Rampant cheating in PVP?

[verdres]

I don't think the devs intend to add any kind of anti-cheat, and everyone would probably experience performance issues if they did that. I think a reasonable solution would be implementing battle logs that get recorded server-side, and then when players submit complaints during that match they can review the logs for "illegal" actions. For example, instantly swapping to a low-hp target, consistently and abnormally quickly. Or using skills with precision timing that does not align with normal behavior. Or even very egregious behavior such as using skills that ought to be on cooldown.

They also can look out for players exploiting existing bugs, such as the WHM heal charge bug. OR they could, with the cooperation of modders, get examples of mod behavior that can make it easier to automatically cross-reference from battle logs.

[Commander_Justitia]

reasonable solution would be implementing battle logs that get recorded server-side, and then when players submit complaints during that match they can review the logs for "illegal" actions

I think that is simply just too much work, only their SpecialTaskForce does that and they don't have that many of them and only do it when there are tons of reports. Would be easier if they can develop an AI that detects those patterns compared to using human resources, which are just too expensive for all those games.

They should shift ranked mode probably into party vs party only and not have it in solo anymore. Cheating will be less effective against coordinated teams and SE only would have to really check those games and not every solo game, which is a lot less trouble to deal for their GM's.

And they should probably let at least 2 people queue up together for the normal/unranked queue, I can imagine many people who would like to play with at least 1 friend.
In The Feast you were able to as well and there were only 4 people in a party.

[Taliriah]

I think that is simply just too much work, only their SpecialTaskForce does that and they don't have that many of them and only do it when there are tons of reports. Would be easier if they can develop an AI that detects those patterns compared to using human resources, which are just too expensive for all those games.

Yeah, that's exactly the kind of situation where a client side anticheat can actually work wonders. Sure, it's possible to circumvent them but serious products can update themselves on the fly while the game runs. The client integrity checking code that they use can be randomised, and changed whenever they want, and that makes it much harder to defeat durably. It would be better if there was also some server side validation, wow for instance despite having authoritative clients movements (which feels a lot better for the player), have the server detects some type of anomalous movements on the server side (with some well known false positive that result in instant disconnections, mostly involving gliders), but even in the absence of this a good client side anticheat is a lot better than nothing.

Something like battleye is very turnkey, the integration is pretty painless and then they send you the IDs of player who have been detected to cheat, they can even preemptively ban them (by preventing the client from starting when battleeye connects) so you can verify the reposrts and decide what to do with them (perma ban them, give them a warning, have your own policy regarding re-offence, etc). I'm sure other similar anticheat solutions are equally easy to implement. Heck, I'm sure that you could enable it only during pvp duty for the sake of harmless mods used outside of pvp.

The downside is that it's costly. But it's not like they can't afford it. I'm going to assume that whatever reason they offer for not doing that is window dressing for "we don't feel like investing money to solve the problem"

[verdres]

Would be easier if they can develop an AI that detects those patterns compared to using human resources, which are just too expensive for all those games.

Definitely agree, at a glance it seems like a very efficient solution. Depending on how predictable bot/addon behavior is, they could even flag a specific sequence/timing of actions to get a good idea of who to scrutinize.