What makes a map good?


Hi everyone, I've started a tutorial series on making maps. You can find it on the FAF Youtube channel. I'd like to hear thoughts and opinions what I talk about in this video, where I talk about how you can reward behavior of players by placing your resources strategically. In particular:

That depends on your layout. Personally I like to think in safe extractors, which should be at least 4 / player. Having less is unusual and requires a special build order. Next in line are expandable extractors, which are part of an expansion which should be at least 3 / player (have at least one clear expansion with 3 or more extractors somewhere for each player). This should reward expanding. Then I think about raidable extractors, which can be between 2 - 6 / player depending on your layout. These should be relative far from a base and easy to raid if left unprotected. This rewards having presence and radar / intel. And last contestable extractors which can be between 1 - 2 / player depending on your layout again. This last category is all about map control and rewarding it by having a slight advantage in mass.

Does this make sense, or would you describe it different? Or do you have a completely different view all together?

A work of art is never finished, merely abandoned


Some stray thoughts (assuming 1v1):

  • Relatively clear mex placement, only a few % of mexes should be contested at max. The further away you place mex from spawn the more volatile the earlygame bo phase gets.
  • The bases themselves should be raidable so provide a lot of open space and avoid mex clumps in base. Ie. You'll want the majority of mexes to be within short walk distance with an engi, but the base itself to be open and spread out. This reduces the potential impact of the first bomber or lab, while still allowing for raids later.
  • Provide enough (uncontested) mex so that the game doesn't become dreadfully slow. About 16-24 per player.
  • If you provide a lot of natural reclaim, don't make it contested. Again, big massfields placed far away make the game very volatile. Ie. avoid situations that would lead to an early ACU vs ACU fight over reclaim and/or bomber engi lab fight over reclaim. Good example: Open Palms, Loki
  • Generally you do want to provide at least some reclaim to keep the game interesting. Reclaim speeds up the earlygame while leaving the lategame unchanged, so don't add too much unless that's your intention.
  • Avoid forced transport play on land maps (again, it's leads to too much volatility). Again you have to consider distances here. Eg. Almost any 20x20 landmap where you spawn near the corners will lead to forced transport play (assuming you place any mex/reclaim outside of main base).
  • Good map sizes for land play: 7.5-12.5. In the case of 7.5 you should place spawns in the corners, and on 12.5 you should place players away from the corners. Remember: Too large distance between mex and base: forced transport. Too large distance between players and enemy mex: very large defender's advantage. Too short distances on the other hand tend to limit the game to gun+t1 spam unless you do something exotic with reclaim/mexcount.
  • When you do force transport play (mostly on water maps since navy play asks for more space), make sure that the transport spots can't easily be contested. Example of how not to do it: Pelagial (which is my map btw), Sera glaciers, PoR.... Ie. try to place transport locations an unequal distance from each player. It's also totally fine to place some backward expansions etc. Also making a good path for the acu to the expansion can mitigate issues with transport play. Eg. Roanoke is fine (1st expansion is closer to you than opponent, and second expansion you send ACU)
  • Take extra care to ensure that a decent % of mexes are frigate raidable on navy maps. If it's not worth building frigs to harass then generally a weird dynamic arises where you want navy to prevent opponent from entering navy so that you can eventually build cruisers, but at the same time investment in t1 navy is a waste of mass because it can't attack anything... I prefer just having both players build frigs from the start.
  • Good map sizes for navy: 12.5-20. Again, if you are doing 12.5 then players should start from the corners and have immediate access to navy, but if it's a 20x20 you'll want to place spawns closer together so that defender's advantage doesn't become overwhelming. Eg. White Fire is at the edge of being good, because the players spawn so far away and mex are along the edges, however 10x10 is just a bit too small. Roanoke on the other hand is pretty decent, because the players aren't placed that much further apart even though it's a 20x20.
  • I want the Map to feel natural, yet balanced (most challenging part I guess is an asymmetrical, balanced map)
  • I want to know where the starting mex belongs to.
  • I want more than one bottleneck(attack vector) to fight over. Ideally, 1.5 Lanes per player (4v4 would result in 6 possible Land/Sea attack vectors)
  • I want to decide what strategy I use (Rush, Tech, Adaptive, Turtle) and the map should not dictate only one possible effective strategy
  • I want air battles to keep as moderate as possible. For me, there is no fun having 500ASFs attacking each other

Pretty skybox

put the xbox units in the game pls u_u


@jip said in What makes a map good?:

Or do you have a completely different view all together?

Pretty much every map that does not exclude navy, does not have everything within TML range and actually consistently fills quickly.

If you host a "non popular map" you will get people to join and within 2 minutes leave "because nobody is joining". Within 20 minutes you have 2-3x the required players pass through the lobby...


Big issue with the newer maps I see is the reclaim placement. Often there are 20 spots with 200 mass in rocks. This forces you to manually reclaim them, which is horrible game play. A very good example on how to handle reclaim is cadmium green (asides from the trees). There are 3 patches of different size of rocks per player. This makes it easily reclaimable by attack move and two of those patches are raidable which adds to the dynamic of the game.
If you want to look at a couple more good example of maps, look at the ones that biass did. Most of them play quite refreshing with a distinct artistic style.


Visually clear pathability and significant terrain differences-eg I dont want to find out that tanks can travel through the tall looking cliffs or that my cerberus turret cant hit the enemy t1 arti because I didnt zoom in close enough to see that the flat looking area is actually a hill

M27AI developer; Devlog and more general AI development guide:


What makes a map good is highly subjective. Astro crater has a very simple layout, does not require expanding all around the map, can run at an okay speed on weaker computers, etc. So, some people prefer to play on Astro crater and consider it a good map, while some others think it is an abomination. Many players really enjoy playing on Dual Gap, but many hate it. The same could be said for many other maps. The perspective on what makes a map good can also vary a lot between groups like high-rated ladder players, wonder/hilly/canis/mapgen/etc team game players, Setons players, etc.

So, for example, the matchmaker team has its own general set of standards that it uses when evaluating maps, and even then, things like player rating/skill and team size significantly come into play when judging if a map is good for X. Additionally, there are undoubtedly lessons to learn about what makes a good map (for a lot of players' standards) from maps like Dual Gap, Setons, and Astro that have had such lasting popularity, even if some of those lessons strongly conflict with the standards of the matchmaker team/mapping veterans/ladder pros/etc.

There are many spectrums that contribute to players' opinions about maps; turtle to snowball, clustered mex distribution to spread out mex distribution, low mex to high mex, low reclaim to high reclaim, flat to mountainous, simple to complex, detailed aesthetics to simple aesthetics that require less processing power, large to small, low player count to high player count, no water to entirely water, etc.

While there are strong merits to many of the pros' standards, I have found it funny how big of a discrepancy there is between what most of the 'mapping/ladder pros' think makes a good map and what most players actually want out of a map. Specifically, a lot of players like more eco-heavy/clustered spawn/simple map design with more relaxed gameplay that focuses less on t1 spam-based map control than what the 'pros' generally prefer.

Despite that, I have written a mapping checklist that covers a lot of what makes a map good for a lot of people in a way that is theoretically generally compatible with the pros' standards. However, even it would still exclude a lot of popular gameplay. So, make of that what you will, but here is the checklist:

  • proper slope indication (where is passable vs impassable and buildable vs unbuildable is clearly communicated)
  • no technical problems (slots are odd vs even, proper naming/versioning is used, the map/code works properly, there are no broken mexes/hydros, the map folder is <30 mb, and the map is not too laggy)
  • legal (work is not stolen, not used without permission if permission is required, and not used without attribution if attribution is required)
  • proper texturing (textures don't showcase repeating patterns too obviously, are sufficiently blended, have proper albedos/normals/scales, have sufficient variety, and fit a sufficiently cohesive theme)
  • cohesive aesthetics (terrain/texture/decal/prop/marker/unit/lighting/water/skybox usage and placement work cohesively to form an overall map that doesn't make any of its components look too out of place)
  • tolerable aesthetics (map is not too hard on the eyes, not too unpleasant to look at, and doesn't have any sufficiently significant aesthetic mistakes (such as excessive floating trees))
  • balanced (the map contains no cheats, intends fair balance, does not too heavily favor any factions, and does not confer a much greater level of asymmetric imbalance than Seton’s Clutch does)
  • quantity and placement of mexes/hydros/reclaim supports a variety of each of the following being reasonably viable: strategies, unit compositions, locations of battle, layers of battle (air, surface navy, submersible, land armies, long-range indirect fire), tech levels of battle
  • usage of different slopes/terrain features/water supports a variety of each of the following being reasonably viable: strategies, unit compositions, locations of battle, layers of battle (air, surface navy, submersible, land armies, long-range indirect fire), tech levels of battle

PS: I think moreso in terms of percentages/ratios of mexes rather than concrete mex counts, for what is safe vs raidable vs contested.

waiting for a mapgen-only TMM queue and Godot