Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Retropunch]

An RTS game where all you need to do to be amazing is click quickly in obvious ways is not a good RTS game. In fact, it's not even really an RTS at all, since the "S" means strategy, and if you can win by clicking quickly with little advanced strategy, then it isn't one.

I've yet to find many that aren't! I mean of all the standard RTS games (supcom,SC1/2,C&C and so on) most devolve into just being able to build and deploy troops quickly upon multiple fronts - something the computer can do very quickly.

As Darkmere says, it's all about it being reasonable and the AI acting within believable parameters. As players can beat other players this must be possible, but as mentioned, it's much harder to do than just giving the AI loads of resources/cheating.

[nenjin]

I never thought I'd say it, but skills trees. It's not that I don't like skill trees. It's just a buzzword now though, and honestly, few of them really live up to the name. Not, like, Path of Exile's skill World Tree. So when I hear it in a pitch I just sigh and wonder how badly someone is going to not meet my expectations.

[GavJ]

insanely complex skill trees like Path of Exile's tend to be impossible to balance like small ones IMO, and yet decisiosn still carry almost as much weight, so you end up with local maximums in the state space, where people stumble upon a couple of possible trees that work really well, then those are just the best, because they were impossible to predict and balance fully. Then the whole tree goes to waste as most people just use the best trees for whatever playstyle and never bother delving into the rest of the possibilities.

I think my ideal (no examples in mind of this actually existing) is a tree just sufficiently complex enough not to be entirely grok-able to most players all at once (but not completely bonkers like Path of Exile), and then combine that with somewhat different enough procedurally generated situations that they demand different trees to excel at the different circumstances.

Then you have to actually use your noggin every playthrough to get the best tree for that playthrough, and you can't pull existing best solutions off the net. And due to moderate complexity, this is not trivial to do well, but also not so difficult as to essentially just be rolling dice.

[Trapezohedron]

1. Procedurally-generated Worlds

A world five times as large as the earth filled with random repetitive crap is no good world at all. I'm looking at you, Rogue Legacy.

But what about DF? It has procedurally generated Worlds too.

I am willing to give DF an exception, because for me, it is Procedural-generation done right. Discounting the fact that it has proper ore, stones and foliage placement due to raw settings and thus not part of RNG, I don't know of another game that generates the landmass according to logical landmass placement, generates rivers, causes those rivers to erode the pathways they cross as time passes, generate proper rainfall shadowing, create a barebones history that also affects the world ingame instead of just being decoration, and other stuff that I can't remember at the moment.

[Sonlirain]

Random generation usually is not random at all. Its generally just several premade peaces with a rearanged order and the "technology" is as old as the first rougelikes and went more or less unchanged though the 90's to this day.

In theory it increases the games longevity because the world is different every time and offers a different challenge.
But honestly? Once you see all the building blocks you pretty much saw it all.

[GavJ]

DF could use much better procedural generation, I think. It doesn't have enough variety to actually feel different enough game to game. Changes that would fix this (while also explaining what I mean more):

1) Caverns just KILL diversity. Every. single. map. I am guaranteed huge forests and almost always water and a full set of reliable crops and stones and huntable creatures. No matter where I begin. Which is just very counter productive as it undermines almost the entire map gen process. Caverns are fun and all, but should be randomly scattered around only. Maybe something like 20% of 4x4 embarks have any at all, in total, and 60% of 4x4 sections of mountain connect to caverns.

2) The high redundancy of crops makes it uninteresting which crops you get. Every biome that grows anything grows fiber, booze, and food crops in triplicate or whatever. And due to #1, you also have all types again subterranean. I admit food and booze must go hand in hand (you can ferment almost anything), but fiber and many other products in the future (like alchemy stuff or medicines) could be very biome-dependent, leading to different stresses in different games.

3) Not sure what to do about it, but: almost every embark has a an awfully large variety of metals and magma safe stones, to the point where worrying much about these is fairly optional.

4) The temperature underground is always 47 degrees fahrenheit everywhere in the world all the time ever. Which minimizes the amount I care about diverse surface temperatures (which themselves already aren't actually that diverse because there is no increased dehydration, hardly ever any frostbite in any reasonable timeframe). It would add realism and more importantly gameplay diversity, if temperature was more acutely felt requiring crop differences and clothing differences, and if the underground temperature correctly mirrored the climate (20-30 feet below, it is a constant temperature seasonally, BUT this constant temp will be much much lower near the arctic than near the equator). Also, magma tubes only slightly warming rock 3 feet away is more than a little silly and contributes to maps all feeling uniform, with a volcano for instance having shockingly little influence on gameplay other than the convenience of not building a stack.

5) Other aspects of map diversity similarly don't matter much, but I'm rambling on too long already. Elevation affects basically nothing other than making it annoying to follow creatures around. Rainfall has no effect on crops or anything, thus is irrelevant outside of evil biomes. Rivers and oceans are bizarrely usually LESS useful than ubiquitous aquifers. Etc.

It's certainly not BAD. But there's a ton  of room for improvement. And this is one of the better games for procedural gen.

[Spacefaye]

Games in general make me groan.

[Neonivek]

Games in general make me groan.

Heh!

"Hmmm game features... Action packed, 100 weapons, 200 monsters, ohh and online multiplayer... so far so good. Wait... 'it is a game' UGH!"

[penguinofhonor]

This might fit in the general procedural generation grumbling, but randomly generated enemies.

It usually amounts to "Hey, you know how much players love lazy recolor enemies? We made a system that makes lazy recolor enemies for us!" When your generator creates the equivalent of a game feature that players usually don't care for, that is a bad sign.

The worst is when they just have random ~elements~ on enemies. Oh, this is the same except it could take 25% extra damage if I have the right weapon? And it could do 25% more damage if I have the wrong armor? So exciting! The damage numbers that pop up might be 25% higher! That is a quality game mechanic.

Slightly less lazy is randomized attacks or slightly changed mechanics. You have a normal frog enemy that jumps around, a red frog enemy that also shoots a fireball every once in a while, a green frog enemy that also poisons you if you touch it, and a blue frog enemy that moves faster. You also have a bat enemy that flies at you, a red bat enemy that also shoots a fireball every once in a while, a green bat enemy that also poisons you if you touch it, and a blue bat enemy that moves faster. Hint: "We have 75 different enemies!" and "We have 15 enemies that can have one of four extra minor attacks!" are not the same thing.

I feel like I've seen this done right, but I can't remember when.

[Neonivek]

Just randomized in general. I think we highlighted it well.

Terraria is what ultimately killed "randomized worlds" for me. While a lot of people look at Terraria and go "Wow, so different" I just look at it and find that every word is essentially the same. There are no interesting features being generated and everything ends up looking bland and uninteresting. All the caves look alike, all Blue Mushroom forests look the same... is just rearranges it. Everyone saw completely different worlds and I immediately saw the same world.

Don't get me wrong, some games do that well... IVAN just took rooms and rearranged them, but that worked to its advantage because it allowed you to have cursory knowledge of the floor while not knowing exactly how things were laid about.

Spelunky also worked rather well because the randomization engine worked together with the gameplay. What it built was purposely meant to combine with the gameplay. If you had to jump down a pit, it was always barely deep enough to safely fall.

I guess it depends in the gameplay and random generation benefit the gameplay... or if it is completely separate.

Minecraft it just doesn't matter.

I feel like I've seen this done right, but I can't remember when.

Note: The quote is about randomized enemies.

I liked how Diablo 2 did it, and heck how Diablo 3 did it (Pre-hell difficulty). Yet it was just a minor feature for the game, they didn't outright advertise it. Enemies would just randomly have a few extra qualities that could inform you how to deal with them, and some of these really changed things up. What made it work is as I said, the game isn't about it so it doesn't need to focus on it.

[Darkmere]

Speaking of random enemies, what about ToME's wandering elites? They pick two player classes and get a smattering of abilities from each. It can be the worst, most disjointed pushover mess... or completely end you in a couple turns if you don't take a look at it beforehand. Unfortunately, far more of them were the former, so you were never prepared for the latter.

[Neonivek]

Speaking of random enemies, what about ToME's wandering elites? They pick two player classes and get a smattering of abilities from each. It can be the worst, most disjointed pushover mess... or completely end you in a couple turns if you don't take a look at it beforehand. Unfortunately, far more of them were the former, so you were never prepared for the latter.

Its a roguelike. It being completely unfair and killing you because you weren't prepared 2 years in advance is fairly typical.

Then again this is ToME which as a whole has a huge feeling that it basically tries to avoid the typical ToME terribleness so anything that is fairly typically roguelike is felt even more.

Yes the Wandering Elites are a rather lazy feature. It easily could have been rectified had they put any real time and effort towards it... but I have a feeling their view of it is "Hey its a roguelike".

Hey did I ever tell you that I think Roguelikes are fun but the constant apologetic nature of the rogue-like community and developers holds back the entire genre and as such many roguelikes use the typical RL features like a crutch in place of genuinely fun and engaging gameplay or don't actually support those aspects well? I say this because my disdain is rather obvious.

Actually add that to my list

-Roguelike: Used to be a feature that I loved, honestly the first time I heard of a game where if you don't wear protection when picking up a poisonous orb, that you would get poisoned. I jumped on it. Heck the dying a lot didn't but me at first. Then however I started learning what you had to do to win those games and found out that for the most part Roguelikes is pretty much a garbage genre as a whole with a few notable exceptions. Even the exemplars like Nethack are just dreadful in how they expect you to play the game.

[WillowLuman]

IIRC, in advanced gen setting you can make caverns more spaced apart.

[GavJ]

However, I imagine DF-wise that eventually, all (or at least almost all, since I'm not sure what to say about the magma-safe stones and metals thing) of these points will be addressed.

btw, just went ahead and made a thread here: http://www.bay12forums.com/smf/index.php?topic=139640.0
If there's any interest I may work on coding much of it myself. Maybe not *shrug*

IIRC, in advanced gen setting you can make caverns more spaced apart.

They're still either guaranteed or not, and I know which ahead of time because I'm the one setting the parameters =/ And when they are, they always have trees and are always the same shape, and are at a depth also known by me even if I change it, etc. etc.

Anywho, continue with your regularly scheduled, less derailed programming.

[Sappho]

I've got another one: permadeath. I used to love the idea of every life actually mattering. In online games, it feels like it should force people to be nicer to each other because otherwise they could die and have to start over. In reality, it's just an annoyance.

In multiplayer, it obviously just allows griefers to ruin other people's games. They don't care if their character dies; they just make a new one and carry on.

In single player, it goes hand in hand with the "hundreds of hours of gameplay" thing. I don't have a lot of time to keep replaying the same things. If I invest my time in a game, I want to feel like I've accomplished something, that I can point to it and say "there, that's what this time was for." If I can just die and lose everything, I just don't want to play anymore. I've started save-scumming on every permadeath game I play. Really, permadeath in single player should always be an optional feature, yet it rarely is.

This is worst in non-randomized games. At least in randomly-generated ones, you're starting something new. But there are some mostly non-randomized games (Legerdemain comes to mind) with permadeath, and that just drives me crazy. It's not "you died! start a new adventure!" It's "you died! Now you have to do EVERYTHING again! Isn't this fun?"

No.