Map ballance

Doesn't mirroring also give you perfect knowledge about your opponents terrain?

There are easier ways to ensure all players have identical starting conditions without compromising the random aspects of terrain.

In the simplest way - design n terrain startup terra templates.
Calculate the spawn point.
Apply template to the spawn points - ensuring the terrain has the characteristics.
Apply random terrain function to the space in-between.

As far as I am aware there is not going to be shroud in the game so you could simply look at their side of the planet to get that knowledge anyway.

And then you can spawn nearer to an exceedingly good place for an expansion or other terrain based advantages that the other player is not close to, that sort of thing is fine for casual play but in a ranked match it skews the results and would invalidate any ranking data unless it factors in the terrain based advantages for the players.

For rankings it is important that a player wins because of how they played and not from advantages they had from things like favorable terrain or random events otherwise it makes the player's skill rating used to determine their rank extremely difficult to calculate. If a player gets an advantage over another player due to the map or anything other than their skill then it would be possible for a less skilled player to win a match against a more skilled player due to that advantage even though the more skilled player played better. That would mean that a win does not determine which player is better and instead you would have to have some AI go through the replay and evaluate how each player played based on their circumstances to be able to calculate if the players should increase or decrease in rank. It's probably not impossible to do, but I highly doubt it is something that would be within the scope of PA, especially given they want to release the game some time in the next few years (it would make an interesting research project though). It's far easier to just have mirrored matches and accept the lack of map diversity than it is to have completely random maps.


EDIT:
I should note however that on average completely random maps would be relatively balanced and would probably be a lot of fun when playing casually, it's just that it's not always going to be the case so it would not work well for something like ranked matches where the outcome of the match is much more important and needs to be consistently about player skill rather than luck.

Aye - rankings/tournaments is its own special case. The game should be designed with the hooks to enable it - but it shouldn't be a primary design consideration.

Didn't they mention launching orbitals to uncover the map in the last live-cast?

Regarding rankings & skill - comparative skill is a better measure - which can only happen over time and should be independent of the map or the proximity to resources (excluding any kind of speed play).
Should be.. it's a loose term
Skill should be measured by how well you deal with the cards you're dealt - anything else and you're basically replaying your preparation - which is a function of memory - not skill.

It's an arguable point - if 2 players play 10 games on the same map - would a "weaker" player with better access to resources win against a "better" player every time? A good player would change their strategy - great players would be able to adapt quicker.

Thinking more generally and off the top of my head - should the end result of a "rated" game be static or dynamic?
We're all comfy with static results - a score, a change in rank/rating - I'm just wondering if some kind of sliding window function. Where the result of the last n games isn't fixed but more "flexible" - where each game played provides a number of attributes that can be used to recalculate or adjust the scores of other games played to give a more representative score.

Maybe PA will allow us to explore these types of ideas in more detail. Would be great if we could explore the data associated with games and see if we can come up with some alternative scoring/ranking approaches that explores "skill" in other ways.

In my opinion, it's far more difficult to balance races than to balance starting positions.
Still, Starcraft with 3 completely different races is considered as one of the most balanced game ever made.
I'm sure Uber can reach a good balance without the need of symmetrical maps.

^This.

It is a fallacy to say that a higher skilled player should always win against a lower skilled player.

If you're talking about the Fog of War, it should be present. In the future. Was mentionned in a previous LiveStream that the functionnality is not implemented yet.

Before continuing on the topic, i must mention that i don't remember seeing any mention about Mass Points.
Mass will be used for sure, but is it certain that we will have Mass Points ? Never saw any in the gameplay that was shown to us until now. All extractors seemed to be planted directly on the ground.
(goal of game is to be about huge gigantic battles, and a lot of ressources are needed to achieve that goal. Could make sense.)

As for the map balancing, i recall hearing clues about it in a LiveStream.
Something about the spawning location beeing choosed by players just before the start of the game.

Add the 2 previous ideas and all Map Balancing stuff is some old story.
No Mass Point so no need to worry about spawning them evenly for everybody.
No pre-defined spawn points so you are the only one responsible for your starting position, and the only one to blame in case of a loss due to a poor choice.

I seem to recall it being mentioned in one of the recent livestreams, that they have not yet implemented discrete resource points. At this pre-alpha stage you can just dump the building anywhere for full resource generation.

The implication being that discrete resource points are intended, just not yet implemented.

I have also assumed up to this point that those points are going to be randomly assigned to the planetary body, and I vaguely recall neutrino mentioning they were still discussing the best ways to handle spawns, including giving the option of self-selected spawn points, and mutually exclusive ranges for players to select a spawn point from. Again, though, at this stage I doubt it is a core focus as much as getting the engine running and basic stuff working; I'm sure people will be adding stuff to the lobby functionality all the way through alpha, and I expect we'll see several iterations of spawn selection options until we end up with one or more that people agree work well :mrgreen:

How many maps for ranked play in Starcraft aren't mirrored? I don't remember there being any, perhaps that's why it feels so balanced?

Shroud is the black that hides all map information in most RTS games, fog of war is the area of the map that is usually greyed out a bit where you can see the terrain but you can't see what is happening on it. I am aware PA has fog of war, but if it doesn't have shroud in theory you could just rotate the map and see what the terrain is for your opponent since there is no shroud to hide it, but you won't be able to see what your opponent is doing there because there is still fog of war.

What Neutrino was saying is that you get to pick from a small set of spawn points, and since you can't pick the same spawn points, one player could have access to a much better spawn point than another, which means the game could be to blame for a loss, not the player.

The developers want mass points because it keeps the player from building a big base. Mass points force players to expand and build forward bases to protect strategic ones. And yes, experienced and advanced players wouldn't make "a big base" and they would make forward bases. But not all players are advanced.

Mass points also limit one type of resource, adding another strategy to resource management. This is why they want to implement them. I would propose a different solution. Well, maybe not propose, but put out another idea. Cut the map into many smaller sections, not too small. And assign a max number of mass points that can be built in each section. Randomly maybe. The sector the player lands in would have a set amount of mass points to be built in it. So the fight for territory still rages like with mass points while allowing for free placement of mass extractors. Some territories may have none, there may be one that is heavily fought for because it has a mass point value of 5.

The sectors would uniformly divide a planet up like in a grid. Then values would be randomly assigned. The player would have an area they could drop into, they choose their spot within that area and drop. The sectors wouldn't have values to them yet. The two players drop, the sector they start in automatically gets a value of 4 for each. Then the sectors around them get randomized uniform values for both sides, then the sectors between them are randomized. Sectors closest to the middle of between the opponents would have a higher chance of higher numbers, equalizing the map. Sectors closer to the player would have a lower chance of higher numbers, preventing potential imbalance.

This is just an idea. Tell me what you guys think of it!

Of course, the balancing of terrain could work like this in a way. When first spawning, the terrain is generated where the player can land. Both sides have the same features of land in the area in which they drop and in some of the sectors surrounding the landing zone. After so far away from the start point I think that the terrain would not unbalance the game too much. What that distance is I don't know.

I really believe this would be an idea to pitch to development. Unless you guys can find kinks in it? Tell me what you think.

It is not about being symmetrical, it is about having a very strict map formula.
If you created an open map in Starcraft the balance between the 3 races would completely break down as Zerglings would just roam the map keeping the other races in their bases while the Zerg player expands unchallenged.
So when PA only have 1 unit pool the maps can have much more asymmetry and still maintain balance because there is no faction balance that breaks apart.

And a large part of that formula is being symmetrical, again, how many maps for ranked play aren't symmetrical? It's the major constant among all the maps used for ladder matches, a lot of the other parts of the formula aren't present in at least one of the maps but as far as I'm aware they are ALL symmetrical.

Actually without the factional differences map balance becomes even more critical. With factions there's usually a unit composition that is more effective than your opponents, which means using those particular units to your advantage you can balance out an imbalance created by the the map. With the same unit pool you can't use favorable factional units to help balance things out, they have the same units you do but you are still stuck with a disadvantage from the map.

Very good point. The map becomes ever more important which is why having a random map without constraints and strict formula will totally unbalance everything.

So is my last post a generally good idea for a map balance formula? It seems to make sense.

That is Starcraft not PA. Starcraft have 3 races. That is 6 different race matchups. If a 1v1 map is asymmetrical there are 9 different asymmetrical matchups. Good luck trying to keep that map balanced across all 9 matchups.
If you want to prepare builds for all matchups with both starting positions with all 3 races you need to practice 18 builds. With just 1 race you still need to practice 6.
In PA there would only be 2 asymmetrical starting conditions.

Now you are looking too much at Starcraft. Some maps and matchups are like that. You have one unit combo that is dominating that matchup and/or that map.
When you only have 1 matchup you can push unit diversity and balance so that there isn't just 1 unit combo that is dominating. For example, do you focus on land, air, sea or do you head for the stars?
Do you make bots, vehicles, hovers, ships, amphibious units or flying units?
The planet terrain dictates how and where you can use them.
Do you got a hilly start? Does is the enemy on flat terrain?
How do you use that to your advantage?
When does an asymmetry become an imbalance?
If there are enough tools for the players to chose from there might not be imbalanced starting conditions it's just a matter of how you take advantage of the terrain with your tools.
Anyway. The important thing with 1 race is that both players can have the same tools and still apply different tools in the same game if you balance it carefully.
Symmetrical maps is just a cowardly way out of having a balanced game with diverse unit sets and strategies.

Literally by definition every time:

http://www.thefreedictionary.com/balance


However usually in games we need to use degrees of balance and some things being asymmetric can only cause a very small amount of imbalance. For instance if a tree near a player's base is rotated such that it could obscure a unit and the matching tree for the other player on the map is not then that is a noticeable imbalance, but it is unlikely to have a large impact on the game so we just accept that as being relatively balanced, much the same as the different factions in Starcraft they aren't really balanced (hence all the patches altering units to improve the balance), but they are close enough that we mostly just accept it.

IMO for ranked games we need to remove as much imbalance as possible so that matches are determined almost entirely by player skill, symmetrical maps do that, asymmetrical maps do not. I know it's arguable that taking advantage of or overcoming an advantage or disadvantage in the map is part of a player's skill but with current technology we don't have a way to reliably measure that so we have to remove that variable when determining a player's skill which requires removing imbalances in the map. But like I've said before for casual play asymmetrical maps are fine, the imbalances would probably be a bit of fun to overcome, but for a competitive environment it's just not suitable.

Rock-paper-scissors is about as assymetrical as it gets.

Are you saying it's not balanced?

I think an assymetric system can be balanced.

The issue with rock-paper-scissors is that you either have to play it straight - and tell players "THIS ONE is the ROCK, THIS ONE is the SCISSORS, and THIS ONE is the PAPER!" or you have to try and disguise it as something else.

Anyways, RPS almost never works in RTSes. It's kinda like having no dice in a Roleplay, nobody ever does it because it's a stupid idea.

I would much rather maps be asymetrically designed to be fair as opposed to by-the-book balanced. That's the point of having separate Naval, Air, Bot and Vehicle factories; if you don't have enough mass, you can always take energy-consuming Bots instead. No, they are not as good for large-scale fighting, but they get the job done if you end up being the subversive player. Or you can gamble and take Air units, possibly murdering your enemy's eco in trade for having no resistance to ground attacks.

rock paper scissors (the game) is symmetrical, you can always go rock paper or scissors.

If you are interested how balance is explained in RTS and multiplayer games:http://www.sirlin.net/articles/balancing-multiplayer-games-part-1-definitions.html

Also a good read:http://www.sirlin.net/articles/balancing-multiplayer-games-part-2-viable-options.html

Designing a game so that the more skilled player always win is detrimental to strategy. Strategy is dealing with the unknown. Strategy is weighing different uncertain outcomes against each other.
The more skilled player should not always win. He can be surprised, ambushed or face a strategy he has never meet before.
Strategy is getting into the opponents head, taking risks, hoping the enemy won't see it coming in time to counter it.
If you design a game where the more skilled player always win you haven't designed a strategy game.

Edit:Something about local imbalances vs global imbalances:http://www.sirlin.net/blog/2012/7/18/a-discussion-of-balance.html

those are good reads, i've seen them before but i'll give them a reread

Read the section on fairness, that is what I'm talking about. You get a starting location that you are locked into at the start of the game and if it has a disadvantage (and on all asymmetrical maps there will be a disadvantage to one player, though of greatly varying degrees) then that is unfair and unbalanced, hence the need for symmetrical maps for ranked games.

And again, just to be clear, all this symmetrical map stuff only really matters for random maps, an authored map can be made to be reasonably (though never perfectly) balanced even when it is asymmetrical, so I don't have any issues with asymmetrical authored maps assuming the author took a lot of time and effort to really make sure the map is as balanced as possible.