Author Topic: A case against 10/15/20-of-everything levels  (Read 1284 times)

0 Members and 1 Guest are viewing this topic.

Offline Strato Incendus

  • He who usually only makes it up to rank two
  • Posts: 645
  • Your graphic-set conversion cleanup service ;)
    • View Profile
A case against 10/15/20-of-everything levels
« on: June 03, 2018, 09:52:10 am »
A lot of forum members seem to be fans of 1-of-everything levels. So am I! :) 3 of everything is also nice, 5 is still fine.

But above that, usually starting with the archetypical "10-of-every-classic-skill level", I just found levels become exponentially less interesting.

The main reason is one that has otherwise been mentioned frequently as a positive of the classic eight skills: The standad eight skills in combination are simply too powerful to work in tandem like that!

Keep in mind, they were designed to be everything you could choose from in original Lemmings - so the skill panel overall had to have this versatility - and then, a lot of levels were determined by which skills you didn't have, or weren't allowed to use. The most prominent case of this happening in original Lemmings is probably at the end of Tricky / early Taxing, where suddenly you can't really trap the crowd with blockers anymore and have to come up with more creative solutions.

Or sometimes, you have to go down without having any miners or diggers, or go up without having any builders. These are the kinds of restrictions that make levels interesting!

However, if you really have 10 of every classic skill, even if you have to save everyone and therefore cannot use bombers, it's still easy to trap the crowd with blockers and just free those with miners or bashers, sometimes diggers at the end. The classic skill panel is a Swiss army knife, and if it's packed with 10 of every skill, you pretty much have a solution for any obstacle you can encounter.

In contrast, a level featuring 10 of every NeoLemmix skill (you'd have to exclude one, since there are nine of them, obviously) is much more restrictive, mainly because the only destructive skill is the fencer, so going down is a problem - and the platformer doesn't gain height, so going up is a problem as well.

I do think though that there are a couple of steps you can take to make your "10-of-everything levels" more interesting; I believe part of the problem is how these levels are created:

Usually, they feature a wide landscape of terrain which is pretty time-consuming to create for a level designer. You end up with a level that looks very pretty, but after having put in all this work creating the terrain, you often do not have the energy anymore to come up with a specific solution. So you just provide 10 of everything and let the players figure it out for themselves.

Unfortunately, navigating through such huge levels is similarly time-consuming for the player, but also not very rewarding on top of that. Because in 90% of the cases, a bit of good judgement and resource preservation (like digging into the ground once and then getting rid of several obstacles with one basher rather than with several ones) are enough to solve those levels.

You don't get that "triumphant" feeling of having cracked a specific code to solve this level, because there's nothing specific to be done here.

Of course, there are exceptions, where a level seems like an X-of-everything level, but actually most of the skills are accounted for, like Nepster's "Final frustration". But the majority of these levels are obviously easier, and therefore their solutions are more generic :) .

So here are my suggestions on how to approach X-of-everything instead:
  • Try a smaller X; 1-of-everything is very restrcitive and thereby most rewarding; 3 or 5-of-everything also still work if 1 is too tight for your level size
  • Does your level have at least X gaps? This binds all the builders to certain spots where they have to be used, making it more of a challenge to overcome the remaining types of obstacles, like traps or walls.
  • Does your level have at least X walls that go up all the way to the ceiling? This means they can't be built or climbed over and at least 1 destructive skill has to be used here (even if you end up building towards or climbing into that tunnel eventually).
  • Is one specific skill provided in much lower quantities than the rest (for example, 10 of everything, but only 1 builder)? This is like the corner stone of a puzzle, the player has to start with thinking about where this one skill goes before worrying about the rest. The downside to this is that it puts all the attention on this one skill, so usually, the spot where it has to be placed can be identified pretty quickly, and then the rest of the level is just rather generic again.
  • Is one specific skill missing from the panel entirely? For example, if the lemmings have to move horizontally a lot, cut the bashers; if they have to move upwards, remove the builders. If crowd control is too plain and simple, remove the blockers; and if you have to save everyone, consider throwing out the bombers for something else, because they don't really have much "red herring" potential in that case anyway.
  • Speaking of something else: Swap out one or two of the classic eight skills for a NeoLemmix skill. It can make a huge difference if e.g. the platformer takes the slot of the builder in an otherwise classic "10-of-every-classic skill" level, because gaining height is much more difficult know. Same with swapping the fencer for miners and diggers, stackers for blockers, etc.

This goes back to the "bias for classic skills" issue, where the new skills receive less attention anyway because people are just more used to thinking and planning with the classic eight skills. While those classic skills have certainly earned their fame, their abundance in 10/15/20-of-everything levels can actually break a level easily. For that purpose, it's actually good to have skills which are objectively less versatile, and thereby less powerful. :)

For example, a swimmer can be seen as a floater that can only fall safely at very specific spots (=where there is water).

If nothing else works, try the following:

Provide yourself with 10/15/20 of every skill while testing
, and freely explore the level as the player would otherwise have to. But rather than solving the level like someone playing on Fun difficulty...
  • ...look for the fastest and most efficient solution, i.e. the one that requires the least number of skills possible.
  • ...try to use skills evenly to get past obstacles, rather than simply using builders for everything just because you can :P !
  • ...if you see an opportunity in the landscape that allows for controlling the crowd without using blockers, use it!
  • ...write down or memorise which skills you used in which quantities.
  • ...and then - guess what? - you remove all the remaining ones!

Now, what started as an open-ended, generic X-of-everything level does in fact have a very specific solution - the one that you found - and the player pretty much has to get into your mind to find it as well. Kind of like with every other, more restricted puzzle ;) .

And that, in my view, is the main purpose of being a level designer:
 
Figuring out any solution that works is the player's job. (This includes backroutes, solved is solved, no matter how.)
But figuring out an intended solution is your job. Do not delegate this task to the player! :)
« Last Edit: July 12, 2018, 08:39:37 am by Strato Incendus »
Ghost Lemmings - help us test a possible new NeoLemmix skill!
My packs so far:
Lemmings World Tour, my music-themed flagship pack, 320 levels - Let's Played by Colorful Arty
Paralems, a more flavour-driven one, 150 levels
Pit Lems, a more puzzly one, 100 levels - Let's Played by nin10doadict
Lemmicks, a pack for NeoLemmix 1.43 full of gimmicks, 170 levels

Offline Proxima

  • Posts: 3370
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #1 on: June 03, 2018, 12:47:30 pm »
I disagree with pretty much everything you've said. And I don't think it was necessary to make your point by shouting so much.

Firstly, 10-of-everything levels are normally used (if at all) for the early ranks of large packs. These levels are not meant to be challenging; they are a gentle warm-up to get the player started without completely stumping them right from the outset. They can have a lot of potential for alternative solutions and challenges (which can now be encoded into the game itself, with talismans). And they can be just the right difficulty if any of us intend to show our packs to people outside the community, or for when new players join.

From what I've seen, 10-of-everything levels in later ranks are very rare, but they do exist. They might be a deliberate "breather level", but are usually either challenging because of the time limit, or have specific features that make the level still challenging even with a large skillset. In this case, it's the level designer's deliberate choice to give a balanced skillset, so as to give the player not much clue about which skills will end up being in short supply and which should be spent more freely to preserve the more valuable skills.

But most of all, I disagree with the idea that a level has to have a specific intended solution to be worthwhile. The "triumphant" feeling you mentioned comes when you surmount a challenge, and levels can be very challenging without being restricted to a single specific solution. A great example is "Labyrinth of Despair" in the Lix community set, which is a 3-of-everything level, and it was considered hard enough to get into the last 10 levels of the pack.

I've talked about this before, but my method for creating levels like this is actually a little similar to what you suggest, except that instead of removing all skills I didn't use, I remove enough to create either a balanced or patterned skillset (e.g. alternating between two amounts). This gives the player a bit more leeway; removes some pointers towards what solution I used and challenges players to come up with their own; and the balance makes the level more aesthetically pleasing (at least to me as the creator). In levels like this, it's not a problem to me if players come up with a different solution from mine, because many of these alternative solutions are ingenious and beautiful; and I know from feedback that the players still found the level challenging and enjoyable.

Online IchoTolot

  • Posts: 2014
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #2 on: June 03, 2018, 12:53:20 pm »
And I don't think it was necessary to make your point by shouting so much.

I think you mistake big letters for shouting.

big letters are used to highlight important points, core statements, headlines or important words in a sentence - just like Strato did here.


Offline Strato Incendus

  • He who usually only makes it up to rank two
  • Posts: 645
  • Your graphic-set conversion cleanup service ;)
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #3 on: June 03, 2018, 01:15:46 pm »
Yes, thanks, IchoTolot, that was precisely what I was going for here ;) , since I know my posts tend to get very long. And especially so when opening a topic.

Quote
And they can be just the right difficulty if any of us intend to show our packs to people outside the community, or for when new players join.

Well, that's the way the Tame rank in ONML went about things: Most of these levels are 10-of-everything levels, and you always only have to use a couple of them to solve the level.

Generally though, I hear a lot of people saying they prefer the Fun rank over Tame, because Fun actually teaches you the skills by restricting the levels to one or two types of skills - thereby leading to specific solutions which the player can remember, and thus they will also remember how each skill works.

Tame, on the other hand, just gives you all the skills and lets you toy around with them for 20 levels. Even to someone who hasn't played our forum packs, but only original Lemmings, this is bound to get dragging pretty quickly. You don't want an entire rank worth of levels which a somewhat seasoned player could just skip over right away without missing out on anything. (I guess that was what Flopsy was expecting from my "Basic" rank in Lemmicks, and then he got caught offguard ^^...)

So I question how much 10-of-everything levels are actually good for the sake of teaching a new player the game. If somebody really never played Lemmings before and starts out with a forum pack right away (maybe because they found NeoLemmix here on this site first), giving all eight skills to them right away might actually be overwhelming.

I find it much easier to create restricted tutorial-type levels that are still a little bit of a challenge.

Quote
From what I've seen, 10-of-everything levels in later ranks are very rare, but they do exist.

My perception may be warped by having played the Sunsoft rank just recently, which definitely includes a lot of these levels ;) . None of them were particularly challenging, though, and Sunsoft is considered the hardest DMA-created rank.

Quote
In this case, it's the level designer's deliberate choice to give a balanced skillset, so as to give the player not much clue about which skills will end up being in short supply and which should be spent more freely to preserve the more valuable skills.

That's what I was suggesting to enforce with my points regarding "are there X gaps" or "X unclimbable walls" etc. ;)

For example, NepsterLems has a lot of X-of-everything levels, and usually the skill that you run out of first is the builder. But one could mix that up a little by throwing in more walls than gaps :) .

Quote
But most of all, I disagree with the idea that a level has to have a specific intended solution to be worthwhile. The "triumphant" feeling you mentioned comes when you surmount a challenge, and levels can be very challenging without being restricted to a single specific solution. A great example is "Labyrinth of Despair" in the Lix community set, which is a 3-of-everything level, and it was considered hard enough to get into the last 10 levels of the pack.

There's a huge difference between 3-of-everything and 10-of-everything ;) . That's why I explicitly excluded 3-of-everything levels from my criticism right away in the first line of my opening post:

Quote
A lot of forum members seem to be fans of 1-of-everything levels. So am I! :) 3 of everything is also nice, 5 is still fine.

Look at it this way: X of everything also means X of the most powerful skills (builder and miner). A level from my first pack rightfully received the criticism by Nepster (who does seem to really like X-of-everything levels, after all ;) ) that the abundance of builders made the level pointless, because one could simply build over all the obstacles, even though it would have been possible to e.g. bash through them.

He demonstrated this to me by solving the entire level with builders, even though that solution took much longer than the intended one. :) That to me was a sign that I had made the level too open-ended. You can leave some freedom to the player, of course - it is very hard to enforce one specific solution anyway. But you shouldn't leave the players so much freedom that they can pick and choose how to get through your level.

In this example, spamming builders in places where other skills could be used as well should lead to the player running short on builders at another point in the level where they are actually needed.

Quote
I've talked about this before, but my method for creating levels like this is actually a little similar to what you suggest, except that instead of removing all skills I didn't use, I remove enough to create either a balanced or patterned skillset (e.g. alternating between two amounts). This gives the player a bit more leeway; removes some pointers towards what solution I used

That is something I like to do as well, thanks for bringing that up! :) I guess one could put it among personal "level designer OCDs", as Nessy calls them. Or one could consider them red herrings to distract from the intended solution (which Arty likes to do a lot).

The problem occurs when the additional skills we throw in to merely conceal the solution end up breaking it. Especially when buffing up the number of builders to that of the other skills, this can happen easily. ;)
Ghost Lemmings - help us test a possible new NeoLemmix skill!
My packs so far:
Lemmings World Tour, my music-themed flagship pack, 320 levels - Let's Played by Colorful Arty
Paralems, a more flavour-driven one, 150 levels
Pit Lems, a more puzzly one, 100 levels - Let's Played by nin10doadict
Lemmicks, a pack for NeoLemmix 1.43 full of gimmicks, 170 levels

Offline Proxima

  • Posts: 3370
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #4 on: June 03, 2018, 01:34:11 pm »
Well, that's the way the Tame rank in ONML went about things: Most of these levels are 10-of-everything levels, and you always only have to use a couple of them to solve the level.

Generally though, I hear a lot of people saying they prefer the Fun rank over Tame, because Fun actually teaches you the skills by restricting the levels to one or two types of skills - thereby leading to specific solutions which the player can remember, and thus they will also remember how each skill works.

Out of 30 levels in the Fun rank, 21 have either a 20-of-all or 30-of-all skillset. Sure, it has the tutorial levels to isolate each skill and show you what the skills do; but then you learn things like blockers for crowd control, dealing with RR 99, and analysing the terrain to work out a route from these 20-of-everything levels. Tame is less interesting than Fun because it doesn't push the boundaries beyond the really simple "pick one or two skills to complete the route" seen already in the very first level.

Quote
I find it much easier to create restricted tutorial-type levels that are still a little bit of a challenge.

This is the route I've gone with the first 10 levels of GemLems. Only the time limit tutorial is a 20-of-everything level, and all the others are restricted and maybe a bit more challenging. But then I doubt this would be suitable for absolute beginners. And yes, making levels that really are suitable for easing beginners into the game is a very difficult art. At the moment, a group of us in the DROD community are working on a beginner-focused level set, and we've gone through a huge amount of testing and feedback to get everything just right.

Quote
There's a huge difference between 3-of-everything and 10-of-everything ;) . That's why I explicitly excluded 3-of-everything levels from my criticism right away in the first line of my opening post

Sure, but then you contradicted this by saying "But figuring out an intended solution is your job. Do not delegate this task to the player!", which would also disallow 3-of-everything levels.

Quote
He demonstrated this to me by solving the entire level with builders, even though that solution took much longer than the intended one. :) That to me was a sign that I had made the level too open-ended. You can leave some freedom to the player, of course - it is very hard to enforce one specific solution anyway. But you shouldn't leave the players so much freedom that they can pick and choose how to get through your level.

And why not? Going back to solved levels and finding different solutions is fun too!

Online IchoTolot

  • Posts: 2014
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #5 on: June 03, 2018, 01:50:11 pm »
I am not against or the biggest fan of "10+ of everything".

They do a bad job in teaching the skills -- tutorial levels are way more suited for that.

They do a good job of training the player basic solving and pathfinding in Lemmings.

So my approach would be: 

- Tutorial levels first to introduce the skills.

- Move over to easy "X of everything" to teach basic solving and pathfinding with lots of freedom for beginners.

- Throw in easy, more restrictive puzzles after a bit to slowly shift over to standard puzzle solving. Also things like RR mechanics, timers, easy tricks and pick-ups come into play.

- In later stages throw in a few "more or less" open-ended hard "X of everything" levels like the "Labyrinth of Despair" or "Final Frustraintion" type.


Also I see quite some similarities in the way I produce a part of my levels to your suggested "Provide yourself with 10/15/20 of every skill while testing" or "how to approach X-of-everything instead" techniques. ;)

Offline Proxima

  • Posts: 3370
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #6 on: June 03, 2018, 02:59:53 pm »
So I went out for lunch just now, and while I was out, I had a train of thought.

Suppose we remove the first seven levels of Fun, and imagine a new player going through the game. What specifically would they learn from each level?

Fun 8-30, level by level (click to show/hide)

In short, pretty much every level teaches the player something new; 19 and 26 are the only levels that are pure filler. And all this teaching is done by allowing the player to discover the mechanics for themselves, and (most of the time) giving them enough leeway that it doesn't matter if they try something that doesn't work out. The Fun set actually turns out to be a really good introduction to the game, which is funny because that happened more or less by accident -- the hard levels were designed first and then the devs just made 20-of-everything repeats to pad out the easy ranks.

(Of course, it's not perfect. A couple of absences stick out: there's no level that introduces steel, until you hit Tricky 9, which is oddly late. And there is no level at all in the 20-of-everything portion, even including Tricky, that introduces miners or encourages their use. So the player may well not get the practice they need at estimating where a miner ends up. Maybe that's why "How do I dig up the way?" was considered suitable for late Taxing?)

Now that I've written this post, I think that making a pack that's suitable for absolute beginners and does the same job of introducing a mechanic or concept with every level would be a very interesting challenge. I might try that at some point :P

Offline Nepster

  • Posts: 1830
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #7 on: June 03, 2018, 05:58:40 pm »
1) The skillset simply has to match the level
I don't think we should outright discourage 10-of-all levels. A 10-of-all level that you can solve with 7 builders and three other skills is most likely uninteresting. But so are 5-of-all levels that you can solve with 3 builders and 2 bashers or any of the Tame levels as a 3-of-all level.

2) A good X-of-all level forces the player to use a lot of different skills
Yes, a level where you can simply build over every single obstacle is not really great. But if you have 5 obstacles each requiring either 3 builders or 3 bashers or 3 miners (or some combinations of such), this can be a lot more interesting.

3) I am not a fan of a X-of-each level with exactly X gaps to build over
This makes it obvious where to put the builders, and if you can have a worker lemming preparing the path, then this only prolongs the level without making it any more interesting. In this case I would just remove the builders completely (or at least reduce their numbers) and fill the gaps with terrain.

4) Swapping some of the classic skills with new NeoLemmix ones does not solve any problems in itself.
There have been quite a few 10-of-everything levels with fencers instead of diggers, platformers instead of builders or gliders instead of floaters. Unless you create a level with nice obstacles, they are still boring.

Upshot is: I think most of the 10-of-everything levels simply lack obstacles and offer too much freedom in their solutions. The big challenge is to allow some freedom while not creating an any-way-you-want level.



Offline Strato Incendus

  • He who usually only makes it up to rank two
  • Posts: 645
  • Your graphic-set conversion cleanup service ;)
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #8 on: June 03, 2018, 06:50:51 pm »
Thanks, Nepster, you're the expert on this type of level, I guess! ;)

My suggestion of "gaps equal to the number of builders" isn't a one-trick-fits-all approach, just like I'm not arguing against all 10-of-everything levels in general - otherwise, I wouldn't be looking for ways to improve them ;) .

It may just be helpful as one possible solution to a level that is currently still too boring to make the number of gaps close to the number of builders - so that there are fewer builders which can be placed freely (but not necessarily 0).

Also, gaps do not always need to be as obvious as an actual abyss ;) . Sometimes you can have gaps that actually are filled with terrain to some extent, but there's no getting out of it, or doing so requires more skills than building over it.

This means that for the intended solution, only a builder can go there - but the player won't know this right away ;) .

Of course one has to use good judgment when swapping out classic for NeoLemmix skills in this scenario. ;) In a level where you have to go down, having platformers rather than builders may actually be an advantage; fencers would be less obvious. And vice versa: Fencers help you go up, so if you want to create a challenge, provide platformers instead.

How did you go about creating "Final Frustration", if I may ask? ;) Did you place each piece of terrain already knowing which skill should be used to get through / over it?
Ghost Lemmings - help us test a possible new NeoLemmix skill!
My packs so far:
Lemmings World Tour, my music-themed flagship pack, 320 levels - Let's Played by Colorful Arty
Paralems, a more flavour-driven one, 150 levels
Pit Lems, a more puzzly one, 100 levels - Let's Played by nin10doadict
Lemmicks, a pack for NeoLemmix 1.43 full of gimmicks, 170 levels

Offline nin10doadict

  • Posts: 322
  • Guy who constantly misses the obvious
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #9 on: June 04, 2018, 02:48:23 am »
Creating an X-of-everything level where X is a fairly big number and still have the level be good is rather challenging. So far I think my best attempt at this has been "Once upon a Lemming" which gives you 16 of various skills but only one builder. The builder's location is pretty simple to figure out, but I think the presence of the talisman that asks you to clear the level without using the builder at all is what makes the level good.
Honestly, restricting the number of builders in these kinds of levels seems fairly important just because of how strong the builder is.
It helps to come up with some way to make the level interesting to navigate or create obstacles that have to be approached in some specific way, or else the level will be forgettable.

Offline Nepster

  • Posts: 1830
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #10 on: June 04, 2018, 04:18:08 pm »
Also, gaps do not always need to be as obvious as an actual abyss ;) . Sometimes you can have gaps that actually are filled with terrain to some extent, but there's no getting out of it, or doing so requires more skills than building over it.
Yeah, in that sense your suggestion makes a lot more sense and I fully approve now of this suggestion. When I first read your suggestion I just thought instinctively of bottomless pits.

How did you go about creating "Final Frustration", if I may ask? ;) Did you place each piece of terrain already knowing which skill should be used to get through / over it?
First I created the left 70% of the level with some path in mind, that uses rather few builders and more of the terrain removal skills. My guideline was to use about 3-4 builders per obstacle (and of course some more in the middle), with about twice that amount for terrain removal skills. For the right part, I had some basic idea. But at first there were rather few obstacles there, to make sure the level was solvable.
Then I spent a lot of time optimizing my solution, mostly adapting the existing solution to require less and less of the terrain removal skills. This allowed me to place more and more additional obstacles on the right side.

But I have to stress, that this is by no means an exemplary X-of-all-skills level! The level was created for a contest that required a 20-of-everything level with a 75% save requirement, where it is possible to save every lemming. And this level was just an attempt by myself to make an as hard as possible level within these restrictions.
So I don't think we should take a level that tries to answer the question "How hard can such a level get?" as a guideline for your average X-of-everything level. If you have to take one of my levels as a guideline, then use one of my 5-of-everything or 10-of-everything level from the first two ranks. In my opinion some of them are much better levels than Final Frustration.

Offline 607

  • Posts: 385
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #11 on: June 05, 2018, 06:39:41 pm »
This is quite an interesting discussion, I think!

Offline Crane

  • Posts: 671
    • View Profile
Re: A case against 10/15/20-of-everything levels
« Reply #12 on: June 06, 2018, 08:20:49 am »
Here's one of my earliest levels that's a 10-of-everything level but not traditional in the sense of just using the tools to forge a path to the exit.  This version is a bit old (although it has the increased save requirement), as my latest version has some extra crystal pieces over the metal so it looks iced over, but I don't have that version on this computer.

For my own take on it, the 10-or-20-of-everything levels make for nice breather levels, are good for practice and experimentation and encourage a different form of thinking, like "how can I get all the lemmings up there" or "I wonder if I can save 100%" and the like.