Different Types of Levels

[WillLem]

Since there's been a lot of effort put into ridding fanmade lemmings levels of things like time limits, execution-based stuff, multitasking, and (in Lix) even a release rate control, it's about time this topic was created.

The push to make all levels basically limited skillset puzzles has become somewhat dreary. Almost every pack I've played recently has been made up of "intended solution only" levels, and it honestly feels like you're being made to paint by the numbers all the time. The genius of Oh No! More Lemmings is that it intertwines the "find the designer's solution" levels with other types - disjoint unions, both large- and small-map multitaskers, timed bomber fiascos, the superlemming level and the flow control level (both ideas that could and should have been developed into other levels featuring these mechanics); many of these ideas are no longer part of the game of some engines, whilst others have been basically rendered way too easy by the engine's UI.

As a community, we shouldn't seek to reduce the game to just the level types preferred by a handful of players/designers. All the best fanmade stuff presents a good mix of ideas (to name a few - GeoffLems, NepsterLems and some of the Lemmings Plus series) and that's exactly what makes them stand out as compelling, interesting and ultimately enjoyable packs.

[namida]

There is still plenty of room for variety even without those things.

Firstly - multitasking isn't inherently bad. It's when it's done in a way that's excessive or annoying, that it gets bad. Examples would include cases where you need to get things really precise at one point for timings to work out later; or when frequent scrolling between opposite ends of a large level becomes necessary; or when there's too many lemmings (especially if you can't easily have them all on screen at once) to keep track of.

Likewise, time limits shouldn't be seen as forbidden, simply as "don't slap them on just for the sake of it" - and for that matter, "execution difficulty doesn't become justified just because a time limit is the cause of it". Much like just because you can put a trap in every level, or because you can use the basher on every level, doesn't mean you should do these things - time limits, too, should be used because they add value to the level, not used to make a level annoying or just for the hell of it. (There is some grey area around cosmetic reasons, like Tricky 21 or Mayhem 3 where the time limit fits a greater pattern on the level - although in the latter case I'd also argue it adds to the level in and of itself too. My personal thought here is that it's fine if it clearly allows plenty of surplus time, but should have quite a clear thematic reason to it, not an obscure reference - the reason for this being so that it doesn't reduce the impact of seeing a time limit on other levels where it might actually be relevant.)

Release rate, I absolutely understand where Lix is coming from on that. If there was not the concern of breaking existing content, I would certianly be simplifying NL's release rate system a lot - not to the extent of Lix where there's no adjustment whatsoever, but more likely either (a) reduce the number intermediate steps between slowest and fastest, or (b) instead of allowing free adjustment, only allow switching between the initial RR and 99. However, these ideas - especially the former - would break a lot of existing content at this point, so even if it weren't for the "no new changes" situation now, the change couldn't really be justified - indeed, I'm pretty sure it came up in the past. However, again, it really comes down to that it depends how it's used. Slapping a tight time limit (including a de-facto one) just to force the player to RR99 within a certain time, where no other reason exists to force this, is annoying; levels like, say, Flow Control from OhNo, are good.

Execution difficulty (where it's done for the sake of execution difficulty, rather than because the solution needs a trick that's inherently difficult to execute) is the only one of those things that I'd say is always bad design. I get that it does appeal to some people, but ultimately, levels focused on execution difficulty is one thing that does fall outside NL's niche, so it's still bad design in NL regardless of personal opinions on it in a wider lemmings context.

As for limited skillset puzzles, there are always ways to add more variety to this. In particular, some limited skillset levels - even very hard ones - with multiple solutions, do exist. One particular example to come to mind from my packs is Sharp 19 of Lemmings Plus Omega II, which has two solutions, both of which require very advanced tricks (one of them I haven't actually seen required in any other level, though I've seen challenge solutions using it), but are very different from each other. It's also worth noting that limited skillset doesn't always mean "exactly the skills you need" - Lemmings Plus II in particular, and especially the Cheeky rank, has a lot of levels that give more skills than you need, but still far from X-of-everything.

As a specific counterpoint that you're very familiar with - let me point out that your Lemminas pack was quite well received despite having very few "specific solution" levels. Yes, a few individual levels got called out as bad, but the pack as a whole was very well regarded.

[IchoTolot]

I think you are missing a big point here: People are already creating and playing the packs they want to create or play! Even if they go against guidelines that usually don't hold them back anyway trust me.

Just go ahead and create what you want and if you want timed bombers - use the Lemmix editor and create Lemmini levels!

Nobody is stopping you! If people enjoy that type of content they will come.

Still, a lot of people enjoy the Limited Skillset Puzzle type and therefore a lot of them are being created.

Demand and Supply.

As a community, we shouldn't seek to reduce the game to just the level types preferred by a handful of players/designers.

If you do not enjoy it --- create different content! The right audience will come!

But somebody needs to start!

One thing though:

Since there's been a lot of effort put into ridding fanmade lemmings levels of things like time limits, execution-based stuff, multitasking, and (in Lix) even a release rate control, it's about time this topic was created.

This is only partly true for NL and Lix!

All versions of the original games, Lemmini, SuperLemmini and to a certain degree Lemmix still cater to execution centric gameplay.

Just because two engines chose differently doesn't mean that it is being ripped out from all engines. The majority of engines still is execution focused.

How about the other way around: Why must every engine cater to the execution centric side? Why can't there be two engines that divert from it?

I was a Lemmini creator at first, but everything that annoyed me about Lemmings back then was centered around execution. All my ideas were hindered by it and with the development of newer engines I finally could play how I always wanted to play. As a result, I switched engines.

In Lix and NL I can mostly be sure that the focus is on the puzzle. If I wanted it the other way around I would have stayed with Lemmini.

Today I would outright refuse to play execution focused packs and therefore I focus on the engines that do not support them. Let's say all the genius you mentioned from Oh No! More Lemmings I am very glad I do not need to do it again.

It is more the case that over the past people just enjoyed the other puzzle centric style more and as a result more content was created for it.

Not every engine must include execution centric ideas! Here I would advice for the variety by saying that NL and Lix should mostly stay free from it.

[WillLem]

cases where you need to get things really precise at one point for timings to work out later; or when frequent scrolling between opposite ends of a large level becomes necessary; or when there's too many lemmings ... to keep track of

For me, it depends on the type of level as to whether these things are annoying. If the level has an extremely limited skillset and only one possible solution, and requires any of the things you've mentioned, then yes - it's a bad level. However, if the level offers many skills and invites the player to craft their own solution (which may still be difficult depending on the layout of the level, etc), then this can be an interesting and engaging kind of level to play through IMHO.

reduce the impact of seeing a time limit on other levels where it might actually be relevant.

I've come around to infinite timers for most levels tbh, mainly for this exact reason; it's better to be clear to the player when time is a relevant factor.

That said, I do still enjoy those moments when you realise that there's more to the level than crafting a solution, and that other steps must be taken to also beat the timer. It's part of the experience of playing through a pack, a little bit of cheek from the designer can be compelling if done tastefully and in good spirits.

Release rate, I absolutely understand where Lix is coming from on that
---
either (a) reduce the number intermediate steps between slowest and fastest, or (b) instead of allowing free adjustment, only allow switching between the initial RR and 99

Why? Honestly, the latter sounds like a terrible idea (and one that doesn't remove the "get the RR to 99 just to beat the time limit" issue). The former could be workable but seems unnecessarily restrictive.

Sometimes, variable spawn can be useful for more than just switching to RR99. You may want a couple of lems to be closer to each other than the spawn rate (but not quite RR99), and variable spawn is a quick, easy way of achieving this that doesn't cost any skills. It gives both the designer and the player more options than just those presented by the skillset.

I can understand most of the arguments against "non-NL/Lix-friendly level design" even if I don't always agree, but the RR thing is something that genuinely baffles me. I really don't see why Simon is against including it, and I have read and re-read his reasons multiple times. I think I'd need a practical demonstration showing that it isn't needed, but even then it's so easy to counter things like "use Jumpers instead" with "yes, but then you're using up skills". I really just don't get it at all

Execution difficulty (where it's done for the sake of execution difficulty, rather than because the solution needs a trick that's inherently difficult to execute) is the only one of those things that I'd say is always bad design

I can understand why some players don't like this tbh. NeoLemmix (and indeed Lix) have done a great job of making the game far less execution-focused, at the cost of (perhaps) increasing the prevalence of fiddly pixel-precise stuff.

As for limited skillset puzzles, there are always ways to add more variety to this. In particular, some limited skillset levels - even very hard ones - with multiple solutions, do exist

Whilst I don't doubt that this is true, my issue is more with every level in a rank or pack relying on limiting the skillset to increase the difficulty of the level.

It's also worth noting that limited skillset doesn't always mean "exactly the skills you need"

I know, I guess I'm just using it as a shorthand for any level which enforces a particular solution by limiting the skillset. And, whilst these levels are sometimes amongst the best any pack has to offer, too many of them back-to-back can start to feel repetitive and off-putting.

Lemminas pack was quite well received despite having very few "specific solution" levels.

True, I appreciate the reminder I guess the community was ready for something a bit easier and I tried to deliver that with Lemminas. I'm hoping the sequel will be even easier, but more solid in terms of overall quality.

I think you are missing a big point here: People are already creating and playing the packs they want to create or play! Even if they go against guidelines that usually don't hold them back anyway trust me.

Just go ahead and create what you want ... Nobody is stopping you! If people enjoy that type of content they will come.

Fair enough, that's a good point. Supply and demand is definitely a good gauge for these things.

I guess I just get a bit disappointed when I dive into a new NL pack only to find that the designer wants me to find just that one solution to every level in the pack. I need a bit more persuasion than that to spend the time that it takes, and I find that a few easier levels or ones which test skills other than puzzle solving can be just the ticket.

How about the other way around: Why must every engine cater to the execution centric side? Why can't there be two engines that divert from it?

There can, clearly! And I personally use NeoLemmix more than any other engine as my go-to for Lemmings gameplay generally, so I'm not completely averse to the "no execution stuff" mentality. However, since I also enjoy that kind of gameplay from time to time, I'm the one who must be running two or more Lemmings engines side-by-side (I currently run 5!). The ultimate solution (for me) would be one engine which caters to all, and both NeoLemmix and SuperLemmini have the potential to do just that in user preferences.

In NeoLemmix, in fact, it is now possible to remove all player-assists and play in a much more old-school way thanks to namida implementing requests for these to be toggled off if the player so wishes. If Superlemming and Frenzy were still a thing, then it'd be the complete package! But, we can't get everything we want

In Lix and NL I can mostly be sure that the focus is on the puzzle
---
Not every engine must include execution centric ideas! Here I would advice for the variety by saying that NL and Lix should mostly stay free from it.

Well, you got your wish! Both of these engines are ideal for more finely controlled, puzzle-focused gameplay. The only thing I would say is that the cost of this is pixel precision, which is now way more of a problem than it ever was on Amiga or Windows, even if the level designer doesn't intend for it to be.

[namida]

I really don't see why Simon is against including it, and I have read and re-read his reasons multiple times.

One thing to keep in mind is that Lix's philosophy is that the game should specifically revolve around assigning skills to lixes; and that the player should have no other way to influence the physics (with nuke - which IIRC is purely cosmetic in Lix - basically being a fancy "quit" button rather than a gameplay feature). This differs from NL where release rate and nuke have also been decided to be proper gameplay mechanics in their own right.

As a general rule: It's usually safe to assume that any philosophy NL has in this sort of direction, Lix has an even stronger version of (at least for singleplayer).

[Forestidia86]

While singleplayer Lix is generally puzzle oriented. Lix has an execution heavy multiplayer. And it shows with the skills like jumper, batter, walker. (Theoretically you can play a singleplayer level (with some tricking) in multiplayer mode to get rid of the quality of life stuff like pause and framestepping.)
So Lix' design is even to some degree centered around making execution possible and worthwhile to support multiplayer.

[WillLem]

I've changed the title of this thread to make it a discussion about different types of level in general. So far, the following have been identified. Note that whilst it is possible for a single level to encompass several of these types simultaneously, most levels can be identified as predominantly having a main type:

1. Limited skillset puzzle
2. One-possible-solution puzzle
3. Open-ended/multiple-solution puzzle
4. Limited time challenge
5. Execution-based challenge (timed bombers, superlemming, etc)
6. Disjoint/fake disjoint union

So - for example, the level Inroducing SUPERLEMMING is a limited skillset, execution-based challenge. Also, whilst it can be backrouted, it (most likely) only has one intended solution. So, it can be said to incorporate types 1, 2 and 5. However, it may be agreed that it is predominantly a "type-5" level, since the execution-based challenge of playing the level at superlemming speed is its primary feature.

My stance is that it's possible for all of these level types to be equally interesting, compelling and playable; ideally, packs would feature several of each type of level in order to offer the player more variety.

[IchoTolot]

I've changed the title of this thread to make it a discussion about different types of level in general. So far, the following have been identified. Note that whilst it is possible for a single level to encompass several of these types simultaneously, most levels can be identified as predominantly having a main type:

1. Limited skillset puzzle
2. One-possible-solution puzzle
3. Open-ended/multiple-solution puzzle
4. Limited time challenge
5. Execution-based challenge (timed bombers, superlemming, etc)
6. Disjoint/fake disjoint union

My take on this is create what you want, have some variety, but do it well!

1.) I don't mind easier Open-ended/multiple-solution puzzles and they can be really enjoyable, but do not give me a huge monoton builderfest fot example. Mix it up! Also, if they look amazing it's a big plus and you have something to look at while you are plowing through it.

2-3.) One-possible-solution puzzles and Limited skillset puzzle should be fair as they are already hard enough and should not have addirional weight like hidden elements, unnessesary timers and unnessesary pixel-precision.

4.) Limited time challenges should really make sure that the timer is felt and one of the key elements of the level! Not just a reason to release the crowd earlier.

5.) Execution-based challenge (timed bombers, superlemming, etc) should still be careful that the solution is not too much of a trial and error. Like be careful that the assignments not only function at an exact frame.

6.) Disjoint/fake disjoint unions should make sure that the effects of one part of a level are felt globally and not locally. Do I really need the builder in this part?


With these points certain level types won't stick out badly and the variety aspect is highlighted more. Also, the chance of "oh not another X type of level" is reduced.

[Proxima]

I've changed the title of this thread to make it a discussion about different types of level in general. So far, the following have been identified. Note that whilst it is possible for a single level to encompass several of these types simultaneously, most levels can be identified as predominantly having a main type:

I don't agree with this. I think there is a scale of open-endedness, in which we could identify the following gradations:

1. Unrestricted (20 of everything, generous or no time limit, generous save requirement)
2. Slightly restricted (e.g. 20 of everything except builders, or 5 of everything, or a tight time limit)
3. One main idea but plenty of room for variation in how the idea is carried out
4. One specific solution with every skill having an intended place to use it

The other categories you've identified don't fit into this scale and are orthogonal to open-endedness. A disjoint union, for example, can be anywhere from unrestricted (Tricky 10) to slightly restricted (Fast Food Kitchen) to one specific solution (Synchronised Lemming). Still, it's an interesting point that if a pack (or rank) has a lot of single-solution levels -- or for that matter, a lot of open-ended levels -- having an occasional disjoint union is one thing the designer can do to add some variety.

In fact, I would say that the real moral lies in this direction. You've made a topic before contrasting the Tame levels with the Fun levels. There are a lot of reasons why the Fun levels feel better-designed than the Tame ones, but one of them is that all 20 Tame levels feel the same -- there is very little variety in what you have to do. The Fun levels are (except for 13 and 18) all open-ended, all 20-of-everything, but they have a lot of variety. Similarly, one pack (or rank) full of single-solution levels can feel a lot more interesting and varied than another such pack, by changing what you have to do from one level to the next -- one may be large and full of teleporters, another may be a disjoint union, another may be a thin vertical level.

Conversely, a pack may choose to do without such extremes of variety, by restricting itself to (for example) small, compact levels, like Simon's "Miniatures". Such a pack could still have a lot of variety if the levels use different tricks and ideas! While there is definitely a place for all types of levels (except ones that don't work in a particular engine, like hidden-exit levels in NeoLemmix), one single pack doesn't have to cover all bases.

[WillLem]

1.) I don't mind easier Open-ended/multiple-solution puzzles and they can be really enjoyable, but do not give me a huge monoton builderfest fot example. Mix it up! Also, if they look amazing it's a big plus and you have something to look at while you are plowing through it.

Agreed! Interesting maps is key to these levels working well.

I think there is a scale of open-endedness, in which we could identify the following gradations:

1. Unrestricted (20 of everything, generous or no time limit, generous save requirement)
2. Slightly restricted (e.g. 20 of everything except builders, or 5 of everything, or a tight time limit)
3. One main idea but plenty of room for variation in how the idea is carried out
4. One specific solution with every skill having an intended place to use it

The other categories you've identified don't fit into this scale and are orthogonal to open-endedness. A disjoint union, for example, can be anywhere from unrestricted (Tricky 10) to slightly restricted (Fast Food Kitchen) to one specific solution (Synchronised Lemming). Still, it's an interesting point that if a pack (or rank) has a lot of single-solution levels -- or for that matter, a lot of open-ended levels -- having an occasional disjoint union is one thing the designer can do to add some variety.

I see what you mean, my level types include an open-endedness scale of sorts alongside level features or elements which are independent of this (which wasn't my intention, but I can see why it would be interpreted that way).

Using your scale, an ideal pack would rotate between all 4 types of level throughout, with difficulty being determined by the puzzle or the map. Many packs seem to be exclusively focused on type 4, with many a few type 3s thrown in. I praised ONML for its variation in the OP, but looking at it again, I can see that it is mainly made up of type 3 and 4 levels This proves your point that the other elements in my list (timers, superlemming mode, disjoint unions) which don't fit into the open-endedness scale can help to make a pack full of (mostly) type 4 levels feel more varied.

So, maybe it's the open-endedness of level that needs a bit more attention. Each of the types you've listed on your scale can vary wildly in difficulty, even type 1 (given a complex enough map!), so it's absolutely possible for all difficulty levels to be catered for.

[Forestidia86]

I've changed the title of this thread to make it a discussion about different types of level in general. So far, the following have been identified. Note that whilst it is possible for a single level to encompass several of these types simultaneously, most levels can be identified as predominantly having a main type:

1. Limited skillset puzzle
2. One-possible-solution puzzle
3. Open-ended/multiple-solution puzzle
4. Limited time challenge
5. Execution-based challenge (timed bombers, superlemming, etc)
6. Disjoint/fake disjoint union

[...]

My stance is that it's possible for all of these level types to be equally interesting, compelling and playable; ideally, packs would feature several of each type of level in order to offer the player more variety.

Not sure if this would always be a good idea to put everything in one pack since execution based and (hard) puzzle based levels can maybe appeal to (partly) very different audiences and need different moods to tackle them. At least if it is on the hard side of both. Dedicating packs to different play styles would be the other option. 

[WillLem]

execution based and (hard) puzzle based levels can maybe appeal to (partly) very different audiences and need different moods to tackle them ... Dedicating packs to different play styles would be the other option. 

Personally, I like all play styles. So, an ideal pack for me would be one which offers some puzzle-based stuff, some action-based stuff, and some completely new stuff!

I've seen a lot of interesting and fun ideas explored in the world of fan-made levels, in both puzzle-oriented and action-oriented styles. Many different types of levels are possible, not just "find the intended solution" ones; it would be sad to see Lemmings be reduced to a series of picture puzzles that can be solved without actually playing through the levels.

I just like to see designers continue to keep exploring all of the possibilities that a game such as Lemmings offers, and my favourite packs tend to be those which either offer a good deal of variety, or explore a particular theme or concept.

As Icho and others have said - each to their own. All designers will make levels and packs which they themselves find pleasing, and the likelihood is that others will also enjoy them as well.

The goal here is mainly to promote a bit of variety, and hopefully inspire designers to try something different from their usual level-making style. Some designers won't need this, but others may benefit from it if they're experiencing a creative block, for example.