Level Design and Keen

Discussion and analysis of graphics, story, levels, and so on.
User avatar
XkyRauh
Posts: 1113
Joined: Sun Aug 31, 2003 9:14 pm
Location: San Diego, California

Level Design and Keen

Post by XkyRauh » Sat Jan 10, 2004 8:41 pm

The average platforming game has many instances of single-brick platform jumping. That is to say, over any given cliff, there are likely to be a multitude of leaps to be made on platforms that are barely as wide as the character you're playing as.

More often, these platforms are 2+ bricks wide, but the screen is scrolling at a set rate, or there's an enemy placed in such a way as to make the jump difficult. In other cases, there's a treasure item placed in such a way that the easiest and fastest route from platform to platform doesn't offer much reward.

Keen is a different game from these. It's very difficult to make a level difficult by Keen standards without making it too long, or too reliant on a single obstacle.

Looking at Keen1's first level, there aren't many places to actually DIE. There are 4 different places where clam-plants can munch you (6 tile's-worth) and 1 pit of green spikes (2 tile's-worth)... 8 death tiles for an entire level that takes about one and a half full minutes to complete with every item grabbed.

Looking at Keen1's last level, there's a similar lack of apparent danger. There are several spikes that are nearly impossible to reach, two guard robots, a half-dozen fire tiles, and one nearly invincible Vorticon. The only real challenging part of this level is if you attempt to get EVERYTHING--and even then it's only difficult for the FIRST drop into the garg-pits.

My question then, is this: Where is the line between easy and hard? Where is the line between lengthy and too long? And what are your guys' opinions on ammo placement, treasure pacing, and enemy density?

This is similar to the poll that was written long ago asking about Big Room, Maze, etc... except that now I am specifically questioning each aspect of it. And of course, any miscellaneous comments that you feel are fitting go here as well. :-)

--Xky

KeenRush
Patch Maker
Posts: 1988
Joined: Sun Aug 31, 2003 2:52 pm
Location: Sand Yego
Contact:

Post by KeenRush » Sat Jan 10, 2004 9:19 pm

Hmmm.. Those levels are so easy.. Wow, haven't thought that before!
Where is the line between easy and hard?
Simple! First answer to question where it the line between good and bad! :)

Well, here are my answers to those questions:
Easy: You can play that almost your eyes closed. Not many enemies, dangers or anything. Just nice jumping.
Hard: Many enemies and dangers. Think twice before doing anything. You have to try it some times.
Good example of what is easy and hard - and everything between you can find from that level pack Test your Keen skillz!
Length: And too long level is full of stuff that shouldn't be used much - everywhere too long hallways to walk. It starts to get annoying if you need to walk long way.
Ammo: It depends a lot of your style of mod. If there's lots of enemies and it's meant to be classic Keen styled (with enough ammo) - use ammo. But I like levels with none ammo, and have used that myself..
Goodies/treasures: I love collecting those - don't be afraid to use them! :) Secret areas and places that are hard to go are good. *drool*
Enemies: Avoid using too much of them. Try to look at the enemy use of Keen 4 for example, and use in easy levels less enemies and in hard levels more of them. And it depends on enemy too. Group of four gargs is quite easy - but when fife vortininjas attack you..
Y-I-K-E-S-!-!

User avatar
XkyRauh
Posts: 1113
Joined: Sun Aug 31, 2003 9:14 pm
Location: San Diego, California

re: Good and Bad

Post by XkyRauh » Sat Jan 10, 2004 10:17 pm

*Laughs* KeenRush, your Good-Bad answer made me laugh--thanks. :-)

As far as enemy use goes--would you rather spend an entire level jumping over static, stationary killboxes (like the spinners in Keen5) or an entire level dealing with, say, Vorticon Grunts in Keen3? They've both got their advantages and disadvantages. With the killboxes you can flat out say "This area is impossible to get through" and leave it at that. With the enemies, you can have some luck-based elements, or even strategy (lure lure lure).

You say that long hallways are a disaster--what, then, constitutes a long hallway? Keen1-lv4: there are some long hallways left-to-right on your way to the Yellow Keycard. Keen1-lv16: there are some long 'hallways' on your way up to the "ID" made of goodies--all those ice platforms. Keen1-lv13: arguably, the entire thing is a long hallway... hehe

Now, the first example, Keen1-lv4, is more tolerable because there are points to get--you have to jump to reach the points, and so there is still something to DO in the hallway. You can choose to not jump at all and rush straight for the card, or you can slow down a bit to snag some points (not much though--they're all balloons!) The second one, however, Keen1-lv16, is much less interesting because you're not only inactive, but you don't even have to touch the controls! And all of the wait is worth barely 5000 points--the same amount you get from dropping down one little garg-pit far below! :-)

Is there some sort of system you guys can think of for balancing risk-reward? For example... a series of two-brick platforms in the air (5) leads to 1000 points (Keen1-lv1). Or an extended risky section leads to a full-on extra life (Keen1-lv13, lower left corner). Are there any rules you follow when designing your levels? Do you set down and say "As a standard, all levels need a possible total of 18,000 points, but only 8,000 of them should be attained in a straight-line minimalist path to the exit." or something similar?

Keen only needs 20k for an extra life, and this can get heavily abused. Check out Keen2-lv11, everyone's favorite FUCL level: After a series of impossible-to-die-on platforms, you head left to face 2 Vorticons. They're easy enough to take out, and you're given over 6000 points and 5 shots for doing so. Depending on the route you take down the left side of the level, you could snag over 2000 points and an additional 5 shots, taking out 4 Vorticons along the way. Then, a simple pogo trick using the Scrub nets you--get this--44,000 points. Two extra lives. AND another 5 shots. Fall back down to the Scrub and head right, you encounter the only real challenge of the level: the unpredictable Vorticon Elite. If you manage to get past him, another 10,000 points are yours. This puts the total gain for the level at over 60,000 points and 15 shots (you'd have to shoot 11 times to kill everything, which is still a net gain of 4).

Even if you died somewhere along the way, you've got a 2-life loop that earns you ammo. One could theoretically head straight for the 44k-and-5-ammo thing and die there repeatedly, boosting their score and their munitions. Is this kind of design encouraged or discouraged by you all? :-)

--Xky

KeenRush
Patch Maker
Posts: 1988
Joined: Sun Aug 31, 2003 2:52 pm
Location: Sand Yego
Contact:

Post by KeenRush » Sat Jan 10, 2004 10:50 pm

Ooh. Too long post! :o

1. Both - moving and not moving enemies.
2. I meant with the hallways something like for example in some badly made user levels where you have to walk first 90 tiles to left, then 90 right, 90 left, 90 right.. :) Or too long mazes, walk 30 tiles - nothing, walk them back, walk 70 tiles - ah, keycard - walk 90 back, climb 55 up..
3. Yeah..
4. None use I.
5. Remember - Keen doesn't always need 20k to get extra Keen, the value he needs is going higher and higher until you can't get even one life no matter how much you try.. :)
6. Not encouraged..

Joseph

keen level design

Post by Joseph » Sun Jan 11, 2004 2:19 am

5. Remember - Keen doesn't always need 20k to get extra Keen, the value he needs is going higher and higher until you can't get even one life no matter how much you try.. :)
Not for Keen1-3 it's always just 20k.

I'd keep in mind for this thread guys that each of the Keen games is different and none of them are perfect.

Guest

Mega Wacky.

Post by Guest » Sun Jan 11, 2004 7:24 am

Quite a tough question, but an important one nonetheless.

Commander Keen: Monky Business will likely be notorious for out-of-balance levels. There are enemies, obstacles and points everywhere. I try not to leave corridors empty, and I ensure that the player needs to collect as many lives as possible, as there are a lot of pitfalls and stumbling blocks that shall lead to some levels being repeated many times before they're finally completed. There are plenty of lives available in the bonus (gateway) levels, but you need to do some work to actually get there.

One thing my mod is most in need of is backdrops. Rather than relying on points and enemies to fill in the vacant spaces, I can then use scenery to balance out the emptiness. But then, the nature of the world in which Keen finds himself is one where Essence is abundant everywhere--I reduced the scores for the first four items so I can put more Essence in, but now it looks like I've overdone it... so I put more enemies in to balance it all out... it's a vicious cycle. When it comes to the full release, I'll put more thought into an actual system behind my level design.

---

I'd compare good levels with well drawn-out movies. Some movies just work well, and everything seems to be there for a reason. Others, while just as enjoyable, may be over-reliant on action sequences, sex scenes, long dialogues or comedy relief, just as levels can be over-reliant on points, enemies, perfect-timing sequences and long hallways.

The best way to do this, I guess, is to start with the larger picture in mind. Your technique, Xky, of setting yourself some basic rules as to how you go about specific aspects of level design, is great for this, if the rules are set organically enough (ie. there's a reason behind them that fits in with the storyline and the purpose of the level in the game world).

That hallway you used as an example in Background Query blew my mind. The hallway has a purpose, the challenge is simple, the reward is ample, and the subtle changes from the first shot make the whole thing look impressive.

And Grelphy's mod looks like it's making good use of empty space, with its strategically placed backdrop elements and point placement.

But take a look at most of the CK:MB screens, and you'll find that my levels are mostly an endless action sequence. There's stuff happening everywhere, and the player has a lot of choices as to which approach is taken to beating each part of the levels. I intend to continue with the open-ended style, but I need to cut back on active level elements and give my levels some character rather than loading them up with filler.

---
Even if you died somewhere along the way, you've got a 2-life loop that earns you ammo. One could theoretically head straight for the 44k-and-5-ammo thing and die there repeatedly, boosting their score and their munitions. Is this kind of design encouraged or discouraged by you all? :-)
I have been working to avoid this kind of situation as much as possible, but my random approach to level design has probably left a lot of loopholes available to be exploited. IMHO, this kind of loop defeats the entire purpose of the high scores list, as it's simply a matter of how much patience/spare time a player has as to how high a score can be achieved.
Is there some sort of system you guys can think of for balancing risk-reward? For example... a series of two-brick platforms in the air (5) leads to 1000 points (Keen1-lv1). Or an extended risky section leads to a full-on extra life (Keen1-lv13, lower left corner). Are there any rules you follow when designing your levels? Do you set down and say "As a standard, all levels need a possible total of 18,000 points, but only 8,000 of them should be attained in a straight-line minimalist path to the exit." or something similar?
I've been considering the importance of structured systems in projects and general activities lately, so this point gives me a lot to think about. The main flaws in my level design are an imbalance of interactive elements and a lack of basic direction as to the layout of the level itself, so everything seems very random and unpredictable.

Here's a system I've just decided on, which may add to the believability and general coherence of any mod:
Oo Level map designed from ground up (ie. purpose of each area defined; then the level design elements will attest to this purpose).
Oo Basic structure of each level drawn up, the purpose of each section defined, and again the more specific level design elements that will attest to this purpose, including basic outline of challenges/rewards.
Oo Appearance of each section outlined, taking into account the theme of the level, so that specific tiles can be decided on when the time comes rather than simply searching through the list randomly looking for something that might fit.
Oo Individual sections integrated (ie. passageways, platforms, doors/shards, teleporters if relevant).
Oo Level constructed directly from outline, then further details added in to improve appearance.

Well, that's the basics of it. I'll probably create a list of challenge combinations/level concepts/level section ideas, and so forth, with which to work from when using the above technique.

This is what I've needed to read. My random approach to level design has really hampered my progress thus far--I'm constantly reworking challenges to make them easier/harder/more balanced, so I end up with levels that are an incredible mess.

I'll have to make use of the 18,000/8,000 idea, or at least a variation of it. And secret/difficult-to-access areas shall abound (my levels, despite their open-endedness with respect to the order in which each challenge is approached, are still markedly linear).
As far as enemy use goes--would you rather spend an entire level jumping over static, stationary killboxes (like the spinners in Keen5) or an entire level dealing with, say, Vorticon Grunts in Keen3?
I guess, if you've got a specific purpose for your level, you can get away with anything. I may have a library in which I can place a heap of monks in one room (the ones replacing the butler robot) with perhaps a gem or something over to one side... or a storage room full of Essence jars defended by higher order monks that fire... and elsewhere I'll have processing plants in which can be found pits full of Essence along with extractor cannons. If you've got a purpose for the level, then it's all about context.

Well, that's what I figure.
Keen1-lv16: there are some long 'hallways' on your way up to the "ID" made of goodies
I love that "ID" made of goodies... it's worth more than the points just to go through the effort of getting there. That could be an interesting strategy--creating bonus locations that have a higher value than their point total due to the way they're shaped, or because of a particular background component, etc.

---

Woog. Now I have some new ideas.

>Commander Spleen

KeenRush
Patch Maker
Posts: 1988
Joined: Sun Aug 31, 2003 2:52 pm
Location: Sand Yego
Contact:

Post by KeenRush » Sun Jan 11, 2004 8:24 am

Great post Spleen! :) By the way, why not logged in?

Anyways, I like system; "Here's a system I've just decided on, which may add to the believability and general coherence of any mod"
I wish I had done something like that..

Your enemy-use answer is very good. I'll try to design enemies next time that way, that I can use eniemies depending on level-style.

Ilsoap
Posts: 197
Joined: Tue Sep 02, 2003 4:04 am
Location: Canada
Contact:

Post by Ilsoap » Sun Jan 11, 2004 8:57 am

The more unknowns a level has, the more difficult/annoying it is.

Take a room that's one long hallway (say, 50 tiles) with a door at the end. That's one unknown; "How long is the level? Is there something hiding further down?" Now, put a garg at about tile 15. \

Now, there's an extra unknown, because after you shoot the garg, you now have more of a reason to believe that there's another enemy waiting.

Now, instead of a straight line room, make it go right-down-right. So at about tile 25, the room stops, and you have to fall down a pit to continue. If the pit is 20 tiles long, you have a third unknown. How far does this pit go? Will I land on something?

Simplistic examples, I know. But based on this, you can determine how easy/hard a level is. The first level of Keen 1 is easy, because even though there are things that can kill you, they're all visible. You see them coming, and as long as you don't stray from the easiest route, you'll get to the exit without incident. The first level of Keen 6, however, has multiple dropoffs into the water, bloogs, spikes, a babobba, and blooglets, who, although they can't kill, will push you around, which takes away your feeling of control over the situation (unless you stun him).

Enemies that can't be killed are often considered the most annoying, because their kind of unknown is "Since I can't kill it, how am I supposed to get out of the way?". Giks (the yellow guys from Keen 6), jacks (from Keen 3), and dopefish are among the most annoying enemies for that reason. So, if you're doing a Keen 1 mod, you'll want to use the tankbots sparingly.

That's my theory.

User avatar
XkyRauh
Posts: 1113
Joined: Sun Aug 31, 2003 9:14 pm
Location: San Diego, California

re: Ilsoap, Spleen, and KeenRush

Post by XkyRauh » Sun Jan 11, 2004 10:02 am

When I asked which you preferred, the stationary versus active--that was not intended to be answered "both" :-) ...I was asking if it had to be one or the other, which would you prefer?

Ilsoap, your theory of unknowns is fantastic--and it speaks of your mod very well--many times in your mod I found myself asking "how big IS this level?!" ;-) which was both good and bad. Good because it kept me guessing and my memory active (okay, green keycard is lower left...) and bad because I had a lot to look forward to on that return trip.

Writing up outlines for each level and putting point guidelines are things I've used since my Keen1 mod--I just set the points standard far too low for it. Originally I had about 25000 points in each level, total, and I'd thought that was too much--I was wrong. I should have kept all those points I erased--by the time my Keen1 mod was released, the average poitns per level was around 12000... which lead to quick game-overs and complaining. Better to have an overabundance of points than difficulty. Unless you're gonna release "easy" and "hard" versions of the game.

...

But once again, all of this discussion has brougth me back to my first question (the one KeenRush refused to answer!) How big is too big? How hard is too hard? And in Ilsoaps case, how much Unknown is overwhelming?

Often times I think it is impossible for a level designer to gauge such things. Having played Keen for so many years, and become familiar with the game engine, we all play Keen on a far different level than a casual gamer.

...

I asked my little brother to playtest some of these levels (he's 18) and I was shocked to find him missing jumps, unable to control the pogo stick, and firing his gun accidentally on a consistent basis. The jumps that I had designed for the "Easy" level completely stopped him!

Later, I had CheesyDave playtest the same levels and breathed a sigh of relief as he passed about 70% of the level as I imagined--but he also found a few errors I had passed over--shortcuts, abuses, and loopholes that I'd not thought of because I had the level's path in my mind already.

...

Probably the most beneficial thing I can do as a level designer is ask people to playtest frequently, and tweak often. Starting with a formula or design document help a bunch, though.

Without wanting to rant too much more, lemme just say that no game is ever perfect. There will never be a "perfect" level... but there are certain aspects of any given level that are better than others--and those strong points are what I'm looking to extract. All of your insight has been invaluable... I feel better and better about my mod! :-D

--Xky

User avatar
CommanderSpleen
Posts: 1014
Joined: Sun Aug 31, 2003 12:11 pm
Location: The Land of Sparkly Things
Contact:

Large Camels!

Post by CommanderSpleen » Sun Jan 11, 2004 12:03 pm



KeenRush
Patch Maker
Posts: 1988
Joined: Sun Aug 31, 2003 2:52 pm
Location: Sand Yego
Contact:

Post by KeenRush » Sun Jan 11, 2004 12:52 pm

Mmmmh.. This thread is giving me some mortimerish ideas I'm going to use later! :)

And ok, I prefer more (if I need to select between enemies and dangerous tiles) moving and intelligent(?) enemies. It gets a bit boring if you always know where your enemy is (in case using only tiles). Though, sometimes it's annoying if you fall down and land on a garg - which happens me every time I play your mod's last level Xky! GAAARG! :)

Hmmm. Too big is something that gets boring, and you feel that you can not beat it ever.

And you know your level is too hard if you can easily pass Keen 6 on hard without dying once nor saving once, and you'll keep dying in your own level time after time. :P

Remember, these are very hard because everyone has own opinions about these things.. ;)

User avatar
grelphy
Posts: 219
Joined: Fri Sep 05, 2003 12:41 am
Location: the Armageddon Machine

Post by grelphy » Sun Jan 11, 2004 2:10 pm

Oof... this is quite impressive. Who would have thought Keen fans could be so philosophical about level design?

Okay, a couple of things. Moving enemies are always better than stationary tiles. Always. Even Vorticon kids and shocksunds. Well, maybe not them, but...

I tend towards lots of invincible or "tougher" enemies... Gargs, Grunts and Meeps never seem tough enough. My favorite sprite: the guard bot. Muhahaha!

Actually, meeps coming toward you can be insanely difficult, since their "singing" obliterates your shots, and then keeps on coming. Yeek.

On the subject of points/ammo: There is no such thing as too many point tiles. There is such a thing as too many points, but if you restrict yourself primarily to low-scoring items, you shouldn't have too much of a problem. (After all, how many levels have 200 lollipops in them?)

Ammo, on the other hand, should be carefully regulated. Players (or "victims," as I prefer to call them) have a tendency to misuse those extra rayguns... so keep it down to a surplus of just one or two shots per level. Keep in mind all those baddys that can be dodged: gargs in pits, for example, needn't be counted, nor should baddys guarding point pits. Just remember, each level should be possible starting fresh, with no shots or, in the case of Keen 1, no pogo. Evil is fine, but impossible isn't. Also, you are well advised to keep an eye on your level borders; my standard strategy for beating K1-level 4 (joystick) is to let the vorticon fall off the edge of the screen, thus conserving 4 shots.

Too big or too small are really abstract concepts. I would say that if a level takes more than maybe ten minutes to complete, its too big; less than one minute, and it's too small. Difficulty is equally abstract: levels near the beginning should be generally easier than levels at the end, but it's seldom a nice smooth gradient. Difficulty is in the eye of the beholder. Beginners (namely: my dad) often have trouble with the first level (!), while experts such as us often glide through even the toughest of levels. When you can shoot ten gargs coming down an incline with ten shots, you know you're starting to get too good.

Okay, that about wraps it up, I hope... can't think of anything else. I'm sorry I missed this for so long, but...

xtraverse
Patch Crafter
Posts: 290
Joined: Tue Sep 02, 2003 6:42 pm
Location: Easter Island
Contact:

...

Post by xtraverse » Sun Jan 11, 2004 7:25 pm

Arg, you guys are so lucky to have friends that would actually have an interest in playing and testing your creations, I don't know a single other person that would care a wink about ck-related stuff.

Anyways, too big and too small in my mind more refer to how the level feels. If you have a very small level that's empty and barren, I'd consider it too big, but maybe I'm just weird. :P

Anyways, you guys seem to be pointing out every flaw of my first mod. I made massive killing rooms of Vorticons which would come out of no where. You couldn't see them, and suddenly they'd appear and kill you. I also made annoying levels where you had to repeat the same stretch of difficulty over and over again. If anyone wants a guide on how to make bad levels, play Norp the Yorp 1 :D

A level style I really like that unfortunately I only used once is sort of a level that builds up and gets really exciting. In the second to last level of Norp the Yorp 3 (the one right before Princess Lindsay), You start out in a sort of open area with points and ammo scattered around, preparing you for what's to come. There is a few keydoors guarding something, and a hole where you can go underground. When you get in the hole, the only thing you can do is explore a large but interesting maze for keys. I never tired of exploring, because I thought it was big enough but not too big.

Things get more tense as the hardness of enemies builds up. When you finally get all the keys, you open the doors up top for a final key. You must return to the bottom for a grand finale. There are two doors that lead to another fall. Opening one will give you to Ankhs, the other, none. When you drop in with 20 seconds of immunity, you fall into a pit of Vortininjas--things get really exciting as you need to decide whether to go right or left--the wrong way could result in your death. Going right will get you a necessary key, left will get you to the keydoor before the exit. You really need to play it to get the feel, but in my opinion this is the best level I've created.

So I'd say levels that build up tension are great, and once in a while a grand finale is effective.

KeenRush
Patch Maker
Posts: 1988
Joined: Sun Aug 31, 2003 2:52 pm
Location: Sand Yego
Contact:

Post by KeenRush » Mon Jan 12, 2004 4:24 am

you guys are so lucky to have friends that would actually have an interest in playing and testing your creations
Well, you have here in the forums..

Well.. That's exactly what good level is about, xtra! :) Too bad DosBox is so slow on my computer, I would love to play that level in "real-Keen-time". I'll try to create that kind of level(s) in the future.

And Norp 1 wasn't that bad, I liked it, thought some places were hard and I didn't bother to get every point because it would have taken too much time. :)

User avatar
XkyRauh
Posts: 1113
Joined: Sun Aug 31, 2003 9:14 pm
Location: San Diego, California

re: Playtesting

Post by XkyRauh » Mon Jan 12, 2004 5:22 am

KeenRush said:
xtraverse said:

Arg, you guys are so lucky to have friends that would actually have an interest in playing and testing yoru creations, I don't know a single other person that would care a wink about ck-related stuff.
Well, you have here in the forums ..
Totally! :-) Playtesting Lego Keen for Ilsoap was a blast, and of course we missed a few things here and there, but it was not only a chance to give feedback on likes/dislikes regarding level design, but it was a unique opportunity to help refine someone else's ideas. Personally, I feel that Lego Keen is one of the most polished mods the Keen community has produced--the Previews screen and Ending alone are fantastic! And it was all helped along by a little playtesting from us.

Don't be afraid to post looking for playtesters--we're here, and we're eager! Ilsoap kept a few of his secrets to himself when he asked us to betatest--he left out a few key levels, and cleaned out parts of his patch file, so that we couldn't get nosey or obnoxious. You could do that, or just let us have your raw work-in-progress... Half the time I ask CheesyDave to playtest my levels, I've got a half-dozen graphical errors just laying around because I haven't bothered to fix them yet. :-)

That being said, if you're giving a product to somebody to playtest, don't just say "Well, how was it?" Give them specifics or probing questions--Did they find any single aspect of the level annoying or enjoyable? Were there too few or too many points? Was the enemy placement and pacing okay? Did you ever feel lost? Etc, etc, etc. As a beta tester, there are so many potential things to comment on--from the shallow "I don't like what color you made your platforms here" to the more productive "This Vorticon presents too large a risk--maybe if you spawned him further right."

Good luck, and happy level designing!

--Xky

Post Reply