Competitive Map Making Discussion

and what makes a good map

Announcements and discussion about community-run events.
User avatar
Blackened
Posts: 347
Joined: Sat May 21, 2016 6:27 pm

Post by Blackened »

Gave is a quick look and I agree with pretty much pizza already said. I even like the capturable civilian buildings to provide sight. Without any oil derricks it isn't necessary to go rax first, but in doing so you get easy map control and long term static vision (should help with soviets!)

Considering its a WIP it's pretty well done and I think would produce great games.

Only other minor concern is the 5 ore mines for each player seems on the low side to me, but that would just need actual gameplay testing.

Well done Omnom :)

So if I mathed correctly, the spawn distance is 62.23, average ore patch is 47.45, and that gives it a ratio of .76?

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom »

Blackened wrote: So if I mathed correctly, the spawn distance is 62.23, average ore patch is 47.45, and that gives it a ratio of .76?
Yep, I spent a while moving the ore and spawn distances around to come up with a map that has a ratio of around 0.75-0.76. When I was testing all of the current maps, I didn't come across any maps that had this particular ratio, so I thought it would be something worth exploring; maybe both tech and blob play are viable on this map?

As far as aesthetics and money situation goes, I think those two are going to be very easy to change as I test this map (if I ever come across any time to log on).

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom »

I've managed to play a few games on my map, and I've quickly mapped out the most important expansions/locations on the map. The red spots are the spots which I believe high value, the blue spots are of medium value, the orange spots to be of lesser value, and the grey spots are locations of interest. The size of the circles have no meaning, they're just there to mark the general area. For comparison, I've analyzed Mo's new map and Patches in a similar fashion. Both my map and mo's map have a spawn-average mine distance ratio (R) of around 0.75, but mo's map is about 33% bigger than my map. Patches has an R of 0.62. Using these locations/zones, I'll also quickly comment on terrain and how you can predict the progression of the games on these particular maps.

Image

In terms of areas of importance, it shares a lot of similarities to Sidestep. In a mirror-color match up, where both players go to the same spot relative to their position, the game follows a linear progression, much like how the old meta of double-expanding used to play out. However, one thing to note is that there is a Rock>Paper>Scissors dynamic with the red, orange, and blue spots, respectively. If Player A sets up at the Player B's red location, and Player B sets up at his natural orange location, Player A player is going to have an advantageous position. However, if player B sets up at his natural blue location instead, he will be able to turret/tesla camp that middle ore while gaining another ore mine. But if player B went straight for the blue location, and player A goes for any orange location, he will have a double mine advantage (I've changed those patches to be double ores). Sorry to be confusing, but I hope it makes more sense with the picture.

There are a lot of possibilities and ways the game could develop on this map because the only zones that are next to each other are the red zones. I'm trying to think of a way to design a new map where I make it so that none of these zones are touching each other...as people have said before, perhaps breaking this symmetry will make for better maps.

In addition, the terrain between all locations of interest is varied and dynamic. There are lot of different ways to move around the map, and each spot has its own advantage, with no "outstanding" advantage that defines a location as "must get there first". In fact, it's actually much smarter if you're able to scout where your opponent is going to expand to (the capturable civ buildings helps a lot with this) than to just quickly take your own expansion. In addition, each location allows unique access to at least 2 different locations.

Image


The single middle road in Mo's map makes the red locations of do-or-die importance; if I were to get my MCV in the opposing player's red location, I'm going to win more games than I would be losing. I have enough money to send my 2nd MCV straight to that location because I start with 4 ore mines (questionable, but interesting). The other zones play no significant role until the war over who controls the middle passage is decided.

Also, the terrain between each location of interest is extremely limited. Some direct passageways are only accessible by naval or air, which requires a significant amount of time and resources to use, and some do not grant any access to other locations at all (they have 0 map control). In addition, the progression of locations of importance facilitates passive play because the players are encouraged to develop away from each other (Red first, then Blue).

If the game were to progress in a mirror-colored fashion, then the game becomes extremely campy -- there is nowhere else to maneuver. Everything is funneled into the middle, the north bridges, or the south bridges, which would most definitely favor high tech and naval play. It may not be intentional, but the "Singles Effect" would most definitely happen if the game is not ended quickly. Some may not consider that to be such a bad thing, but in my opinion, I would rather just surrender than wait 10 minutes for a nuke. In my opinion, this is why maps should not cater to only high tech/naval play -- we should not have to be forced to spend 5+ minutes diverging massive amounts of resources to expensive units, only to lose all of them in 30 seconds.

Image

Let's take a look at patches. The red zones here are way more important than they are in Mo's map; if your opponent gets to your red location, you're not going to win the game. You get an main ore mine camped, he gets to place a refinery behind his tesla/turret camp for a net +2 difference in ore mines, and you're forced into 2 moves -- all in at his main, or all-in at your red location. Chances are, you'll get out-eco'd before either move works. All 3 maps have extremely important red locations, but the value of the red zone in Patches is terrible map design (not to mention the fact that your harvester is going to bug out and long distance mine to that red zone patch if you don't place a refinery there).

The terrain on Patches is also not well designed. These chokes are just asking to be camped by static defense, and with the low-eco situation, it makes this area even more attractive.

In a mirror-color match up, the game then turns towards the 2 middle patches. However, these patches are extremely vulnerable...most of the resources you get from these patches go towards pillboxes/flame turrets to shore up your weakside. Now, you're faced with the opposite situation that you have in mo's map -- you have too many places to go to, but neither location offers much else other than "dangerous" money. However, you need to get this money, because if you don't expand, and you lose your main army, you wont have enough money to rebuild. If you expand, but you lose your expansion, it's going to be extremely difficult getting that expansion back. As a result, the game becomes extremely fragile on Patches, as one false step is going to set you back tremendously. This is not something I would consider conducive to positive gameplay...players should be allowed opportunities to recover, not forced into "do or die" situations.

Patches, to me, represents one extreme of how maps could be designed, and Mo's map is the other extreme. Some people like playing on these maps, but most people don't. In summary, I think map makers, before placing a tile on the map, really should consider how to mitigate the "get MCV there first" syndrome without forcing the map to become a mirrored matchup. I don't mean to say that my map is the paradigm map, but I designed it purposely to create as many scenarios as I could. Maps like Patches, NW Passage, Singles, KotG2, and many other maps are too linear in design and creates "one" type of game, repeatedly, without fail. The game already progresses in a linear fashion -- lets avoid making maps that reinforces this linear design.

User avatar
Doomsday
Posts: 199
Joined: Thu Nov 10, 2016 11:45 am
Location: Helsinki

Post by Doomsday »

This thread certainly gives me inspiration to spend more time in map editor.
The supreme art of war is to subdue the enemy without fighting.
-Sun Tzu

User avatar
Sire
Posts: 26
Joined: Thu Jan 12, 2017 3:03 am

Post by Sire »

I sorta got back into mapping for OpenRA and decided to try and make a reference for Ore Fields and the respective money amounts. I used the included editor's function for counting how much money is on the map to determine these values. I do not know if they are accurate or not, but I hope it proves useful!

Edit: Fixed Image

Image
Last edited by Sire on Fri Mar 31, 2017 4:40 pm, edited 1 time in total.

abcdefg30
Posts: 641
Joined: Mon Aug 18, 2014 6:00 pm

Post by abcdefg30 »

Why is "6x6 + 2T" $200 less worth than "6x6"? Looks suspicious to me.

User avatar
Sire
Posts: 26
Joined: Thu Jan 12, 2017 3:03 am

Post by Sire »

abcdefg30 wrote: Why is "6x6 + 2T" $200 less worth than "6x6"? Looks suspicious to me.
Thanks for catching the error! I don't remember how I got the 6x6 value, but it is actually supposed to be $8,600. I fixed the picture.

OMnom
Posts: 255
Joined: Fri Oct 28, 2016 4:04 am

Post by OMnom »

Now that I think we've somewhat settled on the size and features of a good map, I think it's time to turn our attention to the last, major part of map making....the money.

Lately, a lot of maps have been experimenting with adding extra money in various different ways. From what I can see, there are three ways to change the amount of money on the map: 1) adjust the total ore on the map, 2) adjust the number of ore mines on the map, and 3) adjust the number of expansions on the map

I've sampled a few maps and measured how much money are on these maps, as shown below.

Image

Unfortunately, I don't know how much money is generated by growth...couldn't find it in the yaml files, and I have no way of measuring the growth as well. Also, I'd like some verification the starting ore on Rocky Ravine...every time I measured it, there was a 1.9k difference between the starting spawns.

I believe that as a general rule, players should start with around 18k in starting resources in order to allow for different types of games. For reference, a 2ref SD build with 2 harvesters costs 11.3k, and a 3ref 3 harvester build costs $13.8k, just in infrastructure. A map with 18k in starting resources will give the player approximately
400 seconds of total mining time after a 3ref 3 harvester start. With 6 harvesters, this amounts to approximately 1 minute of full mining capacity before the starting ore mines run dry.

Some things of note:

-Sidestep spawns start with 32% of the total ore on the map, and has the least number of expansions (4) to fight over.

Continuing to use Sidestep as reference map, it seems that giving players a larger percentage of the total ore to start with might have encouraged more aggressive and reckless play in the beginning. Also, since there are fewer expansions on this map, the value of each expansion increases by simple math. In combination with the previous discussion on size and map features, this means that with each MCV movement, you control at least an additional 25% of the remaining expansions on the map. Depending on what definition of control you abide by, you could argue that with 1 MCV move, you could control more than 50% of the map (camp the center ore and stop him from expanding top = 50% of the remaining money + the % money remaining in the center).

For comparison, on Patches, each spawn starts with 22% of the total ore, which means that the remaining 56% of the ore is distributed among the other 8 patches. This results in each MCV movement giving you 7% of the remaining ore. This further supports my previous claims about the "red zone" in Patches being of ridiculously high value relative to the rest of the expansions on that map. since that is the only expansion that can give you more than a net 7% of the map.

-Maps with more than 2 double ore expansions create "too much money"

I took the liberty of figuring out approximately how much money a player needs to sustain their basic production facilities for reference:

Image
Also, by "income", i meant "expenses"...sorry for the confusion.

As you can see, and as many of us already know, the magic number of harvesters has always been around 6. I'm not fan of these new "super eco" maps like Agenda, but in theory, maps with excess eco should provide more than enough money to allow players to tech.
However, the excess amount of money allows players to continuously stream out static defense structures without pause and without any repercussions. In a typical map, such as Behind the Veil, there is a finite amount of resources to be spent; you can't continuously pump out pillboxes before going broke. Rocky Ravine, on the other hand, has so much money on it that you'd be stupid not to place 10 pillboxes in the center, 5 up top, and 5 elsewhere. Even though I have more money to tech, I also need to spend more money on tech, because there are just way too many pillboxes on the map. "Super eco" maps may seem like it creates a better game, but to me, it just serves to highlight the negative aspects of this game.

I'm not trying to say that there should be a "cap" on how much money a map can have, but if you design a high-eco map, there should be some sort of map design (disregarding future balance patches) to mitigate the effectiveness of static defense spam with excess eco. Sirocco is one such map where it has an incredible amount of total ore, but the sheer size of the map discourages players from cramming static defense all in one location.

I think mo may have been on to something in giving players 4 starting ore mines rather than just 3. Taking our findings from the money distribution, with the knowledge from map size/features, I think by creating high value starting bases with relatively low value expansions may create games that would allow for tech and blob play. My next map will try to incorporate this feature in some way.

User avatar
Blackened
Posts: 347
Joined: Sat May 21, 2016 6:27 pm

Post by Blackened »

I'd like some verification the starting ore on Rocky Ravine
The bottom spawn's top ore patch has an extra line of ore on the bottom. Where the top is a circular 3.

User avatar
SoScared
Posts: 1116
Joined: Wed Apr 11, 2012 6:16 pm
Location: Oslo
Contact:

Post by SoScared »

Idea: List up all the played RAGL matches and sort them by maps. This way we'd get a neutral sample as to how these maps have been played in a competitive setting.

There's a break week coming (Week 15), would be interesting to see reviews on a preliminary sample.

User avatar
Blackened
Posts: 347
Joined: Sat May 21, 2016 6:27 pm

Post by Blackened »

SoScared wrote: Idea: List up all the played RAGL matches and sort them by maps. This way we'd get a neutral sample as to how these maps have been played in a competitive setting.

There's a break week coming (Week 15), would be interesting to see reviews on a preliminary sample.
like I did last year? :P

crlf
Posts: 28
Joined: Tue Jan 10, 2017 9:27 pm

Post by crlf »

I suspect there's a strong interaction between spawn-to-expansion, spawn-to-spawn and the eco/early tech question. The further the expansions and the more stable the spawn, the more viable it is to sit and tech a bit first; small spawn-to-spawn means you have a chance to reassert yourself, e.g. threaten the opponent's original base once you have your tech.

User avatar
SoScared
Posts: 1116
Joined: Wed Apr 11, 2012 6:16 pm
Location: Oslo
Contact:

Post by SoScared »

@Blackened: Sort of, only with links to each individual replays - sorted into map categories. Probably tedious work :)

Medium Tank
Posts: 12
Joined: Sat Apr 30, 2016 4:35 am
Location: Arizona

Post by Medium Tank »

@OMnom oops....

User avatar
Blackened
Posts: 347
Joined: Sat May 21, 2016 6:27 pm

Post by Blackened »

SoScared wrote: @Blackened: Sort of, only with links to each individual replays - sorted into map categories. Probably tedious work :)
The German in me has become slightly aroused.

Post Reply