HAWKEN servers are up and our latest minor update is live!
Forgot Password? SUPPORT REDEEM CODE

Jump to content


An in-depth look at optimizations and internals


  • Please log in to reply
6 replies to this topic

#1 Psylo

Psylo

    Newbie

  • Members
  • Pip
  • 4 posts

Posted November 23 2012 - 06:58 AM

First of all, I want to talk about balance before anything else. True game balance is the concept that, in a fight between two players, the more skilled player would win more often REGARDLESS of mech, weapon, item, and optimization choices. The problem is that true game balance is impossible - it cannot be done. Period. Done. There will always be tier lists, broken weapons, and the ilk. This is important to remember; it's a design philosophy and mindful objective, but not ultimately possible.

That said, and understanding that we're still in a closed beta, I want to talk about the optimizations. Specifically, what the current problems with them are, why they need to be fixed, and my suggestions for possible fixes. I'm also going to be talking about stats as a general concept, because it applies to things such as internals as well and weapons and items, so it's all interrelated. With that, let's begin :)

The Problems

Talent trees are nothing new to games - they allow for customization between like characters. Nothing new, and Hawken has fairly standard fare as far as the system goes. You invest points, get more specialized play, and life is great.

However, in Hawken, my biggest critique is that the talent trees are unnecessary. This first and biggest reason for this is based on how little they affect game play overall. The offense tree is an excellent example of this:


If I invest as many points into the offense tree as possible, for the sake of increasing the damage of my Grenadier's Rev-GL, I will have a damage percentage increase of 2.5% rounded up. We'll ignore my increased rate of fire for now. If I give my Grenadier explosive munitions, which alone gives me twice as much damage per shot, then I get a 7.5% increase. In the end, every shot of my Rev-GL now will do 75.25 damage instead of just 70.

The Rev's firing speed is 1.2 shots per second. The increase in firing speed from the offense tree is 1% or an additional .012 attacks per second. If we multiply out the before and after effects of the weapon specializations, we get the following:

DPS with zero bonuses - 84
DPS with all possible bonuses -   90.3 (rounded)

Total difference = 6.3 points of damage

So, if you trust me to do my math correctly, here's a fun table for you to look at showing you how much the dps is increased per weapon in a full offense build:

TOW    DPS difference: ~4.1dps
Submachine DPS difference: ~6.5dps
Assault   DPS difference: ~7.5dps
Grenade L. DPS difference: ~3.3dps
HEAT   DPS difference: ~6dps
Mini Flak DPS difference: ~12dps
Flak DPS difference: ~11.7dps
Sabot DPS difference: ~9.4dps
Hakwens   DPS difference: ~10.3dps
Slug    DPS difference: ~5.6dps
Hellfire   DPS difference: ~3.2dps
Vulcan    DPS difference: ~7.5dps
Seeker   DPS difference: ~6.1dps
EOC   DPS difference: ~3.9dps
Rev DPS difference: ~6.3dps

(Note that these are DPS values, not per shot damage values. Because you're shooting individual projectiles, damage comes in spikes rather than a constant stream. In other words, these are averages over time)

These increases take into account everything - every possible way you could increase damage will give you these results. The effect is that your target is more liable to die from falling off a six foot ledge, splash damage from their own weapon, or the turrets on a ship than from the extra damage increase. Only one in every 100-150 kills, those times when you escaped with literally less than 15 health, would the extra damage have made any noticeable difference.

So, as far as damage goes, a mech with full offense and a mech with zero offense have a nearly identical chance to take out any given target. If you look at armor, the same results appear. Weapon cooling. Fuel regen. Damage reduction. Physical mech movement is the ONLY specialization that has any practical effect on real combat situations. Things such as dodge cool down reduction and increased thrust speed can be invaluable, but alas that's not the sole goal behind a spec system.

Conclusion: Optimizations and internals affect your mech in ways so small that if they were completely removed from the game, hardly anything would change.

So, the question is this - why have them in the first place? If they do next to nothing, why have methods of customizing a mech with different optimizations and internals?

Because it's fun to play with, and that's why we're here :)

The Solutions

Like I said, the point of customization is to make your mech feel different and unique. That's fun, and I feel Hawken is a game that benefits from that. So, on my own and with help from the Black Steel Brotherhood clan, we've been able to come up with some ideas for fixing the problem. Enjoy:
  • Potency: There isn't anything wrong with increasing some of the values in the current system. Each spec is divided into five tiers - give the first, third, and final tiers some progressively hefty optimizations that make investing in them worth the effort.
  • Dependencies: There are certain optimizations that have prerequisites, which is fine. It would be interesting to make some lengthy, well-thought-out chains of prerequisites that result in major optimizations. An example would be, in the movement tree, by investing in fuel regen you can then invest in larger fuel tanks, then make dodges cost less fuel, and finally increase flying speed.
  • Internals only: First, give mechs the ability to have around seven or eight internals; second, make them only purchasable with hawken points; finally, scrap the optimizations altogether. This way, in a realistic fashion, players could buy a wealth of different interchangeable parts for their mechs.
  • Stat-independent: Take away the stats idea and make optimizations and internals into single, situational effects. That is, make optimizations that increase your max altitude, the number of turrets you can carry, or give you the ability to auto-spot enemies you see rather than increases in dps, health, or other 'stats'.
  • Class-unique: In theory A, B, and C class mechs attract different kinds of players - so give each mech class it's own type of optimization tree. C classes should not be dashing around the map like sonic, nor would I want to if I was that large. Instead, give my C class a dedicated tree where I could choose to invest in my siege mode abilities, health amounts, or damage prevention. Same for other classes.
  • Body Shop: As it stands, changing out the chassis of your mech is purely cosmetic, but it could have an effect on which optimizations and internals are available to you. The bonus of this is that it also affects readability, so that when you see some A class mech running around you can know just by its body type if it has the potential to be a major threat or not.
  • In-game: Make optimizations an in-game only aspect that resets every game, akin to League of Legends. Because roles and team composition matter in a game like this, additional optimization/item slots could be unlocked on a time based standard to make up for those who are playing objectives rather than getting kills and xp. It would also increase the dynamism of games in real time.
Those are some of the better ones we've come up with. If you have any more, please post them below. But all in all, the thing I want to stress, is that I want some kind of meaningful customization of my mech that AFFECTS gameplay. That's the point customizing a character in the first place - so it plays differently. Hawken is in the highly unique position that it's an indie project that doesn't need to adhere to the rigid mediocrity of AAA titles - the devs can try crazy stuff and as a good community that loves this game we'll give feedback.

I would love to see some working of the Hawken optimization system that has never been done before because, and this is why I love this game, the Hawken concept has never been done quite this this before either :)

Edited by Psylo, November 23 2012 - 07:38 AM.


#2 Etan

Etan

    Advanced Member

  • Members
  • PipPipPip
  • 367 posts
  • LocationMInnesota

Posted November 23 2012 - 07:02 AM

Realizing how low the values really are is quite depressing xD. But i would love to see what the devs come up with and if any of these ideas make it in =D
Posted Image

#3 h0B0

h0B0

    Non Sequitur Leprechaun

  • Full Members
  • PipPipPip
  • 4,143 posts
  • Location[delete for trolling] --defter

Posted November 23 2012 - 07:50 AM

While i really enjoy your post and agry with most of it. I believe something very important has been left out: the fact that the current optimization trees were severely nerfed because people complained in the difference of power between a lvl 0 and 20 mech.
Making the optimization trees more usefull again might bring the same kind of resentment of the game back.

that being said.

View PostPsylo, on November 23 2012 - 06:58 AM, said:

  • Class-unique: In theory A, B, and C class mechs attract different kinds of players - so give each mech class it's own type of optimization tree. C classes should not be dashing around the map like sonic, nor would I want to if I was that large. Instead, give my C class a dedicated tree where I could choose to invest in my siege mode abilities, health amounts, or damage prevention. Same for other classes.
  • Body Shop: As it stands, changing out the chassis of your mech is purely cosmetic, but it could have an effect on which optimizations and internals are available to you. The bonus of this is that it also affects readability, so that when you see some A class mech running around you can know just by its body type if it has the potential to be a major threat or not.
  • In-game: Make optimizations an in-game only aspect that resets every game, akin to League of Legends. Because roles and team composition matter in a game like this, additional optimization/item slots could be unlocked on a time based standard to make up for those who are playing objectives rather than getting kills and xp. It would also increase the dynamism of games in real time.

Class-unique i very much agree with the idea of having class unique trees but even tho i do not play c-class very often i could not ever imagine playing it with less then max points in the movement tree. Denying C-class the ability to increase its mobility or A-class' ability to increase its survivability could discourage people from playing outside of their comfort  mech class zone.

Body shop: I do love this idea but i doubt it will ever happen. since cosmetics are the main source of income for MTR/ADH. And obviously we do not want any stats attached to items purchaseable with real money.

In game: Although some game modes do feel like mobas i do not feel like unlocking new abilities or stats midgame is compatible with the feel of the game. Not to mention one would have to die in order to spec his newly acquired optimization.

Keep up the good work.

Click me! I dare you.

Posted Image

View Post[HWK]HUGHES, on March 15 2013 - 08:35 PM, said:

Oh don't always listen to h0B0. Lol.


#4 Etan

Etan

    Advanced Member

  • Members
  • PipPipPip
  • 367 posts
  • LocationMInnesota

Posted November 23 2012 - 08:25 AM

As long as there is a level people not at that level will complain about it.


EDIT- With the body shop i think cosmetics would still be there but you would essientally create more internals that would have small effects on apperence. This would allow them to continue making money while giving us mech customization.

At the respecing during a game thing after dying we essientally do this with switching mechs during a match. And to stat it out we have to go to the garage anyways. Example is extra EU containers for your mech

Edited by Etan, November 23 2012 - 08:29 AM.

Posted Image

#5 Psylo

Psylo

    Newbie

  • Members
  • Pip
  • 4 posts

Posted November 23 2012 - 08:43 AM

I'm actually aware that the trees were nerfed because of complaints about the difference between level 20 and level 0 mechs, and I feel kind of silly that I didn't address it in hindsight. But the thought of it is this: If there is no difference between a level 20 and a level 0 mech, then why are there levels at all?

I've played a lot of battlefield 3, battlefield in general, and the truth is you get owned by the higher ranks for quite a while before you get some gear under your belt. But once you do, it feels great. If you really want to balance veteran and newer players by their levels, then you section it off so that only level 15+ can be matched with level 15+. That would be the logical solution. Completely deflating a core mechanic of the game is not a good solution to that problem.

It literally is a stupid argument to say, "I don't like how a max level mech has an advantage over a brand new level zero mech." If you want a solution to that, address the correct mechanics involved :)

In relation to the class specific specs, it's not that an A class wouldn't have access to additional health, but that speed would be emphasized more than health would. You've got a fair point about the body parts and the in-game mode though. It could be interesting in Siege if depositing energy allowed you to purchase items or something of the like though :)

#6 h0B0

h0B0

    Non Sequitur Leprechaun

  • Full Members
  • PipPipPip
  • 4,143 posts
  • Location[delete for trolling] --defter

Posted November 23 2012 - 09:29 AM

While i do understand the logic in thinking the a-class should favor mobility.
I do not see the practicality of it.

Merriam webster: Definition of OPTIMIZATION: an act, process, or methodology of making something (as a design, system, or decision) as fully perfect, functional, or effective as possible...

Forcing a class to invest optimization points in a characteristic it already excels at doesn't appeal to me and seems to be the opposite of the definition of optimization.

Edited by h0B0, November 23 2012 - 09:29 AM.

Click me! I dare you.

Posted Image

View Post[HWK]HUGHES, on March 15 2013 - 08:35 PM, said:

Oh don't always listen to h0B0. Lol.


#7 Psylo

Psylo

    Newbie

  • Members
  • Pip
  • 4 posts

Posted November 23 2012 - 10:09 AM

View Posth0B0, on November 23 2012 - 09:29 AM, said:

Forcing a class to invest optimization points in a characteristic it already excels at doesn't appeal to me and seems to be the opposite of the definition of optimization.

The game design point of having a class specific talent/specialization/optimization tree is that it creates variation where there would otherwise be none. If you take the current foundation, there are three A-class mechs. Without any kind of specialization system, that means there would only be six varieties of A-class mechs you could fight. The trees vary that.

The idea of creating individual class tree would allow for specific focus on a particular trait. Perhaps I should elaborate - I'm not saying the current trees are anywhere near fit for that particular solution; they would need to be vastly expanded. For an A-class, it would make sense to have the possible point of focus be a robust set of options in movement, fuel management, energy gathering, rader breadth, and dodge speed. There would be attack and defense buffs as well, but not nearly as many. Again, that's why I chose an A-class.

Now, if you like an optimized mech, then you would play a B-class. That way, by default, you would have a little bit more defense than an A-class and more mobility than an C-class. In your spec, you would have many options, the most diverse set, including almost equal parts offense, defense, and utility. The expanded tree available to the B-class would support that style of play.

If you had class specific optimization trees, they would HAVE to focus on something. Otherwise, they're not class specific then, are they? :P

Beside, I think the best option by far is a combination of some if not all of these ideas, plus something I'm sure someone will think of sooner or later.




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users