logoalt Hacker News

jsheardtoday at 1:39 AM3 repliesview on HN

Easier said than done for some genres, unfortunately. To catch things like aim assistance from the server-side you'd have to resort to handwavy statistical analysis and somehow thread the needle between catching well-crafted aimbots, but without accidentally banning legitimate players under any circumstances, even if they're extremely skilled and/or lucky.

It's been tried but I don't think it's ever been very successful. The Battlefield series used to use Fairfight, which is based on server-side heuristics, but they ultimately gave up on it and switched back to client-side detection for the more recent games.


Replies

nickjjtoday at 1:05 PM

I used to co-run an online gaming ladder back in the Quake 3 days.

There were aim bots and other client side hacks back then but we requested that folks record and upload demos of themselves for competitive matches. This allowed anyone to replay the game from your exact POV in-game, complete with hearing and seeing exactly what the real player saw at the time.

We all survived back then without kernel level anti-cheat tools with a very high certainty that no one was cheating.

Even if you tried to hide it, it was pretty obvious when someone was cheating. I don't recall a single unsolved case where someone was cheating and got away with it while the community really thought otherwise. This was with over 10,000 registered players and tons of active teams playing every day. No where near the scale of gaming today of course, but it's a big enough sample that the method does work for online competitive play.

Nowadays it would be even easier to detect foul play because with live streaming and human announcers, you're under a lot more analysis by the public in real-time.

asmortoday at 3:13 AM

Cheating creates two problems. Obvious cheats aiming to upset other players and subtle cheats that don't want to be noticed as such. Now, I'm not saying it doesn't matter if people cheat as long as nobody notices or that competitive integrity is not important, but the first category is a much more immediate threat for most games and easy to detect. On a server analyzing locally recorded sampled demos.

MrDrMcCoytoday at 1:48 AM

That statistical analysis with post-facto game recordings could be pretty accurate, and needn't result in bans. In fact, I think banning cheaters is dumb. Instead, we should weight matchmaking algorithms to put cheaters all on the same servers. If they want to cheat, why not let them cheat against each other?

show 2 replies