Since I recently joined the Alpha I've been pondering about how a (competitive) PA ladder match would work. With the heavy emphasis on procedurally generating content I'd assume Uber would want to have this tech drive most of their game, including the ladder. But given how important the map is to the way a RTS match plays out the parameters may need to be seriously fine tuned to make sure a map is fair. There's already been a discussion on the matter here; https://forums.uberent.com/threads/map-ballance.46262/ However, the 'pre-game' (where you pick your starting location) gave me an idea. I believe it is fine to have imbalanced maps as long as there's a fair system in place that determines where players start. I believe fleshing out the pre-game can have the players themselves balance out an outright unfair map. With the added benefit of adding an extra strategy layer to the game. Currently, it is my understanding that you're given a set of possible spawn locations which are entirely different from your opponent's. But what if you have a common pool for both players? It can't work with the players blindly picking their starting locations simultaneously, because they might want the same one. Doing it sequentially isn't much fairer, one player randomly has first pick without the other being able to do anything about it. So what if you could reserve a given number of starting locations, picked in turn, then immediately afterwards get to pick a subset of your opponent's reserved locations and deny those you'd least like him to start the game at? Let me give a very simple example of how it would work; Assume the outcome of a 1v1 game is entirely dependent on the amount of mass points at a player's starting location and nothing else. Now assume the map has 15 possible starting locations and all of them have a unique amount of mass points (MPs) available, they range from 20MPs to 6MPs. This is quite badly balanced, I hope you'll agree. Each player gets to pick 5 locations to reserve and can deny 3 locations of the opponent in the order of 1:2:2:2:2:1 and 1:2:2:1 respectively. In which case the picking might look something like this; Player 1 picks 20MPs Player 2 picks 19MPs and 18MPs Player 1 picks 17MPs and 16MPs Player 2 picks 15MPs and 14MPs Player 1 picks 13MPs and 12MPs Player 2 picks 11MPs Player 2 denies 20MPs Player 1 denies 19MPs and 18MPs Player 2 denies 17MPs and 16MPs Player 1 denies 15MPs Unfortunately Player 1 will always lose this scenario, as the best he can pick at the end is the starting location with 13 mass points, whereas Player 2 can pick a starting location with 14 mass points. However, picking optimally (at least I think the above is optimal) always results in the players ending up with similar, but not equal starting locations. A completely random system will not give such close results as often, though if you're player 1 you'd rather it be random so you'd at least have a chance to win It might not actually be most effective to have 15 possible starting locations, have players pick 5 in this particular way while denying 3. The numbers simply appealed to me as did the picking system. The pregame of a common pool of locations with reservations and denials are what really matters. An asymmetric map with unequal starting locations can be made competitive if you allow the players to do some strategic picking and denying actions in the pre-game. While also adding more strategic interplay from the get go. If you don't have a clue where your opponent can spawn your early game strategy only really depends on your own starting location. Whereas you can't mind game a random draw, you can mind game conscious decisions made by your opponent. This isn't exactly how RTS's tend to make their maps balanced, so its a bit out there.... But then again, it wouldn't be the first RTS convention that was thrown aside, so I don't know. What do you think?
I like the sound of the idea. It might be an alternative mode to a system that generates a fair map by using simple mirroring. I am not quite sure about your concept though. It sounds to me like it still has random elements in it. I am not quite sure how to get rid of them though. Maybe let players pick locations for their opponents. Maybe let them pick locations without knowing who will end up on them. But in the end all of those have serious drawbacks in my eyes.
I like your concept, but I think it has major flaws. With a finite number of known starting positions, you will more or less know, or have a very good idea of where your opponent has chosen to start, especially with denials or a small number of initial choices for the player. Part of what makes especially large planets interesting right now is not knowing where your opponents will be coming from. On the other hand, if there are a larger number of players or starting points, you would have to spend a long time in the setup phase of the game counting, considering which points you want to take, and which to deny rather than spending that time playing. I am not particularly fond of spending that much time just to start a game. Also consider: How would this work with 40 players or even team games? Do team members vote or have a leader? Do they first have to vote for this leader? How long can a player or team think/discuss before forced to make selections? What if they don't choose? My biggest issue is that this seems to add a lot of time and complexity for a minimal gain in balance/fairness. As an optional setting for a small number of players, however, this would probably be alright.
This seems incredibly complicated and prolongs the start of a game before you actually get to start playing.
What kind of randomness are we talking about Colin? For sure, the map would still be entirely procedural, and I suppose the starting location selector might have some randomness to it (I don't know how it works). But in the end it is not random where the player ends up, most of it is up to him, and some on what the opponent denies. Of course it also costs some time at the start of the game. The question is whether or not it would be worth it. If you're fine with having forced mirror maps or pre-made static maps for competitive play, then there is no reason to even consider this idea, its a waste of time. If you'd like to see reasonably diverse or asymmetrical maps from game to game, this could be a step in the right direction. Lets say we stick to my example and give each phase 20 seconds for a player to make up his mind. If all phases took the full amount for some reason you'd end up with 200 seconds of additional pre-game. A little over 3 minutes. Maybe a bit long, but will people actually use all their time? Maybe 15 seconds is enough? Its also something specifically for competitive/ladder play. I don't think a ladder is planned for 40 player games. If it is, then no, I don't really suggest using this. This would only be viable for smaller amounts of people. Also not for, say, 4 man free-for-alls. It simply gets too difficult beyond there being 2 sides with not an unworkable amount of spawning locations/players. I'm not sure why not knowing where your opponent is on a large map is better than having a limited amount of places he might be coming from. Not knowing where he is makes it hard to make meaningful strategic choices, whereas having some, even if it is limited information about your opponent, does allow for strategic choices from an early point in the game.
Why not just have a shared pool of starting locations. Both players choose a starting position. If they are different: carry on. If they chose the same one, remove that one from the pool and let them choose again. There is a really small chance they will continue to pick the same one as the best spots are removed from the pool.
What would you say are the benefits of having as little information on your opponent's position as possible are, ZaphodX?
Good point! It doesn't fit right with me to know where your opponent might be at the start. You should be scouting. That being said most RTS you do know exactly where you opponent starts, perhaps I'm just used to the way it works at the moment in PA but it just feels right.
I think it's been mentioned at some point that you will be able to set the possible start positions on your custom maps while you are building them. If/when a ladder is introduced, I would hope maps that have been designed to have balanced starts within them would be used.
Fully agree scouting is far more important because the fact you know where your opponent is means you can attack without them knowing where you are it's a double bit of defence
I think the best way would be to divide the map in half. Check that the metal and land is roughly even. The player can land anywhere on there side of the planet. There is a small strip of land in the middle that either player can't land on. This way the players have almost no idea where the other player spawns can spawn very near or very far away from the other player It also allows for random planets. I still don't see a problem with the current system though. I would be happy to see both.
