Policy Change needed on "RMT & Services"

[Sythen]

The are indeed easy to ignore, but if we continue to maintain the same stance on all of this silliness, its going to get out of hand.

[SamSmoot]

The are indeed easy to ignore, but if we continue to maintain the same stance on all of this silliness, its going to get out of hand.

I constantly report them in the cities, but it's a futile effort. The same ones have been spamming the same URL for months now. They really need some automatic blocking, using players to recognize the posts.
Say, 100 players report RMT posts, software compares them, and finds 50+ with the same string, such as "Xgold.com", or "Safe & easy gil sales", then block everything that contains that matching string. This would force the sellers to keep changing their ads, until they run out of ways to say what they want.
Right now, they just ban the account, and it's back 5 minutes later on anew account, running the exact same ads.

[Eriane_Elis]

I constantly report them in the cities, but it's a futile effort. The same ones have been spamming the same URL for months now. They really need some automatic blocking, using players to recognize the posts.
Say, 100 players report RMT posts, software compares them, and finds 50+ with the same string, such as "Xgold.com", or "Safe & easy gil sales", then block everything that contains that matching string. This would force the sellers to keep changing their ads, until they run out of ways to say what they want.
Right now, they just ban the account, and it's back 5 minutes later on anew account, running the exact same ads.

Yes, they need to have a reasonable chat log monitor put in place. They could have it automatic on a separate server that queues these messages. They could use machine learning and anything that seems suspicious can be brought to a GM's attention automatically and they could spot ban them. They don't even need to design the system themselves, they could just buy existing software that yes, exist today to do this.

They would just need to have a lot of logs say, 10000 spam entries and 10000 non-spam entries and they could train the A.I. on that. There would be a % of failure but it learns through failure and success and because a GM would review the potential spammy message, no "good" accounts will be banned by accident. It's really not hard to implement. If done from scratch it would take maybe 2 weeks to do by a single developer.

[CarnivalNights]

Yes, they need to have a reasonable chat log monitor put in place. They could have it automatic on a separate server that queues these messages. They could use machine learning and anything that seems suspicious can be brought to a GM's attention automatically and they could spot ban them. They don't even need to design the system themselves, they could just buy existing software that yes, exist today to do this.

They would just need to have a lot of logs say, 10000 spam entries and 10000 non-spam entries and they could train the A.I. on that. There would be a % of failure but it learns through failure and success and because a GM would review the potential spammy message, no "good" accounts will be banned by accident. It's really not hard to implement. If done from scratch it would take maybe 2 weeks to do by a single developer.

This idea scares me because my wife and I trade large amounts of gil between each other since we share all our gil and I don't want either of us getting flagged because of a mess up on their end and it'll end up taking us weeks and maybe months to even resolve.

[SamSmoot]

They would just need to have a lot of logs say, 10000 spam entries and 10000 non-spam entries and they could train the A.I. on that.

They don't need AI to recognize the RMT posts, the players can do that. All the software needs to do is simple text string matching between the reported posts, and if a certain threshold of the reports contain the same string, then block all of them that contain it.

[Eriane_Elis]

They don't need AI to recognize the RMT posts, the players can do that. All the software needs to do is simple text string matching between the reported posts, and if a certain threshold of the reports contain the same string, then block all of them that contain it.

If it could have been fixed with regex alone, the problem would have been solved a long time ago. They need A.I.

[Tiraelina]

If it could have been fixed with regex alone, the problem would have been solved a long time ago. They need A.I.

I severely doubt regex was ever used to begin with. They started filtering specific messages, bots responded by adding randomly generated characters at the end of the messages completely bypassing it and nothing changed since.

The reliance on automation and latent detection ban waves is why we're in this situation at all. I don't know who any of these businesses think they're helping with this system because it's not the players. It's the same garbage in any game that isn't directly dealing with it.

[Kuroka]

The are indeed easy to ignore, but if we continue to maintain the same stance on all of this silliness, its going to get out of hand.

Start a BLM twink.. its already out of hand >_>