1. If you're looking for help-related things (for example, the key rebinding tutorial), please check the FAQ and Q&A forum! A lot of the stickies from this forum have been moved there to clean up space.
    Dismiss Notice

Why randomly dropping "superior" brains is an inferior idea.

Discussion in 'Starbound Discussion' started by gnilbert, Dec 15, 2013.

  1. IcyMist27

    IcyMist27 Poptop Tamer

    This game is Grinding 101. Everything... even the brain zaps are meant to be a chore don't you know? It pads the game length and makes it a 300 hour experience. Extremely riveting stuff.
     
    RynCage likes this.
  2. Mystify

    Mystify Void-Bound Voyager

    Because chess is SOOOO boring without randomness.
    The proper use of randomness in games is something I have paid a lot of attention to, done math on, etc. My conclusion is this:
    randomness should produce unpredictable scenarios and obstacles, but not determine success.

    So, for something like ores,randomness makes them unpredictable and it makes it more exciting to explore and find them. However, they are such a frequent event it doesn't have a significant impact on how much you will aquire, and other factors like different mining strategies are more relevant. Meat and pixels have the same deal going on- they are fairly likely, and it has plenty of chances to happen, so it averages out very nicely. Unpredictable moment to moment, but the overall trend is fairly steady.
    Randomness in the terrain generation creates interesting scenarios to deal with, but does not determine your success in dealing with them.

    However, for something like this, you can't rely on long-term trends to average things out. You only need 1, and its rare, so you instead end up with HUGE variances between players. One guy has it handed to them, the other has to work at it for hours. Why should the second guy have that much of a harder time than the first? Because he is unlucky? The game should reward people who utilize better strategies, not those who are lucky.
     
    HellFireSoul and Beret like this.
  3. Reon

    Reon Phantasmal Quasar

    That is a fine system actually, which I like. The only thing I dislike is the game having to increment two number (in this case, creatures kill with a specific weapon and brains dropped) and run a loop every cycle to check that these two numbers aren't abnormal and goto a line that will give you a free drop/deny you a drop if they are.

    In one instance, this isn't too bad. But if you plan to make a lot of this outlier detection, you're going to bog down your game.

    Still, if one could find an efficient, self-checking piece of code to do this on hundreds of variables (which I can't comprehend of), it'd be great.
     
  4. Mystify

    Mystify Void-Bound Voyager

    The rest need less than 25 kills
     
  5. Reon

    Reon Phantasmal Quasar

    Chess is remarkable random. You are the randomness. Have you ever played chess against your self? It's still very random.

    Now if you're one of the books which gives you the play-by-play of famous matches and just duplicating it... That would be incredibly boring.
     
    JanusZeal likes this.
  6. Reon

    Reon Phantasmal Quasar

    Then the first line should read 136,402. That's the number that needs between 25 and 50.
     
  7. Mystify

    Mystify Void-Bound Voyager

    Player actions are NOT randomness. There is a fundamental difference between "I am facing an opponent and playing against them" and "I rolled a 1, so my queen fails to take your pawn".
    A horse race is not random. Sure, you do not know who will win ahead of time, and hence people can bet on it, but the winner is determined by who runs the fastest, who took better routes through the course, etc. Sure, there are also random events- a horse trips and breaks a leg, for instance - but for the most part it is the outcome of competition.
     
  8. Reon

    Reon Phantasmal Quasar

    To say that there is no randomness in a human brain is.... Idealistic. To say there's none in the body is insane (and able to be proved incorrect), as is saying your physiological condition doesn't effect the function of your brain.
     
  9. Mystify

    Mystify Void-Bound Voyager

    But that is entirely outside the scope of the game, and really starts getting into the philosophy of what thought is.
    I move my piece. It goes exactly where I want it, and it will behave in a perfectly predictable way. My opponent does the same. The ONLY determining factor in the victory is who moved their pieces in a better manner. This is an entirely, fundamentally, different thing than having randomness in the game. It is the difference between "Who is better capable of exploring the depth of this game in order to utilize it to reach a goal" and "who rolled a higher number on the die". It is related to "Who best possesses the manual skills to perform this task better", and the two are often combined. They both are about how well you play the game, which can be pitted against how well someone else played the game, or against a more static obstacle like in a single player game. Many platformers have no significant randomness, it is purely an expression of your skill as a player to get through it.
     
    HellFireSoul likes this.
  10. Reon

    Reon Phantasmal Quasar

    Perhaps I have moved off the subject, my apologies for that. Let me state where I agree with you, as I think I'm coming off too extreme:

    Random drops can be annoying and are replaceable by more complex systems.
    Some kind of detection system for those far outside the norm would be good (perhaps a loop that activates when you fire the brain extractor and auto-drops a brain if it's been X kills since last brain).

    But I do not think introducing a 7% drop rate is a critical, game-breaking mistake nor do I believe that random drops have absolutely no place in games. As I said before, if you want to normalize experience without moving away from random drops or making it take less average time/effort: raise the drop chance higher (say 20%) and add a second random drop item needed for the boss at the same drop rate. The higher drop rate will make the curve fall off faster and the second item will assure it still takes some time.
     
  11. gnilbert

    gnilbert Space Spelunker

    I didn't present a partition, I presented overlapping subsets. Of the original 1million, around 236K needed 25 or more kills. Of those 236K, around 56K actually needed 50 or more kills, etc. In other words, I was using the cmf instead of the pmf.

    Gnilbert
     
  12. Reon

    Reon Phantasmal Quasar

    Still, I show only 175222.8603 need 25 kills or more. Hence me saying one of us made a mistake. Perhaps it was me, but I've checked it a few times and can't find my error.
     
  13. Sariaz

    Sariaz Subatomic Cosmonaut

    I think the drop rates are find but I feel like superior brains should drop from superior monsters such as humanoids or mini bosses. I also feel like there should be a use for inferior brains, possibly replace the steel cost in making npc spawners or some such use.
     
  14. Mystify

    Mystify Void-Bound Voyager

    Things don't have to be game-breaking to be worth doing better. A game comes together from many factors, some of which people consciously notice and many they don't, all of which contributes to the overall experience. Singular random drops have many downsides.
     
  15. Sturmkrahe

    Sturmkrahe Aquatic Astronaut

    Chucklefish, please hire this person. :chucklefish:
     
  16. gnilbert

    gnilbert Space Spelunker

    No worries - it certainly wouldn't be the first time I've made a public math flub. Especially with probability, since it's always a bit hinky.

    I used the probability of success (getting a brain) as .8 * .07 = .056. (P(dropping a brain) * P(dropped brain is superior))
    I used the 1 - CDF(x) of the geometric distribution to represent "probability that you'll need more than x kills before a superior brain drops:" (1-p)^x
    Then I multiplied the result by 1million to represent "expected number who needed more than x kills before a superior brain drops."

    And excel did the math... so I blame excel for that stuff! ;)

    Gnilbert
     
    Reon likes this.
  17. bluesymmetry

    bluesymmetry Void-Bound Voyager

    I'm pretty sure your work is correct though, the probability of P(X>25) of X~Geo(0.056) is (1-p)^25=0.2367553804. Multiply that with 1,000,000 and you'll get your figures. We can also find that P(X<=25)=1-P(X>25)=0.7632446196. The numbers are correct. Unless the p was different. Then just apply the formula. :)
     
  18. Reon

    Reon Phantasmal Quasar

    Yep, it is. I wasn't taking into account the chance of NO brain drop. So I was running P at .07. So... IGNORE ME!
     
    bluesymmetry likes this.
  19. Disig

    Disig Pangalactic Porcupine

    I have issues with certain implementations of RNG. For instance, the brain zapper does not work well for me. This is a mechanics issue, as I'll be shooting at things and unless I hit them at just the right distance in just the right spot they wont get hit. This is a bit frustrating but I am sure it will eventually be fixed as the game polishes up more in beta. But I feel like, in addition to the drop chance, they should have specific mobs that drop it no matter what, like the NPC's or bosses. Make the NPC's more of a challenge to kill (especially with the brain gun) and give us a meaningful reward for taking on the challenge. Other people can take the RNG route and zap level one monsters.
     
  20. MyLittleBurger

    MyLittleBurger Scruffy Nerf-Herder

    The funny thing is, it only took me about a minute or two, to find a superior brain from the 4 monster I killed with the zapper.
     
    blueokapi12 likes this.

Share This Page