How Diablo hackers uncovered a speedrun scandal
Investigators decompiled the game to search through 2.2 billion random dungeon seeds.
But Groobo's record is still listed as the "Fastest completion of an RPG videogame" by Guinness World Records, which has not offered a substantive response to the team's findings (Guinness has not responded to a request for comment from Ars Technica).
Of course they didn't. Guinness has never cared about being correct. People pay Guinness to have a record listed. The record holders are the Guinness records' customers.
They did quietly remove a few of Tommy Tallarico's records, halfway through that Hbomberguy video.
Sure, but I think the key word here is quietly.
They'll try to hide the most embarrassing stuff, long after the fact, but I wouldn't expect them to comment publicly on it. It'd be bad for their business.
Full disclosure, I haven't actually read the article but is it really a scandal? Just create a new speedrun category for preset seeded runs. The old record stays and people can now try to hunt for the most efficient seeds.
That's the thing. They scanned EVERY possible seed, and the run in question didn't match ANY of them.
They also found that a key item found in an optimal location was only possible in a version of the game that didn't match what they claimed to be running.
They also noticed that the lvl 12 sorc was dealing damage as if he was level 20 something.
The group investigating just wanted to see how they would stack up if they could recreate the original streamers perfect luck, and they couldn't even do that. It really seems like the run was faked.
They didn't use a single run but multiple seeds in split segments. They also used illegitimate seeds to get a rare item earlier as well. Also somehow killed some boss with less fireballs but while possible they found the level needed to do it was in high 20s while his speedrun showed 13.
Yes it's kind of a scandal. And it isn't a predefined seed. It's spliced and most likely had changed stats and drops. It was submitted as spliced, which is allowed, but then the splices should have had the same seed, items, HP, etc. Which are all inconsistent.
The article say 15 feb 2025, but I swear I read that exact same story about a year ago.
I've read this yesterday, and it's kinda weird.
The community says they've debunked his cheating and needed months to do so, going through every possible seed to determine that the dungeon in the video doesn't exist and the very important and early item drop is not naturally possible.
But the video description reads weird, it stated that a) it's a spliced run and b) he manipulated the item to drop as early as possible.
...done in 27 segments appended to one file...
...cut down the time almost in half and it's due to even more luck manipulation...
...The most important thing is that I manipulated Naj's Puzzler to drop from the earliest monster possible...
Note that I took these from an archived shot of the video in 2014.
That's not what "spliced" and "manipulation" mean.
A spliced run could play each floor a dozen times and continue from the best outcome. This video jump-cuts to a completely different dungeon.
Luck manipulation means doing weird shit so you can predict a random number. This video straight-up cheats for numbers the game would never roll.
Huh, thanks for clearing that up.
So what exactly is the implication of that description, that the run was made with a fixed seed and maybe multiple tries for some of the levels, but on the same seed?
Isn't it then extremely suspicious, that it's not stated what seed was being used / how it was fixed?
Do you have a link to the archived screenshot?
https://web.archive.org/web/20140314230825/www.youtube.com/watch?v=dW8XorD9RwI
I took the quotes from the HTML file, though, because the drop-down didn't work. The description is the same as today.
