4 ms·
Anyone who even remotely looked at the numbers and followed the situation can tell Dream cheated. The only people I see claiming otherwise are either infatuated
by Scaless 5y ago
Anyone who even remotely looked at the numbers and followed the situation can tell Dream cheated. The only people I see claiming otherwise are either infatuated with his cult of personality or otherwise just don't give a shit, "It's just a kids game who cares"? I care.
I am heavily invested in speedrunning culture, but myself am mostly involved in the Halo community. I have a couple of world records of my own, and I also do routing and other technical things for the community.
Halo has its own long history of cheaters, many of whom were also pivotal founders of the community. In the wider speedrunning scene there are many examples of similar situations. I do not think that speedrunning necessarily has a higher percentage of cheaters, if you compare against other competitive sports and activities the rates would likely be comparable. But that number is far from zero.
I am working on developing TAS (Tool-Assisted Speedrun) tools for the Halo games. With the tools you can create frame-perfect recreations of gameplay that are indistinguishable from actual play.
I am honestly at a stand-still on whether I should ever publicly release a working version of the tools. The main worry is that the tools will be used to submit "fake" runs passed off as human gameplay. I have built in basic safeguards to deter the most basic cheaters, little things that you would never see as a casual player but if you know what to look for would stand out in a video. But there is always the chance a sophisticated user could reverse those changes and fly under the radar.
On the one hand it could open a floodgate of awesome content, tricks, and discoveries. On the other it could potentially completely ruin the integrity of the leaderboards. I am quite conflicted on how to continue with this project if anyone has any insight that might be helpful.
- Olreich 5y agoI can already do this for most popular speed games, and it’s not a big problem. The TAS tools will help the community much more than they will hurt integrity.
- gvx 5y agoDisclaimer: I am not part of the community (for example: I wouldn't even have known about the Dream scandal if it weren't for the video by Matt Parker), so feel free to take my opinion with a bucket of salt. It might be an idea to decide as a community that the time of completely unassisted speedrun leaderboards has passed, and shift the competitive element of your community to technological advancements in TAS tools.
- Scaless 5y agoSorry if I did not explain it well, TAS does not mean just small helpful enhancements. In a TAS, the game is being played by the computer using a series of pre-programmed inputs. The fun in TASing is creating the most optimized set of inputs that completes the game in the shortest amount of time. There are often precise one-frame tricks or inputs that would not be possible for a human to perform but can be done easily when you can fully control what happens on each individual frame of gameplay. Mixing TAS and non-TAS runs doesn't make much sense as humans would never be able to compete.
- gvx 5y agoNo, I meant more like... I don't know if it's possible to verify if a run submitted for non-TAS leaderboard isn't secretly a TAS run and if not, the era of any integrity of non-TAS leaderboards would be over. So all non-TAS speedrunning would be non-competitive in the future, if there's such a thing as non-competitive speedrunning.
- yifanl 5y agoFrame perfect tricks are often fairly easy to spot, and heavy usage of then are pretty indicative of TAS runs. Of course, there are folks who can do those frame perfect tricks, but if it ever does come to contention, then asking them to perform said trick live seems like a pretty easy way to confirm.
- flaubere 5y agoYou just end up with a situation where there's intense competition to make cheatbots as good as possible, within the parameters of plausible deniability. People start meta-hacking to find tricks which don't improve the speedrun but keep it impossibly fast while making it look more convincingly human. People spend time obsessively training not to get good at the game, but to make it seem as though they would have been good enough to do a speedrun which was actually automated. Eventually you start to get false positives as the bar for cheating catches the most skilled and dedicated non-cheating players. Then again some people still enjoy watching professional cycling so maybe it will be fine.
- episteme 5y agoIsn't this true of basically all tools developed for TAS? They are cheats by the normal definition. It would be a shame to hold back on finding out what is possible in Halo because of cheaters, cheaters will find a way with or without your help. If people continue to care about these runs then someone eventually will develop a tool like yours and they may not care as much about people cheating. I would suggest putting some effort into making your tools detectable somehow so that you can at least have a clear conscious about it but don't let the fact that people will always cheat stop you from providing useful tools.
- cwkoss 5y agoI wonder if you could add steganography to the visual output to watermark runs in a way to prevent your tool from being used to pass off live gameplay.
- rcxdude 5y agoPeople have tried to pass TAS runs off as RTA before, and often it's really obvious (especially to someone experienced at speedrunning the game) because TAS runs don't at all look like a human playing them.