Flameseeker Chronicles: Why wave-ban Guild Wars 2 cheaters?

Last week, our very own Bree Royce summed up a response given by ArenaNet’s Chris Cleary to a self-confessed hacker on the game’s subreddit who asked why he had not been banned for his actions. A list of offences was given in the post, and the indignant OP explained that the purpose of this bout of ill-advised honesty was to shame the company into taking swifter and more decisive action against cheaters. Guild Wars 2 Game Security Lead Cleary’s response to the hacker — who used the name MegaWormHole — received as many harsh words as the hacker himself, so I thought I should weigh in on the issue and discuss my views on how ArenaNet handles hacking (and cheating in general for that matter).

Cheating detection is a whopper of an issue for MMO developers to handle, especially since both fairness and equality of resource access are crucial in successfully managing massive online communities. Having said that, if an anti-hacking system is too rigorous, false positives can occur and honest players then suffer for those who decide to cheat. In this edition of Flameseeker Chronicles, I’ll break down Cleary’s response to this hacking admission while discussing batch banning and how adequately this deals with cheating in GW2. 

gw1MegaWormHole’s claim

The user started the complaint by highlighting that members of the subreddit have posted about being wrongfully banned and discussing how unacceptable this is, which commenters pointed out was a bit of a barbed comment coming from someone who causes such systems to be put in place. MegaWormHole was accused of gloating by more than one commenter, perhaps because of the nod to those who are most inconvenienced by cheating detection, but pointing out the fact that false positives occur sets the scene for his complaint and is perhaps intended to press blame on ArenaNet rather than goad those who are wrongly punished.

Most of MegaWormHole’s cheating started fairly recently, if his submission is anything to go by, and the main focus was SAB-related teleporting to accrue gold to be sold externally for real-world cash. One Redditor did a little digging and found that MegaWormHole had previously made a video on exploiting SAB that adds some validity to his complaint by proving the sustained nature of his hacking, which I’ve included below so you can judge for yourself. He also discussed no-clipping through PvP and speed-hack rushing points in WvW that were well behind enemy lines. Apparently clued into how ANet processes player accounts to detect cheating, MegaWormHole closed his submission by asking a fairly loaded question:

If these ANet customer support employees claim the system does not make mistakes, and the decision is final, why have I been hacking for a few months and have apparently been reported by other players countless times, and I’m still not banned?

Chris Cleary’s response

Cleary gave some context to his response by firstly admitting that stories like WormHole’s help him do his job more effectively by allowing him to better identify failures and pinpoint where in the chain things went wrong in cases where cheaters go unpunished. To address the fact that players are wrongfully punished under Anet’s current hack prevention triggers, Cleary pointed out that the anti-cheat system is “fairly forgiving when it comes to flagging players” and that detection is fairly accurate. Why are so many innocent players wrongly identified as cheaters — or vice versa — if this is the case? Simply put, Cleary puts any issues with banning the wrong players or letting cheaters slip through down to human error: As much as the team try not to wrongfully target players or let blatant hackers off the hook, mistakes can and do happen for a whole host of reasons.

Cleary states that MegaWormHole was flagged by the detection system for his in-game antics — meaning that he did not, in fact, avoid detection as he had suspected — but that he had slipped through the last wave of bans because of some human error in enacting the ban. He was apparently placed on a list “of about 300 others” and his ban should have happened back in early April. The details are sketchy in the response with no reason given for the slip-up and no pointing fingers, but Cleary was firm in his closing:

While we are doing our absolute best to handle players that can’t play fair, this means that it wasn’t a problem with our detection, but that we had a slip-up in our process to handle what our detection found. Now I go to fix that process so this doesn’t happen again.

gw0My thoughts on wave banning and the case at hand

Although I don’t know how frequent these waves are, I was rather surprised at the number of accounts on Cleary’s naughty list: I would have assumed that number would be much higher in truth. I’ve most definitely encountered botters and I’ve seen the occasional hacker inexplicably blink around nodes, and I’d imagine that’s the case for most MMO players. Hacking and cheating sort of goes with the territory in MMOs because of the real-world exchanges that occur for hard-earned goods and currency in-game. If there’s money to be made, you can bet there’ll be people ready and willing to exploit a pursuit in whatever way is possible to get their hands on a piece of the pie, and all devs can really do is tread water with such a significant and persistent influx of hacking activity.

Wave-banning is fairly frustrating for the majority of players who decide not to cheat: Seeing other people con their way to the rewards you work so hard for with no immediate punishment in sight is demotivating at best and all out off-putting at worst. The perception of inaction is hard to dissolve when the nature of game security means holding detailed metrics and specifics on hacker detection close to one’s chest. Reading a post such as the one written by MegaWormHole inflames the non-cheating GW2 community’s ire and instigates a wave of frustration at the fact that unfairness has — in this case at least –prevailed in our game. Cheating will always happen, unfortunately, and there’s no plugging the dam at this stage.

Wave banning staggers and occludes the feedback hacking communities usually derive from the usual insta-banning of offenders. People share information on which recently-installed hacks caused their instabans where they can, so wave banning is a good way to deal with hacks without causing some sort of regrouping or intelligence gathering opportunity for the hackers. In terms of PR, it’s impressive to see a wave of bans because big numbers tend to grab attention and act as a deterrent for the average gamer. For the security team, it saves a loss of focus and aids productivity to handle bans in large, closely related batches rather than dealing with flags on a per-person basis.

gw2The solution

I have to say that I never find cheating or hacking acceptable in MMOs and I don’t buy the argument that devs are ones to blame: No matter what a dev team throws at the problem, there’s no perfect solution. The ideal solution is that the client never has authority over information: The client should send commands to the server, but the server should be where all actions occur. EVE Online is a fantastic example of this with its server model that certainly removes client-side authority, but this isn’t feasible for most MMOs. There are relatively few commands issued per second in EVE, and a delay in that context is much more acceptable than it would be in a game like GW2. EVE runs on a one-second tick rate, whereas most MMOs are much higher.

So where does this leave Guild Wars 2? We don’t want to sacrifice the smooth gameplay for additional security, but we’re not happy now. We have to accept a happy medium: Some degree of “sanity checking” on the server-side allows flags to be thrown without such a drastic slowdown and also shows a fair amount of due diligence on the part of the devs. Judging by the example of an erroneous jumping puzzle teleporting bug flagging up three years ago, I believe that’s happening at ArenaNet. I believe Cleary when he says that the human side of the process is where the fault lies and I urge him to look at how the banning process is handled after the point of detection.

Each flag needs a human eye to confirm, which is a significant investment issue for ArenaNet. Autobanning upon flag wouldn’t be acceptable, but neither is throwing a potentially hasty, overworked, or outsourced eye — perhaps that of a person with metrics to meet and no time to spare for the individual nature of each case — at the matter. The main benefit of wave banning has to be that an informed and prepared person sees each case and deals with like matters in a batch, which no doubt is an important way to link users who participate in hacking or cheating and analyse those behaviours.

Over to you!

Not only will false positives be reduced by an unrushed, well trained human reviewing each case, but false negatives such as the case at hand will also reduce as the security team and wider customer service team use their powers to deal with both the immediate problem, swiftly reverse any false positives, and also sniff out chains of unwanted activity to grasp the bigger picture. It’s a goliath task with massive room for improvement across the whole MMOsphere, but it can be done.

Have you witnessed exploitative behaviour in GW2, and if so, do you feel as though it was addressed? Does ArenaNet do enough to stop cheating? How much human error is acceptable when dealing with hacking claims? How would you handle the issue? Let me know your thoughts on the matter in the comments.

Tina Lauro has been playing Guild Wars 2 since it launched and now pens the long-running Flameseeker Chronicles column, which runs every other Wednesday and covers everything from GW2 guides and news to opinion pieces and dev diary breakdowns. If there’s a GW2 topic you’d love to see covered, drop a comment Tina’s way or mail her at tina@massivelyop.com.