Huge Levels: IchoTolot's personal Guideline for doing it right

[IchoTolot]

IchoTolot's Guideline for creating huge levels (V 1.2)

As most of you already know:
I am a fan of huge levels, but this is a double-edged sword as they can easily end up frustraiting or simply boring.  So I decided to make a little guideline about how to create interesting and fun huge levels.
So I will list up my personal "do" and "do not" stuff.
A quick note here: Not every good huge level has all of the "do" and no of the "do not" criterias. It's simply depends a lot of the inner workings of specific levels that decide if people like it or not. But in the majority of the cases: Levels that tend to have a lot of "do not" and not so much "do" attributes are often not in anyones (especially mine) favorite list.
There are some levels (like my "Rise of a Digital Nation") that use a lot of space, but do not take a long time. I don't consider them as huge, as huge means for me they use Space + Time. So I only cover the levels here that use a lot of space And take a longer time to complete (like let's say ~5min or more).
Now let's get started with the "do not" stuff

"do not":

1.)
Use tons of builders/platformers and especially long bridges!
Well this is kind of a obvious one as the "tons of builders" levels are often not liked. There are of course exceptions in levels which use the builders in creative and fun ways.
In huge levels a ton of builders will drag the already long level to an eternity and you also have to babysit the builder/s. So use as few builders as possible! In my own huge levels I try to avoid them where it is possible, like make a ramp if the builders doesn't create sth interesting or should delay the Lemming for sth else.
To give a number: Typically try to stay below ~15 (better even 10 or less). If you have more they should REALLY contribute sth amazing!
So try to be creative in the ways "how to replace those builders" and you may even get a few new level ideas.


2.)
Have a tight time limit!
A time limit should be used only where it is needed and for ~95% of the huge levels it is not needed. If a level takes a long time it usually doesn't make a difference if the player takes a min more to beat it, as this kind of levels are mostly more about execution rather than timing and I can guarantee you: The player will rage if they finally got the ~right solution after X attemds and he is missing a few seconds over a 9min level! So he has to play the whole thing again trying to optimise the solution. I made this mistake before in Lemmini and I won't make it again!

3.)
Make X different levels and throw them together in a huge one!
This is a bit of a disjoint levels thing. Multitasking more than ~3 things in a huge area which do not contribute to each other in any way is simply no fun. Don't make stuff like: Catch floaters in section A, while building a bridge in section B.....etc. If you make a level with different parts they should help and interfere with each other and please don't go crazy on the multitasking thing.

4.)
Create a huge level full of little unintresting parts!
If you make a huge level, don't throw together 10XJust dig!. Make a level that really benefites from the space (and may it be just more decoration)! If the player plays through this he should be entertained the whole time and not bored to death over the 9min.

5.)
Too much multitasking!
If your level needs multitasking at several different and far away places, this might easily lead to player frustration. This holds especially in the following situations:
- Requiring multiple skill assignment within a few frames: Players are then forced to frequently jump from one place to the other and back again, because they won't know which lemming needs the skill assignment first.
- Levels with vertical scrolling: In such levels, the minimap at the bottom does not nearly cover all of the level. Therefore players cannot use it to check how far lemmings in different locations are.
- Having multiple hatches and requiring early skill assignments at several of them: The problem here is, that players do not know the order of the hatches. If the level is covered by the minimap at the bottom, this migitates the problem. But it is definitely annoying when combined with vertical scrolling.


And now here comes the do or "tipp" part

1.)
Combine skills, ideas, level parts....everything!
There are a lot of great levels that use this formula. Try to combine skills to create entertaining and mind-blowing parts for your levels, like for example the now old miner+builder trick. Back in the days it blowed my mind and there are still possibilities to create interesting and surprising parts with the old tricks! Simon joked a little bit over tricks like this as he said: "It's a Lemmini style level so I must use the skills where they go on for as long as possible", but he also found this "Lemmini style" highly entertaing.
Think also about how groups of climbers, floaters, glider, swimmers, disarmers and normal Lemmings could interact with each other and create sth really entertaining.
Also what if the disjoint levels now must help and support each other? Or maybe standing the other group in the way? There are also a lot of ideas to be found here.
But be careful! The tricks you use should not get too repetitive!

2.)
Decoration!
If you use a lot of space it should not look like a boring wasteland! Try to create sth beautiful for brains and eyes! But be careful that the decoration do not create backroutes. Be creative and build some amazing landscapes!

3.)
Rather use long skills or different ones than lots of the same!
Use as few skills as possible for your level idea. I do not mean replace that 5 bombers for a basher. I mean don't make 5 little walls to bash through if you have the choice.  Just use one long one as it reduces the attention you need to spend on that part and shortens the skill list (at the end you could end up with like for example 20 bashers). Here the "Lemmini style" also comes into play   what if you only need one long miner here instead of two?
But be careful here that the long skills don't get too long! X Skills that take an eternity to finish are no fun either. You should keep the player busy!
Also if you are in the 5 walls to bash through situation: Think also of 5 different ways to overcome the obsicles! Or to hint at "do" point 1.) combine stuff to get through these!
A good example would be "The Big U-Turn" from PimoLems. It uses just a few skills, but takes advantage of the whole huge map.

4.)
Try to put most of the pixel-precision and timing at the start of the level and not the end!
In NeoLemmix and Lix pixel-precision is not a great thing anymore as we now have the features to quickly redo the part. But if the level goes on you don't want to have the hardest parts at the finishing line or do you? Most of the harder precision and timing parts should be placed at the beginning and middle part of the level while the more moderate parts of this should come at the end. Also if some of the harder parts must be replayed or you want to experiment with a new idea and they are placed right at the end, you must go through all the stuff before that again and this is not very fun to play!


I stress it once more:
This is just my personal guideline with which I made mostly positive experiences. A lot of my and other great huge levels do not have all Do and none of the Do not
things! It often comes down to the specific feel and spirit of the level, but from my experience in general: If a level has a lot of Do not and just a few Do things it is mostly considered not as great.

I may add more points to the post in the future as I certainly missed a few things I had in mind or you also gave me new ideas (that's why I added ther Version number ).

[namida]

3.) Rather use long skills than lots of them!

I'm not sure that I agree with this one. Certianly, in regards to the example you mention, it makes sense (unless there's some kind of twist - for example, it appears you're meant to bash 5 times, but in fact a couple of them should be dealt with in a different way). But I wouldn't say that's always the best idea; especially if similar tricks are used repeatedly to the point they become predictable (case in point - in Lemmings Reunion, one of the first things I'd often look for after a while was "is there any way I can set up the two-builders-to-keep-a-basher-going trick here - and often, there was).

[Nepster]

One more thing I would add to the "do"-part: Keep players busy!
Players want to assign skills to their beloved lemmings to help them along - they do not want to watch them bashing through a big pile of rock for a minute or see them walking for half a minute to reach the next part of the level. Especially in huge levels it is easy to underestimate distances and time needed to bash/mine/dig.
Of course there is (at least) one exception: Freeing a crowd at the end. Some players like to watch them walking past all the obstacles, and all others can make good use of the fast-forward options the current players provide (without a risk to skip over the next skill assignment).

[IchoTolot]

One more thing I would add to the "do"-part: Keep players busy!
Players want to assign skills to their beloved lemmings to help them along - they do not want to watch them bashing through a big pile of rock for a minute or see them walking for half a minute to reach the next part of the level. Especially in huge levels it is easy to underestimate distances and time needed to bash/mine/dig.
Of course there is (at least) one exception: Freeing a crowd at the end. Some players like to watch them walking past all the obstacles, and all others can make good use of the fast-forward options the current players provide (without a risk to skip over the next skill assignment).

3.) Rather use long skills than lots of them!

I'm not sure that I agree with this one. Certianly, in regards to the example you mention, it makes sense (unless there's some kind of twist - for example, it appears you're meant to bash 5 times, but in fact a couple of them should be dealt with in a different way). But I wouldn't say that's always the best idea; especially if similar tricks are used repeatedly to the point they become predictable (case in point - in Lemmings Reunion, one of the first things I'd often look for after a while was "is there any way I can set up the two-builders-to-keep-a-basher-going trick here - and often, there was).

I think I'll put these points also into account in the   "Rather use long skills than lots of them!"  point as the skills also should not be too long as Nepster mentioned and if there are 5 walls to bash through you can also think of different ways to get through them.
Also adding to the combine point that the tricks should not be too repetitive.

Update: 1.1) version is out. Changed Rather use long skills than lots of them! to  Rather use long skills or different ones than lots of the same! as it should capture the point in a better way.

[Simon]

Very good read, I agree with every single point.

If a level is large, it should realize comfortably an idea that's hard/impossible to realize in a smaller level. Nepster has already mentioned about disjoint unions how you shouldn't make levels long merely for that sake.

Some players like to watch them walking past all the obstacles, and all others can make good use of the fast-forward options the current players provide (without a risk to skip over the next skill assignment).

I never watch the crowd. :-] In NL, I mash the 10-second skip until the results appear. An even more extreme form is asking Icho after a failed attempt, "If I make this precisely defined change in the attempt, will that be the solution?", and if yes, move on to the next level.

Watching and listening to a huge densely-packed bunch enter the exit, that's very satisfying for me though.

they do not want to watch them bashing through a big pile of rock for a minute or see them walking for half a minute to reach the next part of the level

Some good long levels have such long skills, but while the skill is active, the levels entertain the player with other action. In this case, both level and skills are long because the idea requires it.

-- Simon

[Nepster]

Playing a recently released level, I (re)discovered another guideline: Be careful with multitasking.
If your level needs multitasking at several different and far away places, this might easily lead to player frustration. This holds especially in the following situations:
- Requiring multiple skill assignment within a few frames: Players are then forced to frequently jump from one place to the other and back again, because they won't know which lemming needs the skill assignment first.
- Levels with vertical scrolling: In such levels, the minimap at the bottom does not nearly cover all of the level. Therefore players cannot use it to check how far lemmings in different locations are.
- Having multiple hatches and requiring early skill assignments at several of them: The problem here is, that players do not know the order of the hatches. If the level is covered by the minimap at the bottom, this migitates the problem. But it is definitely annoying when combined with vertical scrolling.

[IchoTolot]

Playing a recently released level, I (re)discovered another guideline: Be careful with multitasking.
If your level needs multitasking at several different and far away places, this might easily lead to player frustration. This holds especially in the following situations:
- Requiring multiple skill assignment within a few frames: Players are then forced to frequently jump from one place to the other and back again, because they won't know which lemming needs the skill assignment first.
- Levels with vertical scrolling: In such levels, the minimap at the bottom does not nearly cover all of the level. Therefore players cannot use it to check how far lemmings in different locations are.
- Having multiple hatches and requiring early skill assignments at several of them: The problem here is, that players do not know the order of the hatches. If the level is covered by the minimap at the bottom, this migitates the problem. But it is definitely annoying when combined with vertical scrolling.

Added this point in the "do not"  part as "Too much multitasking!".

[Simon]

Migitation of multitasking: The Clones developers got it right in their level design. Whenever there were several workers active around the map, you had lots of time between assignments to different workers. At least if your solution was reasonably similar to the intended one.

Multitasking can be faster if the workers are all very close to each other.

Clones didn't want the player to pause all the time, and disabled the pause for 3 seconds after unpausing. And they prevented assignments during pause. The Clones devs wanted the pause to be used only for the initial planning, and for unplanned breaks. Clones multitasking should not be made easier by pausing.

Because I made heavy politics against this design, they implemented a second pause eventually, called freeze time. Freeze time could be used all the time, and you could assign skills during freeze time. Each use of freeze time worsened the highscore by 1 second. Fine with me, I didn't care about solution time.

-- Simon