Ironbound

Announcement

Collapse
No announcement yet.

Thinking out loud

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Thinking out loud

    I do not know, maybe I have a paranoia, but this game is already annoying me.
    What's wrong with trigger system i can't understand when Assassins and Witches Mastery procs 12-15 times and for example my 45% Crit chance only procs 1-3 times and no any reflect when i'm last standed and this is not only solitary case, I observe this phenomenon quite often.
    For example if we compare mastery procs and Faith+1 weapon which says it has a 50% chance to be able use again is absolutely do not match, i played with this weapon more than 70 game and did not saw 50% chance at all... a coincidence? I don't think so!
    Also why do you make Witches so strong, gave them imba Orbs which causes enormous dps and you just can't do anything, gave them weapon attack like other melee classes, AP Burn which is unnecessary in this game indeed and ruins whole game balance and yeah Fukcused Mind...
    Muzrub needs to be nerfed -> limit attacks IMHO.
    Crusaders are weakest class in this game, i was convinced after playing for him many times despite the fact that sometimes they have good porcs of crits and they very depending on crit chance because they have only 1 attack in many cases so this must be also taken into consideration.
    Also when you are at 3200+ Rating Bots with 300HP not always easy to defeat (depends on tokens you get and not for every class it's easy) and when you loosing against them you loosing too much rating points because they max lvl is 3100, i mean matchmaking system must be balanced also, if it found you a player he must be close to your rating(which is a real big problem) but if you faced against demigod he's lvl must rise up close to your rating no matter which bot it is, this will be more funny and interesting and will bring in the game variety even low lvl bots with new outfits.
    And many, many more, little bit tired and can not remember all the cases, please add your opinions.
    Thnx.

  • #2
    As someone who has played around a thousand games, in most cases, the proc rate doesn't reflect the percentages. Some say it evens out but I don't see it. If you have a 25% crit, more often than not, you will crit higher than the percentage. If you have 55% crit, you will crit lower than the percentage. The masteries seem about right though. These are just my observations ,and honestly, it really doesn't bother me in the least. You can't win em all unless the RNG gods love you. I have fun playing and will keep playing as long as player base keeps growing and updates keep coming.

    Comment


    • #3
      It's like the parry rate. Way above the percentage.

      Comment


      • #4
        The "chance thing" have been brought up numerous times, and in a lot of different games as well - the main issue it come down to is: computers do not understand the concept of "random"

        From my experience, (12+ years as a Software Engineer), a "random" value is the result of some seemingly complex formula, taking at least one input value, in order to attempt to prevent repetition of the same result within a particular time interval, or number of repeated calls. But in the end, given the same conditions/input values, it will result in the same value. You can try to over-engineer this for the rest of your life, and there's no way to properly get a "real life random/chaos theory" sort of result (maybe possible with Quantum computing, I don't know...)

        Even the concept of the probability itself is not fully understood by many people (my apologies up front here, no offense is intended, just trying to clarify things, hopefully)
        ie. "10% chance" can be "casually interpreted" as 1 in 10 - but it's not, it's: "each time you do X, there's a 10% chance of getting Y", so it can range from never to all the time and still satisfy the "10% chance" label. Neither extremes, nor anything in between, are errors...


        Edit:
        The above kinda makes the % value itself seems kinda pointless doesn't it, things just come down to how "lucky" you are at any particular time.

        For a real life % experiment: Go out and ask strangers for a number between 1 and 100 - how long does it takes until you get one that falls between 1 and 10? I doubt it's 10 times, and it won't be "even" either...
        Last edited by Nhat; 09-17-2018, 01:25 AM.
        Eternium Files - links and details (Updated: 10 Aug 2018)

        Eternium Guides:

        Comment


        • #5
          If you're using classes that can wield weapons in both hands, especially assassin or berserker with their extra attack tokens, you can get a different play experience, by which I mean sentiment, than using a one attack Crusader with no extra attacks beyond chain strike. My Voodoo Needle assassin attacks a **** ton because I have beast fang at 45%, two chain strikes, and a vendetta to use. That's where I get the most "even" feel for the dice rolls. I just responded to a thread elsewhere about this; many turn based strategy game makers have gone away from %. Gerula was bold to design this into his game. Combatting against the negative sentiment of RNG, which inevitably involves "streaks", is a challenge.

          Some players build in lots of risk with crit builds. Others try to mitigate risk with builds that rely less on %'s, like a Crystal Skull/Magic Missile witch.

          Comment


          • #6
            It is possible to design a RNG routine which is more accurate in the short term. I don't know that people would like it though. Most Psuedo RNG are based off of random seeds in maximal length generators. If the length is 32 bits then you can get 31 0's in a row or 32 1's in a row. A suggested way is to use an if statement pair before the random that either forces a crit if the expected number of Crits is too low or too high like if(attacks_since_crit > 250/crit_percent) crit_result = TRUE; else if (crits_in_a_row > 250/(1-crit_percent)) crit_result=false. If you use this code check against db0 errors first. 250 is arbitrary and could be adjusted down. 250 is based on gaussian (normal) pdf for 95th percentile. I am an engineer with 40 years of experience.
            ​​
            Last edited by Jo Dfo; 09-27-2018, 02:38 PM.

            Comment

            Working...
            X