Dream Speedrun Cheating Controversy - Images
HOLY FUCK


Dream Speedrun Cheating Controversy
Critique of Dream Investigation Results - Conclusion


Dream Speedrun Cheating Controversy
Yes, I think Dream cheated. No, I don’t care. Yes, the people who wrote the paper were nerds. But al...


Dream Speedrun Cheating Controversy
Looks like rigging the Minecraft mob vote certainly paid off. didn't it?


Dream Speedrun Cheating Controversy
I can't believe Dream rigged the lottery too


Dream Speedrun Cheating Controversy
All of the day bro


Dream Speedrun Cheating Controversy
Dream Cheating Allegations: Dream's Reddit Post


Dream Speedrun Cheating Controversy
Dream Investigation Results: Conclusion


Dream Speedrun Cheating Controversy
Blaze Rod Drops: Dream vs. Other Speedrunners Graph


Dream Speedrun Cheating Controversy
Piglin Bartering: Dream vs. Other Speedrunners Graph


Dream Speedrun Cheating Controversy
Just to clear up some misinformation regarding the speedrunning thing.


Dream Speedrun Cheating Controversy
Thank you for the totally unbiased, 2 month, 29 page "investigation"


Dream Speedrun Cheating Controversy
My 1.16 run was just rejected after research due to it being "too unlikely to verify".


Dream Speedrun Cheating Controversy
Dream Adresses Cheating Allegations pt. 2


Dream Speedrun Cheating Controversy
Dream Addresses Cheating Allegations Comment


Dream Speedrun Cheating Controversy