Current work update - 11/5/14

If that was mirror map for balance (and the only spawn), it would actually be sort of... creative. You'd sure have to adapt to the situation and think quickly, instead of following some routine plan.

That aside, current random maps are just too random, but there is a lot of potential in it. There are many different approaches that could be used to make them more balanced. For starters, make sure that there is roughly equal access to metal for each spawn. Currently metal seems to be totally random. Instead, one could device an algorithm that places metal equally based on spawns. Or just randomize it 100000 times, evaluate results and pick the most fair one.

should also work on random crashes/pathing, tried to stream today only to have my game crash on more then 1/2 the maps

not just my side, but server side, both players/maps would just crash and die

Yes, that is exactly what I want.

What you call "play blindly" I call "may the player with the best ability to adapt win".

The problem with fixed 1v1 maps is that you already know the exact strategy going into it. There's no need to adapt. There's really no need to even learn or experiment -- just go watch some replays and do everything that person does in the first 3 minutes and you'll already be beating a lot of people (the people who did not watch those replays and do not have the exact right order of actions for the first 3 minutes on that map).

This is where my mockery of 1v1 players begins, in fact. A lot of them are fish out of water if you put them in any situation where they don't have the map memorized. They'll play like fools on a map they're not familiar with because half of their skill was simply memorizing what they saw other players doing and copying that. They are weak in the "adaptability" category because that's a skill that a fixed set of maps does not have much room for.


I say the real measure of an RTS player's metal is how they stand up to each other on a map that neither of them has ever seen before -- unfortunately this is impossible in most games..... but in PA, especially with symmetrical maps and random seeding, we could make this happen.

Chess only has one user-made map and that seems to have quite a few viable strats.

I think you're missing a large piece of what makes PA attractive.

Make these maps multi-planet. We both have the same options, we can pretty quickly find out if we're on the same planet or not, but the direction the game takes can vary wildly based on how we spawned, what the intervening terrain is like and whether or not that intervening terrain includes "space".

MAYBE we should even consider a system of letting players drop anywhere they want within their hemispheres. Your landing zone is all northern hemispheres. My landing zone is all southern hemispheres. NOW we're cooking with gas.


I really don't want PA to take the same path as SupCom 1v1.

SupCom was an amazing game with support for huge maps and lots of options but the 1v1 game came down to mostly tiny maps and it was rare to get a match that ended up in us making the amazing units sold to us on the box. I bought a game that showed amazing experimentals going at it and I ended up winning or losing most fights in a teeny tiny box, often enough with the battle decided in T1 (I especially hated Winter Duel. That map is the epitome of what was wrong with SupCom 1v1. Not that the battle, in itself, was bad, but that it was a far cry from everything SupCom was supposed to be.)

I predict you'll get people develop generalized build orders and use them every single time
Actually that is not a prediction, but exactly what happens so far in PA.

No human is capable of coming up with a fully refined strategy for a given planet within a few seconds.
People may be able to do a few hundred apm if they practice, but what you are asking them to do is far far far beyond that.

and those few capable of that would be the real planetary annihilation gods

there are no such people, that's the point. Not to mention that with the current random generator you actually are very likely to get maps that simply do not over any special strategy. Maps need to be designed for that.

EDIT:
Not to mention that even IF a few people capable of that exist and the maps allow for interesting strategies these few people would be able to play even more interesting games if they had a chance to actually think a bit more about how to play the maps and experiment. So even then you get more interesting games if the players can play around with their ideas a bit.

See my edit for the theoretic situation of having a perfect map generator.
But I doubt that will ever happen as well.

it doesn't have to be perfect but to simply allow for a well number of strategies ... so we don't hear doxerg OP all the time ...

i don´t see how those "ideas" should be any less usefull/intresting on the same/similar terraincombination on other maps/planets than on the one they may have trained

actually no not really. at really pro level almost every single opening is exactly the same on both sides.

and the N°1 players in the world are the ones who deviate the least from the wining meta.

and in particular there is more meta for speed chess. which really deserves the comparison much more, if not to go all the way to speed draughts.

actually it's below that. precicely speaking APM-wise. However it's much more refined than having the highest APM to counter that bot or boom or tank rush that's coming straight for your base because he doesn’t have to scout and neither do you. (all of this is single to dual unit micro, incredibly tedious, requires some of the highest APM, higher still than what you have in starcraft II, to win.)

I know from FA gameplay (as well as you should) that gameplay is more interesting when (albeit using a base building block of a solid build to assure yourself a solid start) adapting to your environment and showing what tricks you have up your sleeve.

If you're scared of gameplay getting to repetitive and unilaterally identical, then I'm sorry to surprise you with this but, we're already there, it's already the case.

players MUST apply a rush using memorised spawn locations to win in ladder. it has prooven the best way. so there you have it. your unilaterally identical gameplay.

proven in the current setup.

therefore you cannot theorise that changing things will lead to the same result. I mean you can, but next to that the theory that your result should change looks a hell of a lot more potent.

I want APM to go down. I said it two years ago I'll say it now. this is bullshit. this game will have no sucess whatsoever competing with starcraft. And why should it? those here wanted something different than starcraft, otherwise they wouldn't be here.

starcraft is not about apm, neither is pa. a certain apm requirement is just the result of being played in real time. That implies that a player doing stuff fast has an advantage over a player doing stuff slow.

If you completely randomize spawns instead you get random wins and losses mixed with people who do some random scout pattern, whoever scouts first gets a massive advantage. Can't see your point.

SupCom:FA is most fun when played on a set of 16 or so ranked maps where you have a specialized build order for the first 5 minutes (or multiple ones, one for air, one for land, etc), maybe a little more. After that so many possibilities for gameplay start to develop that having a planned out build order for it becomes mostly impossible.

It's btw the same for starrcaft and any other good rts. People have build orders and plans of how to play and then those plans crash with each other and it turns out that the plans rarely stay useful for more than 5-10 min ingame. After that whoever is better at adapting the plan to the opponent wins. Even when playing the same matchup with the same map and the same players a thousand times in a row.

how many years did it take for chess to reach that state? Not to mention I've seen people argue it has not reached that state, at least not yet.

1 it has 2 the time it took is irrelevant. just as irrelevant as saying PA is not as good as faf now.
what you care for is the "final" (unless the world has an end the is no "final") result.

other than that I was not advocating for random spawn points but for randomly generated maps. sure it gets diceroll-y but 1 it's been getting better and better since the start of alpha and with linear progression of this embetterment you could hope for completely balanced procedurally generated maps 2 I think symmetric maps are an answer to what you said in and of themselves.

actually starcraft IS about APM and so is PA because guess what? 100% of people wanna win. I can relate.

but Supcom prooved to me that APM wasn't an absolute necessity for winning and even showed me that high apm, wildly over-focusing on a silly squad of units would be my downfall in this game. I found that this way of approaching things was perhaps "better" at least more enlightened, definitely more enriching and made way for thought, strategy, tactics, improv.

you can't blame me if I want to try and dig this up out of PA, I've been baited really hard with FA, don't give me the switch.

that's rather what I said. I agree.

moreover that is not synonym with what is going on, with no mass upgrades, with no adjacency bonus and the likes revolving around eco and it's complexities, all of what you explained above goes out the window, we only focus on Speed. The speed of the reaction of our dox, the specific damage they inflict. The speed at which we can kill the enemy com. And by jove the game throws all the tools in the world into our hands for us to finish it in record time (If you were to make a sex joke out of it it wouldn't be very glorifying). The game is over before reaching any semblance of mass of units god forbid orbital and what is this T2 you speak of? never heard.

you see? not very comparable.

and it IS APM. 100%. Nothing else. What else do you have time for? I mean christ, you can't have all your units on one screen like the strat view. considering the nature of spheres you can't have all of them without five screens.

so basically if you're doing this right you're hopping back and forth between five camera positions (that you constantly have to update except for the base, because, well the battlefield is moving) whereas in starcraft you could get by with two! one for your base so you could queue up units and one for the battlefield.

OH! I forgot! you don't need to queue up units in PA. there’s a mod to make em build in loop by default. Damn...

why ever look back at your base then? WOOOOOO 110% MICROOO

I've covered some of this before.

I've also dabbed some more on why PA is micro or APM intensive In many-a-thread

I like random because I am better at adapting to new things then I am in finding the perfect build for a specific map.
Its a skill worth just as much as any other

Except even this sometimes fails. Like the image elodea posted on page 6 -- imagine you and your opponent both started in identical spots just like that, with 1 metal spot and terrain so rough that there's probably only room for 1 factory before you need to walk a ways.

"Generalized build orders" only work for generally true conditions. Random maps have no generally true conditions therefore anyone trying to use a generalized build order is likely to lose more than someone who takes the time to examine the map (you have a couple of minutes to decide to land) and come up with a rough plan on the fly. What factory are you going to build first? Which way will you walk after that, to get out of that terrain? How much of a chance are you willing to take?

At any rate, you certainly aren't improving things by making the maps non-random. Now instead of a generalized build order that works most of the time and has to be adapted some of the time, you end up with a very specific build order that may severely punish any deviation.


Usually the point when I'm done with an RTS is once the strategies have been whittled down to 1 viable plan per map, and everyone does that. I know some people enjoy it more at this point because then it's all about the execution, but I like it when I have to think of strategy on the fly. The fight to determine which of us can micro our Dox better is not remotely interesting.

From past experiences what actually happened was: "damn this planet is no fun, wanna re?" "yes"

To me you basically want to play on a map where nobody knows what they are doing, so nobody can get owned hard for not knowing what they are doing while their opponent does.

It is never just about the execution. Execution is a very important factor and if your execution isn't even half as good you will default lose, but if both players have the same -very high- level of execution you actually get very interesting high level play.

What you call "no fun" actually looks very interesting to me. It's a starting condition you don't often see. The only reason to abort that map is that you know your opponent probably got a way more awesome start than you. If they had the same start, then game on and may the best adapter (aka, most knowledgeable, experienced player) win.

That's a very minor consideration. Even a casual new player will learn the maps within a month. A dedicated new player might have them down within a couple of days. So the odds of trouncing someone because you know the map and they don't is hardly worth considering.

I want exactly what I've said:
Random terrain that offers a different experience every game. Neither of us has memorized a "Winter Duel" build order or knows the best ways to advance and defend on this terrain because neither of us has ever seen this terrain before. The winner is not just the person with the best execution but also the best seat-of-the-pants strategy.



Incidentally, while I can't think of an RTS that does random maps, it's very common in strategy games. Civilization and literally every space-based strategy game has random maps. Heck, even some strategy board games, like Settlers of Catan, are based on a random map arrangement. It does sometimes mean that within 1 match, it's not always fair, but in the long run the player with the most wins is the one with the best understanding of the game and the best ability to adapt.