DRL > Requests For Features

Balance suggestions

<< < (20/37) > >>

MaiZure:

--- Quote from: VANDAM on March 16, 2013, 06:10 ---And enemies with random health ofc.
former human health: 4d5 (min 5 max 20) and so on.

--- End quote ---

This I'm not completely sold on - the reason being that the randomness should have a meaningful impact the way players handle the game. Finding an interesting weapon or a mod blessing will change the way a player thinks about the run. However, I'm not so sure a player will likely change strategy for each former or lost soul that wanders in to view. Is there a reason that a health 18 or health 22 former would change player reaction? Do we even have a way for them to tell what kind of former it is?

The difference is subtle, but I think important in decide how to handle RNG's interference. Promoting micromanagement of each individual encounter beyond what we already have to do is probably a bad approach

VANDAM:
hmmm, I think you're right.
The difference in weapons will give an opportunity to change run, and also enemies will be more dangerous (they're using those weps).
I think that uniques should have the same properties. Do not know about exotics.
And there was a problem that bosses have the same amount of health, and the only more or less dangerous enemy is Angel of death.
Cybies are predictable and with new set they are useless, Lava Elemental is better but useless when you wear new set. Shambler and Mastermind are predictable too. The main idea is to overrun them and/or find a suitable wall.
They could be more dangerous if they will try to make predict shots (not to fire in you but in the near block) and if they will fire out of LoS (they hear your footsteps and they can miss). And I think that M Apo JC C should be rebalanced (health, damage).

MaiZure:
To qualify the point about different stat weapons - It maybe be better restrict enemies and their drops the base weapon and only the randomly generated weapons on the ground could be of the randomized variety

Why would this be better?
To avoid the inevitable cycle of kill the former, check the drop, kill the former, check the drop. After 100 formers you will likely have the upper-tier shotgun/chaingun in EVERY run. Hence there would be no variety. Ultimately, we should try to avoid mechanisms that become a monotonous grind outside of the already obvious Ao100.

MaiZure:

--- Quote from: VANDAM on March 16, 2013, 09:50 ---Cybies are predictable and with new set they are useless, Lava Elemental is better but useless when you wear new set. Shambler and Mastermind are predictable too.

--- End quote ---

Sorry for the double post.

To the bosses being predictable, I would say that rather than have one AI rule the boss. Have each boss with a set of AIs that the game chooses from at the beginning of the game. Sometimes Cybie tries to hide and shoot, or charges right at you, etc. This way you can't know for sure what to expect every time.

This game, Cybie will charge but the Bruisers will camp their shelter area and wait for YOU to act.
The next game could be the opposite.

Evilpotatoe:
I didn't know extra credits. Very interesting, everyone should watch that !

I think I agree with everything he said, about perfect imbalance, the facts that games should be easy to learn (e.g. with tutorial(s)), that there's no excuse for a counter intuitive interface, or that irreductable complexity is *very* bad. (I didn't try Dwarf Fortress, and will probably never... for this obvious reason)
Though, in doomRL's balance, we're not in LoL or DotA : the RNG can't "counter pick" with lava filled levels when you go tactical boots, or increased number of revenants when you choose DodgeMaster.
That's why we should keep things as much balanced as we can (don't read equal), so that when you find some great items (especially early), it makes you think again about you build. Last time I found a super shotty at level 2, I wondered if I should go for Shottyhead instead of CatEye, but I didn't :
 -Shotguns are great, but unless you find a plasma one or a sniper mod, they sometimes lack power. Plasma rifles don't have this problem.
 -Shotgun shells may sometimes not drop in later levels. With a nuclear plasma rifle and 2 B-moddable BFGs, you solve all ammo problems.
 -Shottyhead doesn't seem worth picking to me. I think taking no mastery is, at least, as good as picking it. It blocks SoB, and shooting speed is only useful in the open... (read mortuary ?).

Of course, my point of view is probably not 100% accurate, but why would I have switched to a weaker build while some special levels ensure me I'll drop the best (plasma) weapons of the game ?
On the other hand, if I had had the option to switch to MAD (which I don't like, but would have made a good use of the weapon), I probably would have done it, just to take advantage of the occasion.

Same problem applies when you find a good gun, which is quite useless if you're not going for SoG... good rocket launchers, well, it's called missile launcher, and can only have some use in N!, to make the arena easier, since you'll get one at Lv7 anyway.
The result is quite obvious : you'll keep on your ammochain build in most cases, just because it's safer and more efficient than the other ones... :(


--- Quote ---There are just not enough variance in boots to make it interesting. It boils down to speed, dodge and lava immunity. Last one is more situational, and not available until much later anyway. And first two are both on one assembly. So, actual choice is just make tac boots before or after wk2
I think we need more boots variation ideas
--- End quote ---
Fully agree. Removing movement penality on all resistances boots (and improving gothic to something usable, like -20% EDIT : nvm, gothic boots have -15%, only armor has -70%) would probably be a first step. Does anyone think it would be too powerful ?


--- Quote ---Why do Tact Boots or any assembly have to have static stats?
--- End quote ---
Because non-assembled modded weapons have static stats. If your assembled hyperblaster can be worse than your AT2 plasma rifle, you'll stop assembling it (or only try one when you have mods to dump).
As for weapon's damage dices randomization, I agree that, "If done tastefully, it could work", but I'm sure it's a wrong way. Having to check every shotgun looking for a 9d3 would be inuintive and annoying, but wouldn't bring much to gameplay. In other words, it would increase game's complexity, but not depth.
In fact... we already have exotics for that purpose.
The problem is that some are guaranteed (like missile launcher or good plasma weapons), while the other aren't (e.g. good guns/shotguns, or minigun), which makes some strategies reliable... while the other have random efficiency.
What makes gun masteries (namely, MGK, MCE and MSh) efficient is that SoG is OP, and thoses masteries are good. Same for MAD, which lets you obliterate any enemy with a simple shotgun, but a Shottyhead or Fireangel build, for example, will be too much impacted by available weapons (not that it's a bad build, but it's not reliable. It doesn't ensure the freewin that other masteries, like MAc would)


If we want to add randomness in the game, I'd rather see enemy's behaviour reworked. Not only for bosses. I think having several AI for enemies (e.g. stand still / wander randomly / explore level / hunt player [in group]) would be nice : you could camp wandering monsters, but not the whole level... and should have to worry more about exploring formers or knight, since their aim would be to run through the level and grab as much stuff as they can before you do.
For bosses, it could be interesting too, but if some AI are more efficient thant the other, we'd end up with an ugly, randomly-refereed balance problem, again.
This can probably be made better, with bosses changing strategies during the fight, but is that enough ? e.g. Cybie could keeps hunting player until he comes close enough, then, choose, after every attack, between charge to melee, normal attacks, aimed attacks, radarshooting, or a few turns of strategical retreat. (melee or retreat/radarshooting would only trigger at close/long range, btw)
Anyway, while this kind of ideas is interesting, it would need a lot of testing. That's why balancing some game elements, which clearly need it (boots, other useless assemblies, F-mod...) looks like a more sensible approach than from scratch remakes, atm.

Navigation

[0] Message Index

[#] Next page

[*] Previous page