|
+ Show Spoiler [Disclaimer] +I previously called this "mixed symmetry" but I've seen people use that term to describe something different, and I also wanted to distinguish this from maps like Strangewood Mire with another weird kind of mixed symmetry, so I felt I needed a new name. My Battle.net tag is Varus, which sounds kind of like Walrus if you know Latin, so that was the obvious choice. Feel free to call it W-sym if you like abbreviations. (By the way, I've had people on ladder tell me there's a LoL player who calls himself Varus and who's very BM. That's definitely not me because a) I'm not BM and b) I don't play LoL. I am, however, the real inventor of the Day[9] transition, first seen in action at Teutoburg back in 9 AD.)
Why this thread?
I've been experimenting with Walrus Symmetry™ since well before the TLMC was announced and I had posted some concepts in the WIP thread. I was just about to post a failed design when I realized this would actually be a great illustration for the typical pitfalls of Walrus Symmetry™ and could maybe spark some discussion. The WIP thread isn't the best place for in-depth discussion on a specific topic anyway, so here we go.
So what is Walrus Symmetry™?
Walrus Symmetry™ is a type of symmetry for maps with more than two spawns (I only consider four player maps in this post, but it could be extended to any number) where one of the spawning locations is "inverted". To better understand what this means, first take your typical four player map (e.g. Frost or Whirlwind) and think of it as a circle. Now divide that circle into several quadrants, one for each spawning location. Draw an imaginary line from the center of the map/circle right trough the middle of one quadrant, and then mirror that quadrant along that line. You now have inverted one of the spawning locations, and your map now has Walrus Symmetry™.
Here are some illustrations, taking Frost and Whirlwind as examples. Note how on Frost every base clearly belongs to one quadrant, while on Whirlwind the edges of quadrants cut through bases:
+ Show Spoiler +
As you can see, regardless of whether you start with rotational or axial symmetry, if you invert one quadrant you end up with Walrus Symmetry™.
Why Walrus Symmetry™?
To answer this question, it is necessary to first understand the real question behind it: Why design any map with more than two spawning locations? Basically, there's two answers for this, and most people seem to support one or the other to varying degrees. The first (and obvious) answer is scouting. Not knowing where your opponent spawns can have a huge impact on the game. Some people enjoy the variety that comes from scouting the wrong base first, but I don't think this alone justifies creating four player maps. The second answer, which in my opinion is the critical one, is the combination of several maps into one map. On Frost for example, which is a map with axial symmetry, every spawning pattern (vertical, horizontal, and cross spawns) plays out slightly differently, so it's basically a 3-in-1 map. Whirlwind with its rotational symmetry is basically a 2-in-1 map (close and cross spawns) plus in non-mirror matchups it also effectively turns into a 3-in-1 map (clockwise, counterclockwise, and cross spawns) because of the inherent racial differences.
With Walrus Symmetry™, I'm going one step further by combining axial and rotational symmetry into one four player map, resulting in a 6-in-1 map, because every spawning pattern plays out different. Of course some of the differences are rather marginal. In the WIP thread, I said about one four player map with axial symmetry but close spanws disabled that it was more of a 1.5-in-1 map rather than a real 2-in-1 map like e.g. Waystation, and of course this largely applies to maps featuring Walrus Symmetry™ as well. Still, I'd say a really good axially symmetric map like Frost can be compared to a pure 2-in-1 map in terms of variety, and that a good map with Walrus Symmetry™ can be compared to a pure 3-in-1 map (although I haven't heard of such a 3-in-1 map yet).
How would I go about designing a map with Walrus Symmetry™?
This is where it gets tricky, and when I say tricky, I mean nigh impossible; because of course you don't just want to create any map with W-sym, you want to create the perfect W-symmetric map, right? Unfortunately, just creating a decent W-sym map is difficult enough. Basically, you combine all the problems of axial symmetricy (A-sym) with all the problems of rotational symmetry (R-sym) and add a few problems unique to W-sym into the mix and somehow try to make it all turn out okay without making the map look 8-way symmetric with no differences between the spawning patterns whatsoever. I mean, you could easily create a map like Red City with Walrus Symmetry™, but what would be the point?
In particular, we need to keep in mind the following issues which I call the Big Four:
- Distance between close spawns
- Positional imbalance between rotational spawns
- Positional imbalance between non-rotational adjacent spawns when expanding to the other side of the map
- Positional imbalance on cross spawns when expanding to identically aligned unused mains/naturals
Issues #1 and #2 are typical issues of axially respectively rotationally symmetric maps, while issues #3 and #4 are specific to Walrus Symmetry™. For some illustration, here's a map I've been working on called Crusader:
(For the time being, please ignore the gold bases as well as any fundamental problems you might see with that map.)
- The first issue is the distance between the closest spawns. There doesn't seem to be any sort of general consensus on what the minimum acceptable nat-to-nat distance is, but on Crusader it's 38 seconds and that is rather short. However, I did three things to help combat problems that typically arise from close rush distances. The first is the placement of the watchtower that overlooks both short ground attack paths. The second is that there's a choke right at the natural base, so you don't necessarily need to defend at some forward choke. And the third is that after the natural, you expand away from your opponent, and because of the backdoor an opponent camping in front of your natural can't prevent you from doing so.
- The second issue is the inherent positional imbalance between rotational spawns, i.e. SW vs NW and NW vs NE. Because of the watchtower placement these need to be considered separatly. The kind-of-symmetric main/nat/third layout helps a lot here, because the player spawning clockwise can pretty much always establish symmetry by taking the backdoor expansion. Still, I'm not really happy with the positional balance here, but I say that about every R-sym map so I'm probably just overthinking things. We'll leave this partly solved for the moment and carry on. After all, this post is mostly about which problems to consider rather than how to solve them anyway.
- The third issue is the positional imbalance when spawning SW vs SE or SE vs NE and expanding to the main/nat/third of an empty quadrant. Luckily, these positional imbalances are rather minor compared to those we just dealt with when actually spawning in those locations, and often automatically solved along those. The biggest difference to consider is that now the positional imbalances mostly concern the late game, so destructible rocks are of minor impact here. Still, because of the largely symmetric main/nat/third layout which most importantly includes a backdoor that makes the main accessible from either side, I think the positional imbalance has been reduced to acceptable levels.
- The fourth and last big issue is similar to the third, except on cross spawns, namely SW vs NE. As you can see, with this spawning pattern the unused mains are aligned towards the same direction — the naturals are both closer to the NE player, and the thirds are closer to the SW player. This is a problem that severely limits W-sym design, because when designing natural and third (or main and natural, depending on the general map layout) you already need to consider that these should be equally viable to take as fourth/fifth/sixth bases for either of the two close spawns. As previously, the symmetric design of the main/nat/third layout with the backdoor solves this problem to some extent, but a certain positional imbalance remains.
In addition to the Big Four, there also exist other positional imbalances in pretty much every spawning pattern. For example, when spawning NW vs SE, the watchtower near the direct path between the two naturals makes the SE player's position more defensible. However, these sort of problems will be different on every map while the Big Four are always major design considerations.
How do I create a W-sym map that doesn't look like Crusader?
Crusader has what I would call the logical W-sym layout, which is 12 bases arranged like hours on a clock with additional main bases in the corners. This seems to combine maximum balance with maximum expansion ambiguity and variety. Sure, there are other possible layouts, e.g. Crusader without the corner bases where the main would be one of the 12 hour bases instead, or a layout like Frost where every base clearly belongs to one quadrant. The problem is that it's really difficult to make any of those layouts work under the constraints of the Big Four. And of course the real task isn't just to design a playable W-sym map, it's to design an innovative W-sym map that would still be great even if you ignored the fact that it has Walrus Symmetry™.
Here are some layouts I've been experimenting with and which could result in decent W-Sym maps:
+ Show Spoiler +Crusader: Inverted Crusader: Execution: Inverted Execution: Weird layout with lots of dead space:
What now?
So this is where we start discussing stuff and hopefully end up with awesome maps. I've pretty much put everything I've learned into this post, but there's one last thing I can show you — a prime example of how not to design a W-sym map:
+ Show Spoiler [Overview] +
+ Show Spoiler [Analyser] +
I'll skip the in-depth analysis and just say that apart from any kind of inherent problems this main/nat/third layout has, the map is also subject to all the Big Four issues except the first. I still think it's a super innovative and super interesting main/nat/third layout, just unfortunately also a super bad one for W-sym.
Also, in case it isn't obvious: The larger a map is and the more bases it has, the easier it is to deal with the intrinsic problems of W-sym. I'm currently trying to design smaller maps than Crusader though, because I don't think we need Alterzim 2.0.
Anyway, I hope this post gave you a few ideas and maybe we'll see some more mapmakers have a try at designing W-sym maps in the future. As a closing question to perhaps get some discussion started: Do you see any alternatives to backdoors for designing a main/nat/third layout around the Big Four? For example, how about a layout where the main ramp extends towards the middle and the natural has two entrances on either side? If you think this is a good idea, how many other bases would there be between adjacent naturals, and how would you keep the map from being 8-way symmetric without introducing considerable positional imbalance?
GLHF!
+ Show Spoiler [Poll] +Poll: Walrus Symmetry?A well-designed W-sym map has a place in tournament map pools. (17) 63% W-sym maps would be fun to play on, but not appropriate for pro matches. (5) 19% W-sym is just a bad idea. (5) 19% 27 total votes Your vote: Walrus Symmetry? (Vote): A well-designed W-sym map has a place in tournament map pools. (Vote): W-sym maps would be fun to play on, but not appropriate for pro matches. (Vote): W-sym is just a bad idea.
|
Gah I clicked the wrong link in the poll.. =( I think this is cool how you put so much thought into it and would definitely like to see how you go about tackling the problems and issues you identified. Thanks for sharing
|
your Country52797 Posts
There doesn't seem to be a lot of breathing room in this sort of symmetry. Otherwise, it's very interesting...
|
This is very, very weird, but it has potential merit. By doing this you give a map all types of symmetry at once, enabling spawn setups like Sidian's 8-spawn map, and then some. The hard part is designing a good map that isn't broken by it, and by hard I mean it's probably something only a couple of people could pull off, I won't be trying it personally. But I like the idea at the core of it.
|
On April 08 2014 05:57 NewSunshine wrote: This is very, very weird You know what's funny? This was the first iteration of Crusader I posted in the WIP thread, and apparently no one thought it was crazy at all. I think people just didn't understand what W-Sym actually means, which is another reason I created this thread.
On April 08 2014 05:38 The_Templar wrote: There doesn't seem to be a lot of breathing room in this sort of symmetry. Otherwise, it's very interesting... It is of course absolutely true that there is less breathing room than in pure A/R-Sym. Part of it, however, is that you'll just notice problems earlier. I see a lot of four player WIP maps posted here where either rotational spawns are positionally imbalanced or close spawns are broken/disabled. It's easy to create an R-Sym map, but it's not that easy to create a balanced R-Sym map either.
I'd say the main problem with W-Sym is that you need compatible main/nat/third layouts. Once you have that, it's actually not that different from creating an A/R-Sym map. I expect Blizzard will eventually get rid of Alterzim but replace it with another Alterzim-sized map, and I think especially that map could profit from W-Sym. Also, on smaller W-Sym maps you could theoretically do stuff like disable close spawns and still have 5 different spawning patterns, so there is some flexibility gained as well.
I've edited the OP to include a few illustrations of W-Sym layouts, maybe others could come up with better ones though?
|
I kinda don't understand why this 2 in 1, 3 in 1, etc. is getting so popular? I am not saying this is bad, I really like that X in 1 concept. But this is nothing new. I remember Ironman's Khalis being called not innovative. So why has this concept now become the real deal?
|
To keep ground distances even to either close spawn regardless of symmetry, you should put the nat ramp on the 45 degree midpoint of the quadrant. Then, include switchbacks and ramps so that the ground distance from the nat ramp to either adjacent quadrant nat is the same.
Here's a possible simple layout, only 12 bases. Main and Nat labelled, third at "3". The heavy black lines are the routes across the map. Looking at this, its clear that nat-nat ground distances can be the same for ALL possible spawn pairs, even with Walrus Symmetry.
|
On April 08 2014 20:07 Phaenoman wrote: I kinda don't understand why this 2 in 1, 3 in 1, etc. is getting so popular? I am not saying this is bad, I really like that X in 1 concept. But this is nothing new. I remember Ironman's Khalis being called not innovative. So why has this concept now become the real deal?
I can think of two reasons. First, it increases the need to scout early, and thus reduces the power of blind cheese. The "2" in two-in-one means that the map may play differently - perhaps very differently - depending on where the enemy is, probably requiring different builds and counters.
Second, and relatedly, in pro-tournaments it offers more variety to viewers and makes it less likely that a map will be overpowered for one race.
|
On April 08 2014 21:07 EthanS wrote:Show nested quote +On April 08 2014 20:07 Phaenoman wrote: I kinda don't understand why this 2 in 1, 3 in 1, etc. is getting so popular? I am not saying this is bad, I really like that X in 1 concept. But this is nothing new. I remember Ironman's Khalis being called not innovative. So why has this concept now become the real deal? I can think of two reasons. First, it increases the need to scout early, and thus reduces the power of blind cheese. The "2" in two-in-one means that the map may play differently - perhaps very differently - depending on where the enemy is, probably requiring different builds and counters. Second, and relatedly, in pro-tournaments it offers more variety to viewers and makes it less likely that a map will be overpowered for one race.
I was not asking for the concept-idea. I was was asking why this is now popular, even tho this idea is old.
|
On April 08 2014 21:01 EthanS wrote: To keep ground distances even to either close spawn regardless of symmetry, you should put the nat ramp on the 45 degree midpoint of the quadrant. Then, include switchbacks and ramps so that the ground distance from the nat ramp to either adjacent quadrant nat is the same. That is, actually, pretty much exactly what I did with my failed layout where the adjacent nat2nat distances are 43, 43.5, 43.5 and 44 seconds. However, the whole point of W-sym is differences between spawning patterns, so I don't think it's necessarily a good idea to keep all nat2nat distances exactly the same. Admittedly, a nat2nat distance that would be okay on a 1v1 map might be too short on a W-sym map simply because not every spawn has such a short distance to worry about. On Crusader for example, the NE player needs to send a worker to the SE main rather early, while a player spawning NW does not have to worry about this and on top of that will be scouted rather late by the NE player. Still, I believe you're fine as long as the closest distance is over 40 seconds.
On April 08 2014 21:01 EthanS wrote:Here's a possible simple layout, only 12 bases. Main and Nat labelled, third at "3". The heavy black lines are the routes across the map. Looking at this, its clear that nat-nat ground distances can be the same for ALL possible spawn pairs, even with Walrus Symmetry. The problem with this kind of layout is the positional imbalance though. Here, when spawning NW vs NE, the NE player would not be able to expand to the third in his quadrant and thus be forced to expand to the (faraway) SE natural instead, which would also give him the SE main for free. Not saying this is impossible to do (I've tried something like that myself) but it is very difficult to balance.
On April 08 2014 20:07 Phaenoman wrote: I remember Ironman's Khalis being called not innovative. Khalis basically has the same kind of symmetry as Strangewood Mire, only less extreme. You could say those maps are the spiritual successors of Antiga Shipyard. On Khalis, the difference between horizontal and vertical spawns lies only in the central area and not in the base layout, at least not until you take a fourth. Strangewood Mire is similar, except there the differences already start with the third base. The point of the Strangewood Mire concept of course is to make an R-sym map into more of a 3-in-1 map like A-sym maps already are. W-sym is something completely different and results in 6-in-1 maps right from the start.
On April 08 2014 21:49 Phaenoman wrote: I was not asking for the concept-idea. I was was asking why this is now popular, even tho this idea is old. As you say, the X-in-1 concept is as old as maps with more than 2 spawning positions are. But I don't think it has become more popular lately, it's just that the current TLMC excludes 1v1 maps, so there's a temporary increase in X-in-1 concepts being posted, and while I experimented with Walrus Symmetry months ago, the TLMC is also the reason why this thread exists now.
|
Very nicely written discussion thread, andG. To be honest the first few times I saw your 4p W-sym maps I thought you had made silly error in copy paste and not realized it. The payoff is so slight in comparison to the standard types of 4p symmetry in any map that obeys the big four that it's difficult for me to see what benefit I should be focusing on and saying "that's the reason why we need to do this!" I admit I like it just because I like cool things, but I can't think of a single reason why you should contort yourself so much to make it work, when the innovation you can get in other areas of map design is so much punchier. What might be interesting is partial W-sym where a certain base location obeys the W-sym but the rest of the map is either milktoast rotational or mirror, so that you get a standard playing field for the early game but the mid-late game has this starkly defined strategic assymetry, which will augment the otherwise standard opening and midgame development. This would give you more freedom in map design since you only have to worry about W-sym for a subset of the map.
I definitely agree that any Alterzim style map would benefit from W-sym. I hope we never see any of the superbig map style anymore though.
Khalis basically has the same kind of symmetry as Strangewood Mire, only less extreme. You could say those maps are the spiritual successors of Antiga Shipyard. I don't think anything anybody here ever made was a spiritual successor to Antiga light-yourself-on-fire-and-jump-off-a-cliff-yard. (Well, I can think of 1 person.) I think the first popular map using shifted symmetry was Shakuras Plateau, which has markedly different distances in the two cross spawn diagonals. imo this is the biggest selling point of shifted symmetry, just the rush distance. Your map is instantly much more interesting when you have 2 long adjacent, 2 short adjacent, 1 long cross and 1 short cross. If you count the adjacent positions as one each where you swap rotational sides, that adds up to 6-in-1. Granted it mostly affects the early game but you see a lot more variety this way on the same terrain, which is ideal for how maps get used in SC2 tournament map pools (such slow rotation, lots of games on the same terrain).
That reminds me to address the fact that I'm pretty sure 90% of pros and casual players will swear on a bible against W-sym without any second thought just on unconsidered grounds of ASYMMETRY DEVIL WORSHIP. Unfortunate but you have a huge stigma to overcome.
|
On April 09 2014 01:28 EatThePath wrote: That reminds me to address the fact that I'm pretty sure 90% of pros and casual players will swear on a bible against W-sym without any second thought just on unconsidered grounds of ASYMMETRY DEVIL WORSHIP. Unfortunate but you have a huge stigma to overcome. I fear you are correct, although I have at least some hope since people rarely complain about the asymmetrical nature of rotational four player maps simply because those maps look symmetrical. So if a W-sym seems more symmetrical than it actually is (Crusader...) then perhaps it won't be burned at the symmetry stake?
On April 09 2014 01:28 EatThePath wrote: What might be interesting is partial W-sym where a certain base location obeys the W-sym but the rest of the map is either milktoast rotational or mirror Isn't that kind of like Antiga/Khalis/Strangewood though?
I mean, this is how I see certain symmetric layouts, with red being axial symmetry and blue being rotational symmetry:
+ Show Spoiler +Whirlwind: Antiga light-yourself-on-fire-and-jump-off-a-cliff-yard: Khalis: (here the red part represents 2-way rotational symmetry, but it's the same effect) Strangewood Mire: (here the red part represents shifted axial symmetry) Frost: Archetypical W-sym layout (e.g. my failed map): Crusader:
I'm not sure what "partial W-sym" would mean in this context, unless you're talking about inverting Crusader in such a way that the bases around the spawning locations are all A-sym (or R-sym) and the middle is W-sym, which would certainly be possible, but I don't really see the additional value compared to the Strangewood Mire layout or an inverted Strangewood Mire layout where the corner bases are A-sym and the middle is R-sym.
Maybe you could post a diagram like EthanS' to show how you envision this "partial" W-sym? Because I don't quite understand how you intend to get W-sym to only apply to one corner of the map.
On April 09 2014 01:28 EatThePath wrote: (Well, I can think of 1 person.) I believe I know who you mean and I really, really like some of his maps.
|
I love those diagrams, haha. I'll try to make an example map because I can't think of it in abstract terms very well.
|
I've been thinking a lot about viable main/nat/third layouts, especially in regards to how it might be possible to have the main not on the symmetry axis between the adjacent mains (like on Crusader) but instead have that symmetry axis go between main and natural (like on Alterzim). Basically, I tried to make the inverted Crusader layout (see first post) work. I played around with a few in-base natural designs with backdoors but ultimately decided that this didn't fit the W-sym philosophy very well. I thought: "It's impossible, I can't have the main on the same level as the natural but still have a ramp between— Wait a minute..."
So I combined logic with crazyness which resulted in this very experimental prototype layout:
+ Show Spoiler +
This layout has a few obvious and a few not-so-obvious potential problem areas that need to be carefully considered:
- The middle high ground should be large enough to accomodate for all the buildings that would usually be built in the main, but at the same time far enough away from the mineral lines that reapers, cannons and air harass units can't shoot workers from the high ground.
- Reaper scouting a two-basing player should be possible, but not too easy — Reapers should probably not be able to circumvent wall-offs by jumping up the high ground directly.
- Blink play: Should it be possible to blink from one middle ground to the other over the low ground at the third?
- The top of the ramp near the main should be within hatchery creep range to prevent three-pylon blocks there, since unbuildable rocks there would just be stupid. Also, one creep tumor to connect hatcheries.
- There should be chokes very close to the main/natural bases for a variety of reasons including close nat2nat distance and attackers bouncing between main and nat, and ideally there should be somewhat more open areas in front of those chokes for Zerg to defend.
You could have a ramp near the base instead of at some forward point, but I think that regardless of which outside third you take you should need to take down rocks to defend it properly, and I think this works better with the forward ramps. Either way, this layout would most likely result in slightly weird things like Zerg walling off their main choke with spawning pool and other early buildings, and if you can blink between the middle grounds then Terran would need to have one bunker at each choke, but could then park their units at the high ground between the bases (where their barracks would be as well).
I think that in principle, this is a good main/nat layout for W-sym that solves many problems. The question is, can this layout be abused in a specific way? Did I miss any points of consideration in my list? Are there maybe any maps with a similar layout?
+ Show Spoiler [Polls] +Poll: The middle high ground should be...As large as possible and cliffs should be near the mineral line (3) 75% As large as possible, but harass units shouldn't be able to shoot at the mineral line (1) 25% As small as possible; production buildings should be built on the low ground (0) 0% This layout won't ever work anyway (0) 0% 4 total votes Your vote: The middle high ground should be... (Vote): As small as possible; production buildings should be built on the low ground (Vote): As large as possible, but harass units shouldn't be able to shoot at the mineral line (Vote): As large as possible and cliffs should be near the mineral line (Vote): This layout won't ever work anyway
Poll: The ramps to the low ground should be...Either is fine; it depends on the map (4) 100% Right at the bases (0) 0% At forward points (0) 0% This layout won't ever work anyway (0) 0% 4 total votes Your vote: The ramps to the low ground should be... (Vote): Right at the bases (Vote): At forward points (Vote): Either is fine; it depends on the map (Vote): This layout won't ever work anyway
|
In this situation I would rather just have a very large ramp leading to the high ground from the main minerals, and then you can wall/defend at the small ramp as usual. The small ramp facing the main CC is just a liability for abusive contain strategies. The only downside of the large ramp is that you get a lot of unbuildable space in an awkward position -- but you already have this with the cliff structure as such. As long as you have an aesthetic theme that properly communicates to the player, it would be fine.
|
I know what you mean, but keep in mind that the small ramp also makes either base equally viable to take as a fourth/fifth in unused quadrants, which is one of the main concerns of W-sym. It's true that you need to keep units outside of the high ground even more than you want to keep them outside of the main on standard maps. The question is, how should this layout be designed to counteract this? Which abusive strategies seem especially strong? Since Terran and Protoss will build most of their unit producing structures on the high ground, this mostly seems to affect Zerg.
|
You know, the original Lost Temple actually kinda had this type of symmetry (the map is generally reflectional, except for the northernmost main/nat.) Granted, this was probably because only 2 ramp directions were supported, but still...
I'd be interested to see a more "faithful" remake of Lost Temple in SC2. (though the highgrounds over the naturals probably would still not work.)
|
your Country52797 Posts
I'm going to try a few maps with this symmetry now that I finished my TLMC stuff, let's see how it goes...
|
|
|
|