Randomness in early game

Discussion in 'Planetary Annihilation General Discussion' started by matizpl, January 25, 2014.

  1. matizpl

    matizpl Well-Known Member

    Messages:
    229
    Likes Received:
    430
    Hello, I wanted to talk about very important issue that I believe has negative affect for early game play. What I mean is randomness of spawns and randomness of distribution of Metal extractors.

    Randomness in RTS game is usually bad mechanic. Imagine tournament when you play for 10000$. You and your opponent chose exactly the same early game build order. You spawn near your opponent but you sent your initial dox in wrong direction and you expand around your main base. Your opponent on the other hand sent his 1st dox in your direction and he sniped your initial 2 engineers. You are very behind and coming back from such situation is almost impossible if you play against someone really competent (let's say Godde for example)

    In Supreme Commander I had specific build order for each map that involved walking distances of commander and engineers. In PA you can only have -somewhat- good build but sometimes it will work perfectly, sometimes badly, depending on the metal extractor spots. Every game you have to basically invent the build order and this is not something good, because it takes away one of the most core determinant of skill. If my optimization of economy is better than yours, then I should have upper hand. Right now this is impossible to not only prepare for game but sometimes you just get flat out worse mex spots where you have less in your main base or your expansion is farther away. Imagine this - you have revolutionary build order that might win you a tournament, but you can't use it in finals because of how mex were distributed. Obviously making them on the fly is a skill as well but this way it takes away preparation for a tournament. You can't prepare for specific map because every game map is totally different.
    You might say that it promotes smart players and not the players that just play a lot. I don't think so, because then it will require even more grind in order to prepare 10 or 20 or 50 schemes of different builds for each possible type of spawn. Pre-prepared build is almost always stronger than any kind of build order made on the fly, even by smartest people out there.

    Most of RTS games have symmetrical maps and while I agree it's really nice to be different that majority, sometimes things become popular because they are simply good. I know PA tries to be different than other RTS games which is great, but symmetrical maps is one of those ideas that are actually very good and therefore it became popular.

    I'm really for spawning on the other side of the planet with symmetrical mex spawns because this way you can guarantee both parties have equal chances in competition. Planetary Annihilation actually has the potential to become the only RTS that is actually perfectly balanced because both players will have the same set of units. (ive been told that other factions will not have different units, is it true?)
    I think it's the best when you spawn on other side of planet, because that's when games become epic and you can fully utilize 3d sphere play. This will also remove problem of commander rushes, they still might be strong in order to secure certain location on map but it will not result in end of game. This is how it was in Supreme Commander when you had to make strategical choice whether you want to secure location with commander or you help with production which was very nice.
    If you want a rush map, then just create much smaller planet, then it will be more important to utilize early units.

    Another problem that random spawns create, is inability to know when you will face your opponent. If you wanted to play defensive in SupCom, you made 3 bots at certain time and you were protecting your engineers with them. Right now you have no idea when to build those protecting bots, because your opponent might find your engineers at 5 minute mark but it also is possible at 3 minute mark. This deletes possibility of creating optimal build orders and promotes build order poker.

    Another problem is that it's harder to construct your strategies that involve abusing certain mechanics of map, like water or unpassable terrain. If you knew the map, you could prepare a strategy that for example requires less units and you make wall to fortify certain position and not allow your opponent to pass through it and you use this to tech up faster or whatever. Right now you can't really do it because terrain is also random.

    Right now It's still not terrible because you scene and balance isnt very developed yet so build orders aren't very detailed yet so it's much more important to have knowledge of what's good and what's not. But in future after release of game, BO's will become much more important and then things that I described above will have big impact on the results of tournament. I really love PA and I think it has huge potential but this is critical in order to allow PA to become competitive/esport.

    I believe removing randomness would result not only in more fairness for players but also better quality of game for spectators, because players would be able to prepare better which would result in higher quality of games
    Last edited: September 26, 2014
    greysuit, beer4blood, liquius and 3 others like this.
  2. carlorizzante

    carlorizzante Post Master General

    Messages:
    1,371
    Likes Received:
    995
    So, what would be the best way to fix this aspect in PA?

    So far I have to admit that I like to be challenged by randomness. In fact I also do enjoy seeing how top players deal with randomness in the early phases of a match. But of course I do understand when you say...

    I do not see myself in that position, so for me hardly it will ever be an issue. Nevertheless, what's your solution?

    Mine would be that, assuming the System Editor will allow us that level of editing, there will be planets specifically designed for that kind of contest, with two or more Axis of Symmetry. While a more average match will have randomly generated environment, 'cos I believe most of us will like better that way.
    Pendaelose, corruptai and igncom1 like this.
  3. igncom1

    igncom1 Post Master General

    Messages:
    7,961
    Likes Received:
    3,132
    I only agree with respect to metal spot position in regards to possible spawn points.
  4. matizpl

    matizpl Well-Known Member

    Messages:
    229
    Likes Received:
    430
    I'd say ladder and tournaments should be symmetrical and custom games or maybe separated ladder could be non-symmetrical. I realize some players might like it and have fun with it and they should be able to play this, but as I said it's not good mechanic for competitive play.


    I'm sure you might also enjoy if there was a guy who came to chess tournaments and randomly took pawns from players, there's no debate about it. The thing is though that's simply not not fair for the players. If you play for fun, then it's not a big deal, if you want to play seriously, then ye, its bad.
  5. shootall

    shootall Active Member

    Messages:
    218
    Likes Received:
    184
    We have to have a planet editor. If we can build maps / planets and use for competitive play, this would not be an issue. Those who prefer random can still play random, and those who want symmetrical / fair / whatever maps can have theirs.

    Is Uber working on one or do we need to make one as a mod? How do i contribute?
    leighzer and carlorizzante like this.
  6. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    This. I believe you can have a distribution of metal points with a reasonable degree of fairness on otherwise random maps. Being able to choose your own spawn point makes a big difference too. Procedurally generated maps is the defining feature of PA for me. Randomness will never be eliminated completely, that is why you have best of 7 games etc.

    This is why I think random maps are better, because it promotes an on-the-fly application of understanding the core mechanics and situational awareness, rather than memorization of a specific build order for a specific map.
    SC2 is boring to watch because of the symmetry of it all. It gets away with it mostly because of the different factions. In fact that is the reason for different factions.. because it makes the game more interesting. In other words, adding asymmetry makes for a more interesting game to watch, rather than the other way around.

    And of course the players who practice most are the ones who will inevitably have the best understanding of the game.. if they are smart. ;) Current top players are used to symmetrical maps in other games and will probably be resistant to any change.
    Quitch, godde, Pendaelose and 2 others like this.
  7. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    Regular chess is asymmetrical as it is, white makes first move, kings on different colored squares. And yes it does give white an advantage statistically. Obviously they strive for symmetry in chess (I guess?), but I don't think chess is a very good comparison point to an RTS anyway.
  8. matizpl

    matizpl Well-Known Member

    Messages:
    229
    Likes Received:
    430

    I think Chess is very good comparison to RTS actually, it has some very similiar aspects. Yeah in chess you also have imbalance, but lets leave it.


    There is reason why all RTS games are balanced by developers. They want to give fair and equal chances to both players. What's the point of balancing game if the core mechanic of the game is imbalanced and it guarantees that you will never have the same chances.
    If you choose to spawn somewhere and I choose to spawn somewhere and you randomly find me first you have advantage because of luck.

    If you don't think that it's bad, I'll give you analogy that works exactly the same way as random spawns.

    At 4 minute mark, engine of the game rolls dice for each player. The result of dice roll is multiplied by number of doxes you are willing to sacrifice for this procedure. Final result is number of how many fabricators you will randomly receive.
    Would you as well say "From spectator view I like to see how players cope with such situation" "yeah it adds depth to the game" "It promotes players who are smart and are able to adapt to situation" "It's boring when all players have the same number of fabricators"?
    No. Just no.
    Last edited: January 25, 2014
    carlorizzante likes this.
  9. FXelix

    FXelix Active Member

    Messages:
    368
    Likes Received:
    116
    Same mex Spots and a spawn on the Other side of the Planet would be the fairest thing ever, but this would remove some kind of tactical play, fighting for mex spots when you are closer for example.
    So this would be fair and you could use your early game tactics and your build order at the beginning, but in my opinion a tactic is not static, you must be clever and must control every random match, this is tactical play.
    And a clever, competive player should/ must can handle every situation.
  10. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    The point is that properly designed random terrain does not break the balance any more than different factions, unless one insists on some mathematically strict definition of fairness.

    Both players have equal chance in scouting.. arguably scouting luck does not (and should not, which can be influenced by game design) decide any single game overpoweringly, and over many games it does not matter. I think it's more of a philosophical/game style difference of knowing where your opponent spawned over not knowing.

    Not random spawns.. player chosen spawns, meaning player skill comes into play. Of course currently you have the option to choose from a predetermined set of random spawn areas, but anyway. Ideally you could just spawn anywhere.. which would lead to short games if players decide to spawn on top of each other, but I think the current system is alright.

    Of course ideally we have all kinds of maps and tournaments can choose whatever they want, which will lead to same old same old.
    godde likes this.
  11. cola_colin

    cola_colin Moderator Alumni

    Messages:
    12,074
    Likes Received:
    16,221
    I think random maps can be okay if the generator is further improved.
    What's never okay is to not know your enermy locations at all. At least show the players the possible spawns of their opponent. Otherwise the early scouting becomes a horrible gambling game.
  12. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    You do know the possible spawns currently by looking at metal spot density.
    Pendaelose likes this.
  13. cola_colin

    cola_colin Moderator Alumni

    Messages:
    12,074
    Likes Received:
    16,221
    No you don't. You can do a guess based on that, but it is far from knowing.
  14. matizpl

    matizpl Well-Known Member

    Messages:
    229
    Likes Received:
    430
    I insist on mathematical fairness. I have played RTS games for 10 years competitively and 3 years of starcraft2 a couple of hours a day. Even slight tiny advantage early on is something extremely huge in situations when you are this competitive. If you play over 15000 games in a game every single ******* tiniest detail matters extremely much.
    Also scouting luck right now might decide a single game overpoweringly, the example is described in OP.
    And it will get even uglier in future if we get better at game.

    By random spawns I mean situation where you don't know where your opponent is and you can't react before you find someone. Semantics, the point is if I don't know where you are I can't set up a strategy and I gotta guess.
  15. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    Well the question is about possible spawns, so of course there is some degree of guesswork.
  16. trialq

    trialq Post Master General

    Messages:
    1,295
    Likes Received:
    917
    That's a good idea. It also means spawns deemed unfair could be reseeded.

    edit: To clarify, opponents could see each others spawns and decide if the map is fair.
  17. matizpl

    matizpl Well-Known Member

    Messages:
    229
    Likes Received:
    430
    Any kind of guesswork is bad for RTS game as I've described before.




    The other solution to this problem and to also build order poker I found in Command and Conquer Generals 2. It was the game that development is abandoned right now but I was able to test it in alpha stage.

    At the begining of game you can see every building your opponent is making. You stop seeing it after your opponent builds command center that prevents vision. This was reasonable idea although it's a bit counterintuitive for newer players so I don't know.



    While I think we are on the same side of discussion even with this I must disagree. I'll use SC2 example so dont lynch me but this is good analogy.
    When you play Protoss vs Zerg holding 3rd base is strictly connected to map design. If 3rd is too far away or it's too opened then toss has hard time holding it, if otherwise, then they hold it easily. If you introduce random terrain then you will have hard time solving stuff like this. Can I tech to t2 air at 4:50 if there is terrain like this? Or I have to wait till 5:10 because I have to make 2 turrets at these 2 locations. This blocks development of strategies because you have so many variables that constantly change every game.
  18. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    As long as you don't know where the enemy is, your priority should be to find out asap where the opponent is and your strategy is based on that. But I agree, scouting luck can be a problem.

    I respect your experience, obviously you know what you are talking about, much more so than I do. And Uber would do well to listen to experienced "serious" players. I'm just saying being extremely strict when it comes to fairness can lead to a more boring game.. in my humble opinion. :rolleyes:
    godde and Pendaelose like this.
  19. abubaba

    abubaba Well-Known Member

    Messages:
    501
    Likes Received:
    385
    Which is why you scout, to reduce the guesswork. ;) Late game scouting is as much guesswork as early game scouting, with equally drastic consequences, like finding nukes etc.

    That does sound like an interesting approach.
  20. cola_colin

    cola_colin Moderator Alumni

    Messages:
    12,074
    Likes Received:
    16,221
    I think that's mostly a result of the way SC2 does balance. It basically evolves around the map-structure. On a flat map Zerg would probably be totally imba.
    PA does not have that issue since everyone has the same unit types.
    The problem with t2 air you use an example can be solved by simply scouting first. Then you can decide based on the information you gathered. At that point it's a pure skill based decision. The only issue with this is that scouting in the early game is luck based right now. That would be solved by showing enemy spawning locations and maybe even which spawn the selected.
    Pendaelose likes this.

Share This Page