave some of the stats completely different rates of growth from the others, which made it mathematically impossible to balance both the middle and the extremes.

In the end I had to completely re-write a few functions, but I was able to precisely balance all builds with equal point investments, without making every option the exact same widget with a different name. Oh and pro-tip, don't say something is 'situationally balanced' against something else, unless you can precisely quantify how often that situation occurs, and whether the value/time of that advantage is anywhere near the value/time of the disadvantage outside that specific situation.


Anyways... PGI, please... every time I see you nudging the damage up on Large Lasers by 1 point (without touching pulse lasers or PPCs) or increasing the ROF on AC2's without looking at their heat, or adjusting heat sinks in any direction but up... I can hear the darts hitting the wall. Look hard at the math. Look at the beautiful graphs your concerned players are posting. Model potential weapon or heat changes against each other.


PS, look at your heat system.

[MustrumRidcully]

Sprouticus, on 07 November 2012 - 09:43 AM, said:

When you are ONLY using math to evaluate the effect of a change on a game, you may be missing something. As a point in fact he is missing stuff because his assumptions are wrong. But Im not going to go into that argument, I explained it in other posts last week when you were all slitting your wrists.

I am not "only" using math, actually.

I also observe the game and the trends. I play my mechs. I observe things. Than I create models of the situation and see if if actually fits. I am prepared for unexpected outcomes.

For example, I was surprised how effective small and medium lasers actually were, both if we calculculated heat neutrality and TET efficiency. I was also surprise that the Gauss superiority was confirmed, but the Ultra AC/5 was actually better in that area, as was the AC/5 and LB-10X AC. I realized that the only thing that is keeping these weapons down may be that single shot value needs to be given more weight (or atleast analyzed after the fact), and that the shotgun behavior of the LB-10 X AC may be the only thing that is holding its down. If we ever get regular AC ammo for the LBX, we may see it being preferred by many mechs, since it requires much less trade-offs than installing the Gauss Rifle on a 65 ton mech.

[Vapor Trail]

DHS aren't supposed to be niche equipment. They're supposed to be the new standard, relegating Standard heat sinks to "niche" status.

[TheCerberus]

One problem. Critical slot space. Double heat sinks, even working at "true" double efficiency, are less efficient for space. Combined with endo steel and ferro fibrous it gets pretty hard to fit any in a 'mech outside of the engine. Add to that that you can't fit double heat sinks in the legs or CT.
So yes, if you abstract everything and just look at the DPS (I lost track of all the acronyms there but I mean DPS at 0 net heat) "true" double heat sinks will be a game changer. But in practical terms things are not so simple. Also, as vapor trail said, they should be better because they are better technology and should be a favoured (if expensive) upgrade just as endo steel is universally better than regular structure.

When the clans invade and they bring their kickass double heat sinks... well then we're going to have a mass of laser-shooting clanners blowing us all to hell. But I'm kind of ok with that, clanners are heavy users of energy weapons and it's fitting that they are the most efficient option.

Edited by TheCerberus, 07 November 2012 - 01:16 PM.

[BA Dillard]

Matist, on 07 November 2012 - 09:14 AM, said:

I think his point was that it doesn't really matter how you tweak the 1.4 number since small and medium lasers are out of whack and therefore throw everything else off.

It's really obvious in play how much better normal lasers are over pulse variants and again his graphs show that off perfectly.

I guess I don't understand what it is you guys want. Do you want it to just be double like in TT and leave it at that? Those graphs make my head hurt. I can't read them without my brain shutting down. lol. Can you explain for the non scientists here?

[VictimEN]

TheCerberus, on 07 November 2012 - 01:14 PM, said:

When the clans invade and they bring their kickass double heat sinks... well then we're going to have a mass of laser-shooting clanners blowing us all to hell. But I'm kind of ok with that, clanners are heavy users of energy weapons and it's fitting that they are the most efficient option.

OTOH, Clanners also got missiles at half weight.

[Tuhalu]

BA Dillard, on 07 November 2012 - 03:23 PM, said:

I guess I don't understand what it is you guys want. Do you want it to just be double like in TT and leave it at that? Those graphs make my head hurt. I can't read them without my brain shutting down. lol. Can you explain for the non scientists here?

This thread isn't actually about whether or not double heat sinks should be double. It's a thread about how some weapons are favored more by having less heat dissipation available on mechs when you up their fire rate.

[BlackSquirrel]

Hmm strong work...

However I would argue make "boating" harder

by either 1.decreasing the HP of small and medium lasers. Upping the hp on large etc.
2. The more lasers fire from a part on a mech heat will build up higher in said component than it would if they were spread.

Specifically this addresses the feared Hunchback build with an Xl engine. Or annoying jenners with 4-6 small lasers. Or Awesome (Though the 7 lasers are fairly spread out in those variants)

Also maps (until LRM spam) were fairly in favor of short range combat given how zoom, and map layouts are.

Honestly seems a little over zealous to nerf DHS because of 2 weapons systems.

Edited by BlackSquirrel, 07 November 2012 - 05:30 PM.

[TheMightyWashburn]

Zyllos, on 07 November 2012 - 07:23 AM, said:

Extremely informative post. It also shows that devs are going wrongly with the DHS.

What should be happening is weapons that are becoming completely crazy with DHS will just need minor adjustments. DHS themselves should only provide the same maximum heat benefits of SHS. Then, there needs to be penalties for even having heat built up, like slowing of the mech, slowing torso turning, slowing arm movement, reduced damage with energy weapons, slowing lock of locking weapons, slower RoF for direct fire weapons, internal damage before reaching 100%, allow internal damage from heat to have a chance to cause a critical hit.

With these changes, then you can allow DHS to be 0.2/per and reduce the maximum back to the same as SHS. This allows for SHS to have larger maximum, thus more heat before receiving penalties but dissipation is less efficient.

This is stupid as hell. Lasers run hot, THEY HAVE TO. Derp

[Targetloc]

TheMightyWashburn, on 07 November 2012 - 06:05 PM, said:

This is stupid as hell. Lasers run hot, THEY HAVE TO. Derp

What if I told you... there's a spectrum between 'not hot enough' and 'too hot' where lasers are not OP?

[Vermaxx]

The change in DHS solved the issues they wanted to solve.

They are flat better for most light and medium mechs, especially when you add in the pilot xp bonuses. Therefore, most people will use them and spend CB on the upgrade.

They are arguably better for many heavy mechs. The brawlers, the boaters, the non large weapon crowd.

They are generally worse for the heavy and assault mechs carrying a lot of large type weapons. However, this is not terrible because they are not horrendously overpowered for light/medium mechs.

The solution is incredibly simple. Add a pilot specialization to the xp system. Players can pick which role to devote themselves to, and that 'class' would give them unique bonuses.

Like increased heat function when in heavy/assault mechs. Bam. Suddenly the people who generally 'need' tabletop doubles get it, while the light and medium pilots get something equally fitting, like faster target identification, or TAG that lasts longer.

[Targetloc]

I was always in favour of engine heatsinks not counting as doubles. Only the ones outside the engine that actually take up 3 critical slots getting the benefit.

Even in TT, double engine heatsinks were just not balanced. They were magic and 'free' and led to a lot of munchy builds. The only downside is that a lot stock designs from 3055 onward count of the free 20 heatsinks to work.

[lshtaria]

DHS at 1.4 is probably the best all round trade off I think. Currently beasting with a 9x mlas Hunch and its 45 point alpha is nothing short of devastating on ANY mech, Atlas included. 2 rear ST alphas on an XL Atlas and it's goodnight Vienna. A stationary Cicada gets insta-legged.

A total of 18 DHS allow me to get in more heat dissipation and despite going up to 50-60% on the threshold, it dissipates quick enough to get in 3 straight alphas. It's also heat neutral when chain firing.

I still have to be tactical and manage my heat well. If DHS were any higher than 1.4 then lasers would be beyond ridiculous.

[Targetloc]

Yeah, but setting them to 1.4 still means there's never any reason NOT to have Doubles on any mech that only uses 10-14 Heatsinks. The upgrade cost is trivial and there's no design trade-off when you can fit them all in the engine. Even 1-2 outside is trivial to find space for.

This just further pushes the balance towards low heat weapons. If you were already running a build that operated optimally (not heat neutral, but decently, like a Gauss+2ML HBK) DHS are a 40% buff for a 1 time 500,000 cbill investment and no other trade-off.



addendum: Double Heatsinks shouldn't have their dissipation nerfed because lasers are too good. DHS need their mechanics adjusted because they were never balanced in TT, and in some builds they have absolutely no trade-off, unlike Endo Steel or FF (which always take extra slots).

Lasers need to be looked at separately. (ie, their ROF is too high, but the only thing holding them back is the game's abysmal heat dissipation, which leads to a huge performance gap between their 'practical' and 'heat neutral' DPS)

Edited by Targetloc, 07 November 2012 - 08:37 PM.

[Draco Argentum]

I don't mind the use of partial ammo tons. In theory you can mount two guns and feed them off the same ammo bin, hardpoints permitting of course.

Your conclusions are wrong though, or at least not supported by the graphs.

Any weapon whose line is always under another's is bad. Having the SL actually be better than the longer ranged weapons when you're up close is a requirement. If it wasn't it'd suck, just like the AC20. Your graphs do an excellent job of showing why pulse lasers and PPCs are bad for the same reason. The presentation of those graphs really drives home which weapons should not be on your mech.