This just makes me appreciate Doom's replay files, which make it virtually impossible to cheat speedruns since every single input can be read and played back at will, making it easy to detect...
This just makes me appreciate Doom's replay files, which make it virtually impossible to cheat speedruns since every single input can be read and played back at will, making it easy to detect whenever something is off. I wish more games had this feature, although I completely understand speedrunning isn't on the mind of most game devs as a priority.
It's a tough balance to achieve for sure, you want to make leader boards accessible (i.e you don't need the fanciest recording setup to submit), but also strict enough that you can catch cheaters. I think for most communities this is "solved" by having a rule like "any times lower than x have to adhere to these stricter rules", to make sure that actual world record times are subject to the strictest criteria.
Of course it will always be a cat and mouse game, and with PC games it gets a lot uglier as this article points out, since modifying the game itself becomes possible, but ultimately I still think as long as you have a solid community and well thought out rules, the majority of cheaters will be weeded out at some point. World record runs are always scrutinized no matter what, and if nobody can reproduce whatever the world record holder did, then that record holder will eventually lose their title.
Would Doom's feature protect against a TAS-style fakery? I mean, no one could go full TASbot and argue they were doing it themselves, but a spliced run with inputs from segments they recorded...
Would Doom's feature protect against a TAS-style fakery? I mean, no one could go full TASbot and argue they were doing it themselves, but a spliced run with inputs from segments they recorded themselves seems like it would get past that sort of check.
You could try to fake it via TAS, and people have tried, but it's very hard to fake human movement with a TAS because every human run is imperfect in unique ways that are hard to replicate. There...
You could try to fake it via TAS, and people have tried, but it's very hard to fake human movement with a TAS because every human run is imperfect in unique ways that are hard to replicate. There are several people who know the ins and outs of how a human player behaves with regards to input in Doom, you would need to be an expert to even begin to think of faking it via TAS.
And even then, in Doom's case there's more working against you since timing is done with in-game time and that only goes to the second, so time saves are done in increments of 1 whole second. Thus, since everything is incredibly optimized by this point since game is almost 30 years old, you would be scrutinized to hell if you actually beat any of the Doom world records, and you are almost guaranteed to be caught cheating.
I'm not saying it's impossible to cheat in Doom, of course, but it is a much harder feat in its case, thanks in part to its replay files, hence my comment.
I know it's hard to fake human movement. That's exactly what I said. I was proposing a player recording their own motions and splicing those best segments together to create a more perfect run...
I know it's hard to fake human movement. That's exactly what I said. I was proposing a player recording their own motions and splicing those best segments together to create a more perfect run than they could maintain on their own. TAS isn't restricted only to programming the button presses individually. Unless there was a splicing error you could view all of the button presses and see that every bit of it was human generated, but that wouldn't show you that it didn't actually happen all in one play.
This just makes me appreciate Doom's replay files, which make it virtually impossible to cheat speedruns since every single input can be read and played back at will, making it easy to detect whenever something is off. I wish more games had this feature, although I completely understand speedrunning isn't on the mind of most game devs as a priority.
It's a tough balance to achieve for sure, you want to make leader boards accessible (i.e you don't need the fanciest recording setup to submit), but also strict enough that you can catch cheaters. I think for most communities this is "solved" by having a rule like "any times lower than x have to adhere to these stricter rules", to make sure that actual world record times are subject to the strictest criteria.
Of course it will always be a cat and mouse game, and with PC games it gets a lot uglier as this article points out, since modifying the game itself becomes possible, but ultimately I still think as long as you have a solid community and well thought out rules, the majority of cheaters will be weeded out at some point. World record runs are always scrutinized no matter what, and if nobody can reproduce whatever the world record holder did, then that record holder will eventually lose their title.
Would Doom's feature protect against a TAS-style fakery? I mean, no one could go full TASbot and argue they were doing it themselves, but a spliced run with inputs from segments they recorded themselves seems like it would get past that sort of check.
You could try to fake it via TAS, and people have tried, but it's very hard to fake human movement with a TAS because every human run is imperfect in unique ways that are hard to replicate. There are several people who know the ins and outs of how a human player behaves with regards to input in Doom, you would need to be an expert to even begin to think of faking it via TAS.
And even then, in Doom's case there's more working against you since timing is done with in-game time and that only goes to the second, so time saves are done in increments of 1 whole second. Thus, since everything is incredibly optimized by this point since game is almost 30 years old, you would be scrutinized to hell if you actually beat any of the Doom world records, and you are almost guaranteed to be caught cheating.
I'm not saying it's impossible to cheat in Doom, of course, but it is a much harder feat in its case, thanks in part to its replay files, hence my comment.
I know it's hard to fake human movement. That's exactly what I said. I was proposing a player recording their own motions and splicing those best segments together to create a more perfect run than they could maintain on their own. TAS isn't restricted only to programming the button presses individually. Unless there was a splicing error you could view all of the button presses and see that every bit of it was human generated, but that wouldn't show you that it didn't actually happen all in one play.