More from The Dotabuff Blog
70 Kommentare
2006.우리가 얼마나

    jopa chlen zalupa

    Dieser Kommentar wurde geändert

      Я НЕ ПОНЕЛ!


        First English?

        Professor de Português

          Espanish please


            Great article, Gorgon.

            Joint Doctor

              nice learning )


                Great to have the understanding of this but I don't know if most people at any skill level would really be able to pay a lot of attention to their procs unless it's a solo Lane. I'll add that I'm in the normal skill level so I might very well be wrong.

                Dieser Kommentar wurde geändert

                  Will test it out


                    Terrific article! Made easy to understand :D


                      Great article, I would really love to see a table to other heroes, not only Slardar, but really well done overall.


                        "Non-hero abilities also use true random"
                        So if you devour a wolf as doom, you'll get true random crit?

                        Fake yaphetS !

                          rly rly usefull tnx a lot


                            чет лень читать ;/


                              This is amazing with the earthshaker shadowblade crit build. you walk into the jungle, you get 5 misses on your crit, you search for opponents, go invis with the shadow blade (ofcourse you use enchant totem before you go invis) and you have a much higher chance to perform a critical strike and destroy your opponent.

                              Dieser Kommentar wurde geändert
                              Sup m8

                                17% Boys


                                  Awesome article!

                                  Dieser Kommentar wurde geändert

                                    Good one.

                                    Just one thing that isn't clear enough about the graph of true and pseudo random.
                                    It's the chance to touch your proc in n "hit" right after your just proced.

                                    Пожилой кролик

                                      Я частенько поебываю гусей.


                                        please russian


                                          Что делать,если не знаешь английский?

                                          Boundless Strike in Gaben...

                                            really hard to obtain any real practical relevance from the article and the subject as a whole.....many more factors to consider then just when your passive is going to game


                                              Well. This looks complex.


                                                Does stacking the chances for proc work if you attack something you'll 100% miss? (e.g. Windranger while windrunning)


                                                  Great article!


                                                    Too much math, my brain just melted.


                                                      17% is really boys



                                                        You can make a crit on a missed attack, so nope (mb for mjolnir it does work however, i should test it ^^)

                                                        Dieser Kommentar wurde geändert

                                                          what about ogre magi's ultimate and canceling the animation on the stun. will it increase the odds or are the odds calculated after the stun goes through?


                                                            Очень интересная статья. Если вкратце, она о том, что можно искусственно получать пассивный эффект (баш, стан, крит) перед гангом или замесом. Для этого можно бить своих крипов, либо отменять атаку несколько раз - тогда следующий удар будет с эффектом. Кол-во ударов зависит от %, который имеет эффект.


                                                              This changes everything :)


                                                                I dont think the right graph on the top you have there is correct. I work with pseudorandom numbers on a daily basis. Generating pseudorandom numbers is a non-trivial task, and years of dedicated research by mathematicians have produced highly sophisticated algorithms that produce random numbers that are almost indistinguishable from true random numbers. Your average household calculator with a rng function can even give you a good binomial pseudorandom number with a 17% chance of an event happening. The notion that the rng in Dota 2 produces a distribution different from the 'real' one is HIGHLY unlikely (if it does indeed use an RNG at all), and even saying it makes you look stupid and uneducated. I'm sorry but this article is trash.


                                                                  I'm trying to understand why in the pseudo random scenario you suggest that after 11 hits of not procing the event (bash, crit, etc) why is there a sharp decline in cumulative odds.


                                                                    Yeah, I guess then crit doesn't work, the others might.
                                                                    Ogre's ult is True Random (as the article says), so it doesn't matter if you cancel the animation.

                                                                    Dieser Kommentar wurde geändert

                                                                      after so many hits (11 for example), you have had so many chances/hits beforehand to proc the event that getting to such a high number without a proc is very unlikely. The graph is simply showing how likely it is for you to proc after each successive attempt (for the first time). It would be extremely unlikely for you to get 15 hits in WITHOUT procing the event, hence why it is such a low possibility.

                                                                      Old Monk

                                                                        my english gud vry..commend


                                                                          Lets say I am playing troll and I went 9 attacks(in meele form) without a bash and now I am saving the 10th attack to hit a hero but then I get a lvl up and I lvl up my 1st skill.Does this mean that the odds for a proc of bash on my next attack also reset or it stays the same i.e. high chance of proc on my next attack?

                                                                          Dieser Kommentar wurde geändert

                                                                            @Cunnilinguistician: You are such a noob and uneducated, the pseudo-random dota 2 uses is different from the real one. Ice frog said so. Period. Btw I'm economics teacher and work with random number too, both pseudo and real.

                                                                            Dieser Kommentar wurde geändert

                                                                              this is really interesting to read


                                                                                really, wasting of time counting unsuccessful chances.

                                                                                typical. 🎃

                                                                                  @Cunnilinguistician - For someone who claims to be so educated, your post sure is a load of utter rubbish.

                                                                                  aikyu3 SuBi ♪

                                                                                    so deeeeep


                                                                                      Commending author.

                                                                                      More game mechanics articles please :)


                                                                                        So much meth, I mean math.


                                                                                          Fyi the hero this is the most exploitable on is CK -- his animation is slow enough to be able to "release" cancel when you're trying to get a crit ; practice a couple minutes and you should be able to release cancel based on sound. I've been called a cheater for killing people with 4 crits lvl 1 if my lane partner has a slow/stun.


                                                                                            And get off my obstacle you fat tub of lard.


                                                                                              Короче, псевдорандом-миф, Габен-гей, и сын шлюхи.

                                                                                              surprise tax fraud

                                                                                                Currently the Enchanted Mango only has a 31.43% win rate. Together, we as a community, have the ability to fix this problem and end this horrific tragedy. Please help. It is our duty as gamers to give this item the respect it is due.


                                                                                                  @Mandalord That's great. You probably work with monte-carlo simulations then. Perhaps you even use stochastic differential equations in your work (i.e. using ornstein-uhlenbeck processes to model interest rates, etc.) Then you should know that all random numbers generated by computers are 'pseudo-random'. If you knew the seed of the RNG, you can predict exactly what the next random number will be. In an effort to obtain 'true' random numbers, some wall street firms actually use data obtained from movements of particles in the stratosphere, remaining true to how 'Brownian Motion' was discovered. There is no such thing as a 'real' random number generated by a computer. Everything is done algorithmically, in very clever ways and using parameters that have been carefully calibrated. If you are curious, you can look up 'mixed linear congruential generator' or 'multiplicative congruential generator' as these are the popular ones.

                                                                                                  Since you are so smart, please enlighten me: the gist of this article seems to say that for some heroes, the likelihood function is somehow different from a series of independent Bernoulli trials. The only way I can somehow see that this article would be correct, is if the random numbers are somehow correlated (i.e. the independence and identically distributed assumption is violated). Would you say that this is what you had in mind as well?


                                                                                                    Я человек простой: вижу иностранный пост - пишу "АААА СЛОЖНА".