Level designing tips

[Shvegait]

I don't think I ever implied that I was, but it's cool to see that this forum exists and there is still such an interest in such an old game There was this girl on my bus in 2nd grade who drew them too, and we shared on the bus or something... that was a long time ago though, when I knew nothing... &#A0;

If anyone wants to try my first (LemEdit/CustLemm) level pack, it's in the level pack topic. Feedback would be very helpful, though I admit these levels are short and most of them are pretty easy to figure out. Then again, it's hard to tell if something is hard to figure out once you have it figured out, you know? Hard to implement is anoter story, though you shouldn't have much trouble with these compared to many others...

[guest]

I somehow seemed to have missed this thread until now.  I think all 4 ideas mentioned in the poll are great, and I'd actually say that you should apply as many as you can simultaneously.

For example, for my recent Cheapo level "This level requires no skill" (I promise to find a method at some point to distribute levels on the web), I started off with a specific base idea, motivated by breaking the level record in question.  I then simply created a minimalist style level that just does more or less what my base idea wants, and somewhat randomly assign statistics.  I then start playing around with the level many times and start adjust numbers.  For example, for this level since I can't use skills, I start adjusting how many lemmings come out.  And in essence what happened is, I ended up applying the "play around with it until you get an interesting/hard/fun solution", and that led to the present form of the level, where I've in effect added a bit to what would've been a simple base idea.

For the particular level in question I'm quite constrained, but in other cases, you can apply the other 2 tips.  For example, maybe the level you have right now is a little too sparse and you want to add stuff.  Why not throw in some stuff you've seen before, add a few personal touches, and voila, you've expanded your level, in effect by linking old stuff together.  Repeats are also interesting, if you're absolutely out of fresh ideas.  Or instead of doing a true repeat, you might choose to do a repeat with slight but critical modifications [think "Just a Minute" and "Just a Minte (Part 2)"].  Or heck, maybe you'll incorporate an entire old level into parts of a new level...Insane Steve's "Honey I Shrunk the Levels" is a brilliant example.

So I think all 4 tips are equally useful and should be applied in tandem.  Then the sky's the limit.

[namida]

I too drew many levels when I was younger, but I've lost them all. They were pretty pathetic anyway. Not a single one was hard.

[Proxima]

Heh, by the most bizarre coincidence just a couple of days ago I managed to find the old set of levels I drew -- must have been around eleven or twelve years ago -- and yes, like yours, they were pretty pathetic. I might re-make some of them in Cheapo when I have time just for the comedy value.

Some of mine are hard, but only because they involve lots of building through a huge level. Bizarrely, though, one I thing I quite often did was draw a pretty hard level and then negate the hardness by only requiring 1 to be saved.........

[chaos_defrost]

My notebooks contain a lot of poorly scaled levels. 8.5x11 inches is not the typical lemmings level terrain. All the straight lines weren't really straight. And some of the levels were probably impossible. A few of the levels were REALLY out of place. (The second to last level in my first two notebook attempts is currently remade in the FIRST SET of my Cheapo remakes.)

[Conway]

Sounds interesting, but also very complicated, both to make and to use. And it doesn't sound all too usefull, unless you're on holiday without a computer or a laptop, and you feel a desperate desire to design a level, which you still can't properly build or test until you get back.

  However, all this information would be useful for creating an exact replica of Lemmings but non-DOS based, like Cheapo, but even closer to the original design.

  Still, it sounds like an interesting project. I just don't think it would serve much of a purpose.

[Timballisto]

Actually, I have a plan.  I'm going to create a notebook.  I've started collecting statistics, such as lemming level playing area (1584x160) and the number of pixels a lemming will fall off of and die (65 not counting the pixel above(When I say this, imagine we all fall down.  If you ran up 65 pixels, the first pixel being directly above the lower platform, then you would land on the top pixel of the top platform.  I'm not sure where the Lemming actually starts falling).  I'm also going to try to dig out of Mike, the formulas for lemming release (release rate (lemmings per second) trapdoors (order of release from)), rate of lemming movement (falling,digging,anything basically), length of dying animations (if a lemming is a bomber, and he explodes during the animation, the effects still occur), pixel positioning of mouse (it is only so precise in where it can be), real time of bomb count down, max time allowed in a level (99?), slant and length of bridges, and several other things.  I will collect these things in a notebook.  then I will print all of the terrain pieces in black onto paper from each set.  These will follow the numbers.  Following will be multifold pieces of paper, that are basically blank paper lemmings levels-you will be able to tell the pixels because the pixels will alternate, light blue, white, light blue, white, etc.  and look like a checker board.  On these, you can trace terrains.  On the backs will be statistics, such as number of lemmings, lemmings saved, etc.  Using the numbers from section one, you can test your levels for possibility.

In short - the ultimate on-the-go level design book!

You could make levels-anywhere!

recap
sec.1-valuable statistics
        -physics statistics
        -physics rules
sec.2-terrain piece black prints
        -object black prints
sec.3-design sheets

Oh, and it'll all be in a binder.  This way, levels and terrain sheets can be removed and put back in.
---------------
On a side note, I've developed a refined process for making levels.

[Timballisto]

Well, actually you could "test" a level-you just have to use a couple of things from sec. 1 to find out where a lemming would be at a certain time (precise timing here based on math-no guesses), how fast they come out of the trapdoor(also math), how long a hole would take to dig(ditto), etc. &#A0;If you have all the right numbers, then you'll know where the lemming will be. &#A0;What helps is, if I'm not mistaken, the lemmings' rate of movement does not change over inclines.

The purpose was so you could design levels away from home. &#A0;Geez, I'd do it in school!

EDIT:First level design sheet has been created.

[guest]

I've also drawn over a hundred levels on paper. &#A0;Many of the Cheapo levels that I make use tricks and other ideas that I drew on paper. &#A0;Often I use my paper levels for ideas for new Cheapo levels.

I have just completed my first set of ten Cheapo levels, but I have not released them yet. &#A0;

It's been over half a year, will we ever see them released?

[guest]

Well, actually you could "test" a level-you just have to use a couple of things from sec. 1 to find out where a lemming would be at a certain time (precise timing here based on math-no guesses), how fast they come out of the trapdoor(also math), how long a hole would take to dig(ditto), etc. &#A0;If you have all the right numbers, then you'll know where the lemming will be.<snip>

The purpose was so you could design levels away from home.  Geez, I'd do it in school!

I think it'd be too tedious and error-prone to really try to hand-simulate the movements of individual Lemmings as you suggested above.  Besides, it's mostly unnecessary anyway.  When a human solves a level, you generally don't try to simulate what exactly happens to the lemmings.  Instead you just have a general idea of what to do, and then try it out to see if it works out.

Unless you're going for some very timing-sensitive moves, I think you can easily get away with designing the levels without trying to simulate the lemmings' movement at all.  You can deal with that later by playing on the computer; after all, you should playtest your level at least once before releasing it.

What helps is, if I'm not mistaken, the lemmings' rate of movement does not change over inclines.

You're mistaken, or at least only half-right.

If the step in front of the lemming is 0-2 pixels high, then yes.  But if it's 3-6 pixels high, what happens instead is that the lemming momentarily turns into a jumper (if you pause the game at the right moment and highlight the lemming, you'll find that it is listed as a "jumper" instead of a walker, unless it has been assigned a climber or floater).  The jumper moves up I believe at most 2 pixels per frame (less if the top of the step is reached), and the game already moves him up 2 pixels during the frame when he transitions from walker to jumper.

So for example, if the step in front of a Lemming is 6 pixels, then it takes 3 frames to get up there, instead of 1.

Of course, most inclines are not too steep, so you won't get steps of 3 pixels or higher.  But certain terrains, such as the staircase in the "columns" set, does have tall steps.

[Argai]

Most of the time, I'll just start building a level without having a specific idea. I think 9 out of 10 levels I am creating that way aren't good enough imo, but I have to play the levels a while before I know that

[Proxima]

Interesting that no-one yet has voted for "Finding ways to link several old tricks"; I very much enjoyed the recent collaborative level using this device.

[guest]

Well it's an old poll, so I'm not sure how accurately it still reflects current practices.  I'm sure there are more levels out there that combines tricks than what the poll would suggest.

Of course, one tricky aspect of combining tricks is that, you have to make sure you can't use the skills intended for one trick to get around the second trick.  In other words, more potential for backroutes if you were not careful.