Announcement

Collapse

Forum Rules

  • No flaming or derogatory remarks, directly or through insinuation.
  • No discussion, sharing or referencing illegal software such as hacks, keygen, cracks and pirated software.
  • No offensive contents, including but not limited to, racism, gore or pornography.
  • No excessive spam/meme, i.e. copious one liners in a short period of time, typing with all caps or posting meme responses (text/image).
  • No trolling, including but not limited to, flame incitation, user provocation or false information distribution.
  • No link spamming or signature advertisements for content not specific to Dota 2.
  • No Dota 2 key requests, sell, trade etc.
  • You may not create multiple accounts for any purpose, including ban evasion, unless expressly permitted by a moderator.

  • Please search before posting. One thread per issue. Do not create another thread if there is an existing one already.
  • Before posting anything, make sure you check out all sticky threads (e.g., this). Do not create new threads about closed ones.
  • It is extremely important that you post in correct forum section.

  • Balance discussion only in Misc.
  • All art related (such as hero model) feedbacks go to Art Feedback Forum.
  • All matchmaking feedback should go here: Matchmaking Feedback
  • All report/low priority issues should go here: Commend/Report/Ban Feedback
  • No specific workshop item feedback. These should go to workshop page of that item.
  • When posting in non-bugs section (such as this), use [Bugs], [Discussion] or [Suggestion] prefix in your thread name.



In case you object some action by a moderator, please contact him directly through PM and explain your concerns politely. If you are still unable to resolve the issue, contact an administrator. Do not drag these issues in public.



All rules are meant to augment common sense, please use them when not conflicted with aforementioned policies.
See more
See less

Pseudo Random Distribution

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

  • Pseudo Random Distribution

    For those of you who played DotA 1 and knew a bit about mechanics, you may have heard of PRD (Pseudo Random Distribution).

    For those of you who don't know, its an algorithm used in the Warcraft 3 Engine (specifically for bashes and criticals) that makes it less luck based and more likely to fit the probability.

    For example... (this is not 100% accurate)

    Let's say you have a 25% chance to bash.

    On your first hit, your chance to bash is ~10%.

    The second, ~23%.

    Third, ~38%.

    Fourth ~55%.

    As you can see, your chance starts out lower than the initial probability, and increases for each UNSUCCESSFUL bash that occurs, once a bash occurs, the counter is reset back to the original 10%.

    Originally posted by MDuh View Post
    a little bump

    So I was really intrigued by this thread and decided to test it myself. I took out my notes, skim through vb.net notes and made a windows service that monitors "Sound levels"(through stereo mix) outputted in the computer.
    The simple windows service appends a line to a log file when a certain sound wave reaches a level peak..

    when a sound reaches n% of sound level is reached, it will log something

    so what I did is,
    1) create a private game and turn cheats on
    2) start the game
    3) chosed Dragon Knight
    4) Leveled him up to "25"
    5) bought 2x Heart of Tarrasque and 4x Bloodstone
    6) spawned 2 dark seers, 1 allied, 1 enemy; picked vacuum skill
    7) spawned Slardar (enemy); leveled it to lvl 7; Leveled bash up to lvl 4.
    8) spawned Witch Doctor (ally); leveled it to 25; Leveled VooDoo Restoration up to level 4; bought: 2x Guinsoo and 4xBloodstone
    9) in every lane of dire and radiant near the tier 3 tower, I spawned 2 fountains so no creeps can push and possibly end the game..
    10) Vacuumed Dragon Knight, Witch Doctor and Slardar in the bottom cliff near the bottom rune (Reason: that place the hast the most quiet ambient noise)
    11) Run a Calibration test to determine the threshold of the sound when Slardar Proc'ed a Bash and normal attacks
    12) Cleared the log file
    13) Let slardar attack Dragon Knight Continuously
    14) Ran the Test while I'm sleeping

    Note: Voices, Commentator are disabled so when a tower is destroyed, no voice will pop-up

    Here is the log file that the program recorded
    http://pastebin.com/download.php?i=nLUxx09Z

    The test logged 16,880 attacks;
    The results was somewhat disappointing.
    out of 16,880 attacks
    Slardar only Proc'ed his bash 2,534 times
    Based on the results of the test, the Actual Stun Percentage was only: 15%
    LVL 4 Slardar's bash is rated at 25% chance.
    Maximum Miss Streak: 60
    Maximim Bash Streak: 4

    Immediately there, I stopped analyzing and wondered, Did my program has flaws? Is my methodology for data gathering incorrect?
    I'm not sure but I spent 2 hours of testing the program to really see if it is reliable for data gathering and not even once it has a incorrect detection.
    So for me, Dota 2's RNG is prolly flawed

    I'll update this for a graph of chance vs stun streak of slardar

    I'm curios to see Diabolic's test (who promised to will release numbers for this) if his test will vary or not..
    Originally posted by MDuh View Post
    SLARDAR LVL 3 BASH TEST

    Steps:
    1) create a private game and turn cheats on
    2) start the game
    3) chosed Dragon Knight
    4) Leveled him up to "25"
    5) bought 2x Heart of Tarrasque and 4x Bloodstone
    6) spawned 2 dark seers, 1 allied, 1 enemy; picked vacuum skill
    7) spawned Slardar (enemy); leveled it to lvl 5; Leveled bash up to lvl 3.
    8) spawned Witch Doctor (ally); leveled it to 25; Leveled VooDoo Restoration up to level 4; bought: 2x Guinsoo and 4xBloodstone
    9) in every lane of dire and radiant near the tier 3 tower, I spawned 2 fountains so no creeps can push and possibly end the game..
    10) Vacuumed Dragon Knight, Witch Doctor and Slardar in the bottom cliff near the bottom rune (Reason: that place the hast the most quiet ambient noise)
    11) Run a Calibration test to determine the threshold of the sound when Slardar Proc'ed a Bash and normal attacks
    12) Cleared the log file
    13) Let slardar attack Dragon Knight Continuously
    14) Ran the Test while I'm sleeping

    Note: Voices, Commentator are disabled so when a tower is destroyed, no voice will pop-up

    Here is the log file that the program recorded
    http://pastebin.com/8atb1kag

    The test logged 14,583 attacks;
    The results is intriguing.
    out of 14,583 attacks
    Slardar Proc'ed his bash 3,183 times which is higher than LVL 4 should supposed to bash
    Based on the results of the test, the Actual Stun Percentage was: 21.5182%
    LVL 3 Slardar's bash is rated at 20% chance.
    Maximum Miss Streak: 34
    Maximim Bash Streak: 5

    Level 4 Bash Test: http://dev.dota2.com/showthread.php?...l=1#post113392


    From the data you can see that sometimes there are large gaps between you getting bashes, PRD can almost completely eliminate the extreme circumstances and make it more accurate.

    As far as I have witnessed, this is NOT in Dota 2, and I am assuming it is intended not to be, but I would like to see other's views on if they should apply it to Dota2 as well.

    In my honest opinion, it SHOULD be. It is still slightly luck based, but not completely. The less complete randomness that happens the better in my opinion.
    Last edited by Blarrg; 02-29-2012, 06:56 PM.

  • #2
    [EDIT] reply no longer relevant
    Last edited by Zanez; 11-29-2012, 03:47 PM.

    Comment


    • #3
      Considering how much I've witnessed bashes and criticals occur in Dota 2, I can pretty safely assume it is not in Dota 2. I'd say I'm 95% sure.


      And I see your logic, reply to the thread with no valuable information or opinion on the matter at hand, I see, I see.

      Please don't waste your time replying if you don't have anything meaningful to contribute.

      Comment


      • #4
        [EDIT] Reply no longer relevant
        Last edited by Zanez; 11-29-2012, 03:47 PM.

        Comment


        • #5
          Whats wrong with anecdotal evidence?

          I don't know if you know how probability and percentage works but there is no possible way to say "There is PRD" based on any data that I would provide, because it is CHANCE based either way. The only way to confirm it is to have someone with access to the code confirm it.

          Now please be a good little feedback provider and actually provide feedback, instead of harassing people trying to suggest improvement. If it truely DID exist in Dota 2 already, then leave it up to a mod or admin to close the discussion.

          Watch your hostility by the way buddy

          Comment


          • #6
            In my opinion this topic is worth to be discussed. I didn't experience this yet, but if OP is right and PRD isn't implented yet it should be implented.
            Thats why Zanez is right too. We need to test this, we need the stats or even confirmation from a Dev to solve this issue.

            Comment


            • #7
              Originally posted by Zanez View Post
              You have no statistics to provide. I have not noticed this being a problem, or even significantly different from Dota1 at all... You have brought absolutely nothing to the table... just your anecdotal evidence against mine. The difference is, you have the burden of proof. The good news is, this is very simple to provide... so be a good little beta tester, go into dota 1, and play some games against bots with the heroes of your choice, and then load up dota 2 with the same heroes against some bots, and compare data. Bring back compiled evidence and data, and record them here. THAT would be something meaningful to contribute. Until you do that, you have nothing.
              Please do us a favor and go read this wonderful discussion regarding RNG and actual data compiled and recorded. It has been proven that RNG is working as intented HOWEVER doesn't make sense (especially for Slardar and Axe) over 16,000 attacks for each level acquired.

              http://dev.dota2.com/showthread.php?t=20582
              Originally posted by Blarrg View Post
              Whats wrong with anecdotal evidence?

              I don't know if you know how probability and percentage works but there is no possible way to say "There is PRD" based on any data that I would provide, because it is CHANCE based either way. The only way to confirm it is to have someone with access to the code confirm it.

              Now please be a good little feedback provider and actually provide feedback, instead of harassing people trying to suggest improvement. If it truely DID exist in Dota 2 already, then leave it up to a mod or admin to close the discussion.

              Watch your hostility by the way buddy
              Have a good read with link above!
              Last edited by Sandy; 02-29-2012, 05:19 PM.

              Comment


              • #8
                I don't think there is anything too off about the RNG currently, except for that level 4 slardar bash.


                The point of PRD is to be appliable to small sample sizes and still give accurate results, testing it in large quantities is pointless.

                Also you can tell by the data given there (miss streaks of 36 and higher), that PRD is definitely NOT in play in Dota 2.

                Actually I'll think I'll put that data in my OP.

                Comment


                • #9
                  Is there such a thing as 'too random'?

                  I think so. Something just feels off about the %to occur things in this game, crits/bashes/counter helixes, in comparison to dota 1.

                  Comment


                  • #10
                    Originally posted by Zanez View Post
                    So your logic is "I don't think it has this, therefore it doesn't."

                    K just making sure.
                    Try looking up how the PRD works in DotA. It prevents (absolutely prevents) intervals of longer that 11 attacks without a proc (for abilities with 25% chance). So no, it is not being used in DOTA 2, at least not for Slardar's Bash (which it was used for in DotA), that much I can say with absolute certainty. See the guide.

                    From what I remember of other tests, I recall similar behavior for other abilities (like critical strike), i.e. going longer than 11 strikes without a critical hit, but I don't have that data in front of me ATM.

                    Comment


                    • #11
                      Yeah I remember seeing that data before. Assuming it's accurate, something is very wrong. Even assuming there's a bug specific to level 4 Slardar bash, a 60 miss streak is insane. 34 miss streak with an ability that is proccing at about 20% is also way off.

                      Comment


                      • #12
                        Essence Aura + OD interaction

                        I believe OD's proc rate on Essence Aura is lower than 40%, but it can;t be confirmed because all I can measure is observed proc rate. At 1500 mana, I have been able to go for 15 proc-less streaks, more than once in a game. I've seen it happen twice, ever since the Feb. 26 patch. At 40% proc-rate the odds of this happening are well under 0.01%, so something might be up. After 1300 mana, you should theoretically never run out except in extreme cases. I'll keep testing it, though.

                        Repro:
                        Rank 1-4 Arcane Orb
                        Rank 4 Essence Aura
                        Get at least 1000 mana, 1400 preferred.
                        Try to run out of mana. You shouldn't, but I have been able to.

                        Comment


                        • #13
                          I've had the exact opposite for OD, but the aura doesn't work for everyone else in all the games I've played. OD can stay above 80% mana in all the games I've played with no problem, but playing as Drow Ranger with cold arrows on autocast my mana drains with no regeneration from the aura.

                          and to say with your repro, at level 4 it's a 40% chance to regen 25% of your mana. so 2/5 shots would regenerate your mana which if math were perfect you could survive on 500-700 mana.

                          Comment


                          • #14
                            Hmmmm, interesting. Anyway, 1000 is the "Magic number" for his mana. You're going to proc 4 times on 10 shots in theory, so those 4 shots had better cover for all 10 you take, which means each proc needs to return 250 mana, which is 1000 base mana, which you almost get for having level 4 aura anyway. Anyway, I'll keep an eye on this one.
                            Last edited by Yasuhiko; 03-05-2012, 06:35 PM.

                            Comment


                            • #15
                              RNG is broken, it works way too much or way too less. In the end you'll have the expected proc but it'll be like proc a lot or proc-o-never.
                              Winter solstice.
                              Originally posted by uebok
                              Muted scum shouldn't have an option to disagree with me. That's why they're muted.

                              Comment

                              Working...
                              X