Randomness in early game

I really enjoy this discussion, points that you are bring are valid and I really respect your opinions.

I think it comes eventually to priorities that you have when playing/designing game.
I have fun from beating other players in good macro games that I can prove through skill that I'm better than someone. And obviously from winning tournaments. To achieve this I want as fair environment as possible where I can feel I've beaten someone not through luck. Also in tournaments when you face worse opponent and he gets lucky and wins it's one of worst feelings ever. And in tourneys if I get lucky and win vs better opponent through luck it doesnt feel so satysfying. (Money and satisfaction from winning game in tourney can make up for it though haha )
This is generally how most of competitive players feel and most tournaments are based on. The better player should win. Right now there is decline in SC2 scene because the state of Protoss in the game and guess what's the problem? Yeah it often feels like Protoss didn't play better but won because of imbalance or abuse of lack of scouting abilities by other races.

Some people have other definitions of fun and well designed RTS game, for sure. That's where arcade or custom games come into play where people play just other stuff, If you have fun having luck-based games, it's fine, but why can't you play it in custom games? Every decision to implement luck in competitive scene, decreases the possibility of succeeding PA in esport scene. I love TA/Supcom/PA and I really want it to be more successful than Starcraft in esport scene because these are better games than starcraft.

Yeah, I shouldn't take part in this discussion anyhow since I don't play competitively and thus my opinion = zero.

edit: competitively meaning in tournaments with prizes etc

Although I think for competitive play randomness should be decreased, at least at the start, to allow for at least fair starting positions. After that, randomness adds to the tactical versatility of the game, and makes things more interesting. My thought is that spawns should randomly generate each time the map is played, and have a basic 5-mex template that is applied under the spawn area. That way each player has access to the same amount of mex to start the game as the other. Spawn points would be generated no closer to other mex then a set range, so that a randomly generated spawn wouldn't spawn right next to a massive blob of mexes. After spawn generation, mex distribution and terrain distribution should still remain random, so that the game remains interesting.

Yeah it makes sense what you say. I just really feel like spawning on other side of planet is situation where 3d sphere map shines the most. I watched VODs of my game vs godde and every caster was happy that we spawned far away from each other. I was happy too. I guess it's matter of preference?

Yeah sc2 is mostly about maps but I think using map features like terrain or water and stuff is good for interesting games. If terrain is random it's harder to use it because you don't have setup plan and you more likely would ignore it.
What I meant with this t2 example is that if I can abuse certain terrain mechanic I could improve my strategy somehow. But this is a process of a lot of games to make it work well and if it keeps changing then it's even harder to introduce strategy that relies heavily on terrain.

Well, If I knew where my opponent spawns it is first step.
Second step must be to make sure both players have equal chances of metal expanding. I believe it's best to achieve this through symmetrical far away positions. If you have other suggestions then I'm up.

3rd step is stabilization and preparation, what I mean is that you are able to prepare for a tournament at certain map and certain setup of mexes and certain setup of terrain. If we don't have this we also have some sort of randomness but not gamebreaking. In other RTS you have mappool, usually 7 maps. You can veto 1 or two for whatever reason. You might think it's imbalanced for other faction, you might not like it because of your playstyle, you might not like it for design, you might be badly prepared for it, you might not like it for whatever reason it's your buissness. Different setups of mexes can be described as different maps, because you have to use different strategies for different setups of mex. If I don't like certain map I veto it, If I don't like certain setup of mexes or terrain or whatever, I'm not able to veto it and I'm forced to play on it because of random generator. (assuming it generates fair which leads us to step 2). This is a bit uncomfortable. It's not gamebreaking but I believe maps play huge role in setting up strategies (mostly setup of mex and terrain) and I really want to be able to prepare, otherwise practicing is a little bit problematic if we reach really highest levels of play. In supreme commander: Forged Alliance what differentiated absolutely top players from just good players was usually usage of map-specific reclaimables and counting walking distances of commander and engineers. I mean obviously not at the begining, but later when we got very good at game it is what mattered a lot. Right now this step is minor, but in future it will matter a lot.

I believe 1st and 2nd steps are absolutely critical for competitive play. 3rd step is I guess not a necessity but I think it would be really good for the scene as well. You guys seem to be against step 3, please consider the points that I made about it though.

I am not against step 3. I liked how in FA you had to have 16 different build orders for the 16 differens maps that were in the ladder pool.
I am just saying that it is not necessary if the map generator is good. So we agree on it I guess.

How do you picture the map generator that wouldn't be in conflict with step 3? I have hard time imagining how would it work

I think it's okay to measure how well a player can adapt to an unknown (but perfectly balanced) map instead of measuring how well a player can replicate a well known and well practiced build.

The entire premise of this game is that it is random and un-sequenced. It forces players to think on their feet instead of relying on preset moves. I'd also like to point out that this is how actual battle strategy and tactics work as well.

As for a defensive strategy, and an offensive strategy, and choosing between them, you will learn with experience how and when to utilize them. This is a game where SCOUTING is just as important as ATTACKING, if not more so.

It adds to gameplay, and for the competitive scene, it makes it even more interesting.

For me it's much more interesting to see how the player would adapt their strategy and tactics.

I think all we need is a 'competition' checkbox in the system editor. You would need to specify the number of spawns, and the system editor would build a symmetrical planet with evenly spaced spawns and normalized metal placement. Then you just need some sort of flag in the lobby that marks the system as a competition map and the number of players it's designed for. Sins of a Solar Empire uses randomness in its maps (planet types and quality change) but they recently added competitive versions of many of them that balance everything fairly.

what jodarklighter said.

The randomness of PA's maps makes PA a more challenging and fun game, but they are indeed not suitable for high-level competitive play. A planet editor or fixed positions of mexes/spawn zones would make the game better for competitive.

so real tournys play best out of 5 different planets everytime........ problem solved

The insistence on mathematical fairness and symmetry from some people just goes to show just how much other RTS have devolved into memory, muscle memory and click-fest games. this is not a skill that has anything to do with strategy/tactics, it is something that young people do better than older typically (this is a generalization) and can be honed to a fine art, but not by everyone leading to elitism and such by people who probably aren't as tactical and strategical as they think they are, and purely down to practiced build-orders and nano-second reflexes.... this is not strategy people!

The beauty of random spawns and even not knowing where your opponents start is it gives each player the chance to adapt and make plans IN the game and not before hand because they know the map or a map like it. Sometimes it wont work out in your favor, but sometimes it wont and you'll still win... and if you do its down to you and your brain, not your set in the lobby build-order and freakish fast mouse clicks.

Indeed.

Then the game would not become a WHO IS BETTER AT PA and instead morph into WHO IS BETTER AT THIS PARTICULAR SYSTEM.

Planetary Annihilation is about this hectic, crazy, random allocation of spawns that ensures a different experience EVERY TIME. This isn't baseball. You don't have four bases. This is more like soccer - you have t be everywhere at once, and the battlefield is constantly shifting.

You want to know what real strategy is? Go watch Ender's Game or read the book. That seat of your pants thinking and execution is true strategy. It's hard and it's real - to a point.

This is the closest game to a direct conflict between two minds that I've seen in many years. It does it brilliantly. Sure, mechanical skill will always be a factor, but it still comes down to who can outsmart who.

High level competitive play should reflect this. It shouldn't rely on things being fair - that restricts gameplay unnecessarily. Uber has been getting much better at making the spawns even - and different. I applaud this and hope it makes it to the competitive scene.

I really fail to see how mathematical fairness and symmetry relates to a pure memory, muscle memory and click-fest game.

Sure if everything is random you don't need any of those skills, but that has nothing to do with strategy either.
You need to give players a fair starting scenario. Random spawns where it can happen that your opponent randomly runs into your first expansions are really bad. They break the game in some cases. PA should not be about a dice role.

The thing with random maps to prevent the typical "I learn a few builds by heart to be better" is okay.
However I anticipate that it actually will not work. The best players basically will have just more to memorize. (like a lot of general patterns that occur often in the generated maps, etc.)

Not to mention that currently the differences between maps are so small that PA has only 1 single build order. The random maps are so "bad" currently that they reduce the strategic depth of the game. In games with fix maps you often get to see vastly different play on different maps. In PA the map is a non existent factor for how the game will play out in most cases.

high level competitive play is completely based on fair starting scenarios by definition. Nobody play competitive dice guessing.

Presumably the lack of depth in maps won't be true forever. And for sure, symmetrical maps most definitely leads to knowing the map and pre-gaming it, the challenge exists only when two players of equal or near equal apm face-off, and if one player knows the map intrinsically and the other doesn't (a rare scenario to be sure). then that player has the advantage. The clicks/actions per second thing is a reality no game can remove itself from without thought controls but it really has nothing to do with strategy and i really hate it when people hold it up and insist it's an important feature to protect and not a side-effect of our control system, and to a slightly lesser extent existing ui in strategy game.

Herb Brooks was the coach of the United States team who beat the *invincible* Soviets in the Semifinal of the 1980 Winter Olympics. Before that game, he told his team that the Soviets may beat them nine games in a row, but they only had to beat them ONCE to win. Were they a better team? A better trained, better equipped team? Not by any means. They were TRAINED to exploit the Soviets one weakness: themselves. And they did it. They won the one game that counted.

Look, I also respect your status in FA matiz. But if every tournament was based solely on who was the better player, we simply wouldn't have tournaments. We'd just stick with the ELO rankings and throw a couple hundred bucks at the top 10 players each month. Tournaments are about winning. You only have to win that ONE TIME.

If you played me more, I'd probably beat you multiple times - but you got me with my pants down ONCE, almost by sheer luck. You won when it COUNTED.

I understand this is the accepted norm for other games - but PA isn't just another RTS. You cannot just practice. You have to learn, adapt, and evolve. Rote memorization doesn't matter here, except in the early game.

Um. Colin.

The only way for me to explain this is to repeat it back to you:

MATHEMATICAL (strict, numerically based rules) fairness and symmetry (Mirror maps, essentially) leads to memorizing extensive build orders and muscle memory to increase your mechanical speed. Kinda like chess, the best example of this. Usually the player who wins is the one who can think farther ahead and memorize more moves and countermoves.

Pa is not about a dice roll. I mentioned above that Uber is getting much better at even spawns - and I am sure it will be getting even more precise and interesting as time goes on.

As for strategy, only the top tier players in both Starcraft 2 and Chess can effectively develop new strategies to surprise the enemy - because of the huge learning cliff to enter the *competitively competent* level.

I have yet to run into maps similar enough in layout and spawns that I can use the same strategy or build order on.

So you are basically saying that while SC2 devs can come up with different maps with markedly different strategies to win, PA is just one map that has the same blanket strategy?
Interesting. I'm not sure I agree, but interesting.

Again, it might be random, but it is even - or at least, that's the goal.

If we wanted to be completely fair we should just start on different planets with the same mex count.

Actually, that's not a bad idea......hmmm

obviously starting conditions should be as equal as possible, but true skill does not come from optimizing build orders. That is just calculus. Memorized build orders take the strategy out of strategy games in my opinion. I stopped playing starcraft for that very reason.

As for "unfair starting positions" being problematic, I somewhat disagree. Part of strategy is making decisions based on the current state of the game. Having a random starting location isn't a binary situation. It isn't a "Oh this spot is better than my opponents, or vice versa" situation; it is more of a spread of possibilities. Those possibilities should be recognized by a skilled strategist and said strategist should react accordingly.

If you are certain that you spawned with fewer metal spots than your opponent, then you should expand sooner than you normally would to keep up. Your decisions should reflect the situation, not an optimized build order.

Actually professional gambling is pretty big. They even air several card game tourneys on TV.


I'd just like to cite Frozen Synapse as an amazingly competitive and fair multiplayer game that uses procedurally generated asymetrical maps. Depending on the game type, you may or may not know where your opponent spawns. In the case of an unfair map, people just agree on starting over, though that's pretty rare.


I think a number of problems with the random maps can be solved by just upping the pre-game time before a match. 60 seconds to look at the map per planet/asteroid before the match starts should be more than enough to decide where both you and your opponent should be fighting over/focusing scouting efforts.