To give the gist to everyone (still watch the video, it's well worth it), if all 10 billion people on earth would be playing minecraft for every second (actually playing the whole speedrun in 1s, but that's details), for 100 years. The chance of his result happening to any one human is 1/1000. Alternatively, dreams result is 10 orders of magnitude less unlikely to have happened than the next "lucky" event that happened which was in craps.
> “Whoever wrote that is either deliberately manipulating numbers in favor of Dream or is totally clueless despite having working experience with statistics. Familiarity with the concepts is clearly there, but they are misapplied in absurd ways.”
> They also wrote that the “amateur mistakes” that appear in the document make them question the overall qualification of the (anonymous) author.”
https://www.dexerto.com/minecraft/dreams-minecraft-speedrun-...
Going by the rest of your comment, you would've learned something.
I can tick the box in my head which says dream is cheating - but it's not a smart move from where is standing.
The video discusses this and tries to provide their own intuition besides the two papers, as both agree on the core odds of the likelyhood being around 2 x 10^22. They contrast this with the number of events if 10 billion (10^10) humans would do something for 1 century (10^2) every second (~3.15 x 10^7): ~3.15 * 10^19. So the odds of the event happening if more humans than we have today do it every second for a century is roughly 1 in 1000.
Lastly, they postulate it doesn't matter if they cheated, but the sheer unlikeliness shows that the minecraft run is not performing up to expectations.
I would put this at better than 1/10^22. If you claimed that your religion was real and showed me an incontrovertible 'miracle' which had a chance of less than 1/10^20 of occurring by chance, I think I would accept your claim.
We all have our own priors, I guess, but personally, I consider one of those to be more likely :D
This is addressed in the video. The report from the anonymous statistician hired by Dream does not actually dispute the incredibly low odds.
I've never heard of Dream before and I have no horse in this race. But the probability of this actually having happened by chance on an unmodded set of runs is so low I cannot believe it. At the same time, the someone who has made a career out of running Minecraft with mods having a great run of luck with the help of mods and then claiming that it was fair does not seem very unlikely.
*If every single human being on planet Earth each completed a Minecraft speedrun every single second for 100 years (no sleep), the chance anyone during those 100 years getting as lucky as Dream just once is less than 1:1000.*
I can believe he's cheating but I don't consider it probable. Dream's success doesn't come from record settings, so I don't think he has any incentive to cheat.
If anything, I believe more the allegations that his manhunt videos are staged, given that's the interesting content.
He makes money by creating long and entertaining videos (about modded Minecraft and his insane reflexes). This ticks off the YouTube algorithm which measures how much of a video viewers watch and how much time they watch a channel, contributing to exponential growth. On top of this, Minecraft fans are a niche that lives on YouTube.
Dream never became popular because of his speed records, he gained initial traction by creating content on PewDiePie (an infinitely more popular YouTuber). This is exploiting the YouTube's algorithm which recommends videos: by talking about someone famous you'll eventually end up being recommended to their huge user base.