ID:152001
 
In a matter of opinion from a player prospective would you like to see a game with a self sustaining admin system for example.

1.)Chat Monitoring(With ban/boot ability)
2.)Active help system(Report a problem then the program will try different solutions to try to fix it)
3.)Bully Protection(The game would monitor the kill/death ratio and determine if the user is mob kill happy or just playing.)
And a few more features.

Plus(In my opinion)
1.)No abuse of power.
2.)No need to always have someone actively monitoring
3.)Plus total fairness between game creator and all.

Negatives(In my opinion)
1.)It's a computer there are bugs
2.)There is not someone actively monitoring
I hate auto-administration. It's too easily abused, especially how you propose it:

2.)Active help system(Report a problem then the program will try different solutions to try to fix it)

This is too hard to do correctly.

Q: "The quest "Retrieval of the Hobgoblin's Crown" is stuck in the second phase. What should I do?"

A: "If you're stuck, you should relog so that you will be respawned in your spawning point."

Q: "No, I'm not stuck, the quest is stuck damnit!"

A: "If you're stuck, you should relog so that you will be respawned in your spawning point."

Q: "..."

A: "Invalid query."

3.)Bully Protection(The game would monitor the kill/death ratio and determine if the user is mob kill happy or just playing.)

Yes, and what if, say, two players are dueling each-other and one player is constantly winning. The other player, pissed off from losing/dying all the time, asks the auto-admin system to ban his opponent, and it gladly obliges because the kill/death ratio between the two is unbalanced. See the problem?
The pros/cons seem to be a bit wrong:

Pros:
1) Technically speaking there is no abuse of power, as a computer is not powerhungry.
2) The game is monitored 24/7.
3) All punishments are equal; different players are not punished differently because of how they appear in the game, although repeated troublemakers may be punished more severely.

Cons:
1) The system could have bugs in it that prevent it from working or make it work too easily. Players may be able to abuse such bugs to try to get others banned.
2) A computer has no emotions, and cannot detect if the players are happy with a certain individual or not. As such, it may ban somebody for mass-killing even if nobody has a problem with this or if this was expected.
3) You cannot reason with a computer, so it can not show any lenience whatsoever.

I think a combination of both may work pretty well: the computer could get rid of players who are for example spamming the server. Spamming the server may prevent human moderators from doing anything, but the computer would still be able to intervene. An automated system against spam and other basic rules can be implemented.

Anything more sophisticated than that, such for instance a "no mass-killing" rule should be moderated by a human.

-- Data
Typically programs aren't good at handling admin issues on their own, and they're worse with GM issues. The former is easy to understand: Much of good administration involves subjective judgment calls and knowing when a person is just getting too unruly. If you have a hard and fast line for punishing any abuse, you'll fail to catch anyone skirting just below that who knows where the boundary is. This is one reason good admins prefer guidelines to rules: Having flexibility gives them the ability to be lenient as needed, and iron-fisted as needed. An auto-admin system is probably still worth using for some things (I question the wisdom of counting a PK ratio), but only for extreme cases so its tolerances should be quite high. In practice they're best at catching rapid-fire spammers, people who admins might not be able to kick out quickly enough to mitigate the problem.

On the GM side of things, a help search or "expert system" is rarely any good, and it's confounded by the problems that 1) some players speak gibberish or don't know how to ask a proper question, 2) you have to anticipate all possible problems, and 3) you have to be able to resolve problems when mere information doesn't suffice.

What you're suggesting is, in principle, not a bad idea for times when your server can't be monitored, but I don't think it's safe to rely on as a permanent babysitter.

Lummox JR
In response to DivineO'peanut
Maybe it would work. Some artificial intelligence datum to monitor everything.

Of course, the artificial intelligence itself would be a problem to make.
In response to Kaiochao
It would not work in the same way that proper monitoring would work. I would only use this in the case of (as I think Lummox said) temporary monitoring while there are no human moderators. The computer would make too many mistakes in the long run for it to properly work.

George Gough
In response to KodeNerd
Thank you for inputs. I do see how having both would make a stronger line of defense then just one alone. Although some things I mentioned I did not quite go into detail with, like I should have. Never the less the idea of having a self monitored system is a good idea. Mixing in human interactions would also be a nice step in the direction. And the FYI the bully monitoring system was more to stop spawn killing..and also the program would be set with many variables in order to determine such things as location, tournaments, dueling, time, and a few other things. But thanks for the input! Have a great day!
In response to Smokymcpot
Smokymcpot wrote:
And the FYI the bully monitoring system was more to stop spawn killing.

There are also other options for dealing with that, like making a player invulnerable for a short period of time after spawning. Spawn killing should be relatively easy to define in terms the computer can understand, hence relatively easy to prevent or punish. You can also have specific spawn points that established players can't just walk back into, and create environmental hazards for players who hang around near a spawn point for too long.

This is actually a much easier problem to solve than server monitoring in general, so you might be better off tackling this one first. The solution you find here might also give you ideas for tackling similar problems.

Lummox JR
In response to Lummox JR
Personally I've used a hybrid(AI + Humans) system in my game for a while, and it works quite well.

While it's rudimentary at best at the moment, I've been devising ways to make it alot better.. At the moment it can only fix people's blackscreens(Everpresent bug that happens randomly due to savefile craziness), reset people's logins to let them recreate their characters, and provide some occasional interaction with players.

To be honest, with a complex enough artificial intelligence, an AI or team of AIs could moderate a game on it's/their own quite well. While Android's post has quite a bit of truth in it, there's some things that don't really impede an AI's work.

Cons
1) The system could have bugs in it that prevent it from working or make it work too easily. Players may be able to abuse such bugs to try to get others banned.

That's why you would constantly improve it to stay a step ahead of any bugs or workarounds.

2) A computer has no emotions, and cannot detect if the players are happy with a certain individual or not. As such, it may ban somebody for mass-killing even if nobody has a problem with this or if this was expected.

That is a problem only with basic AI systems. You could implement a reactions system or a voting-style system to prevent such problems. If you wanted to go more complex, you could even design a system that gives an AI basic "emotions", which would have an interesting effect in the long run.

3) You cannot reason with a computer, so it can not show any lenience whatsoever.

It's not physically impossible at all to code a system that would allow reason.



Of course, I doubt anyone will take this post seriously.

Lummox has a point with the bully monitoring system part.
Things like this work for simple things like spam filters and violations of game rules. But they don't work in place of a game staff, they just make their load lighter.

For example, in one game you have the ability to blow up your team mates if you aim your bombs poorly, so you need to be careful where you're firing them. There's utterly no advantage to killing your team mates, so there's no legitimate reason why anyone would want to do this.

However, unless your team mate happens to be standing where the bomb goes off, they're not going to be killed. So killing team mates happens fairly infrequently. So if someone is racking up kills against their team mates, they either:

1. Are working against their team.
2. Have some personal vendetta against a team mate.

Either one of those reasons isn't good for gameplay, so if someone is team killing too often, they probably have bad motives and the program can detect this and boot that player from the team.

That would be a working example.

But as Lummox pointed out, if players can figure out the limits of what they're able to get away with, they'll be sure to keep below those. So if you can only do 3 team kills in 5 minutes before you get ousted, disruptive players will only do 3 team kills every 5 minutes. That will at least minimize the problem, but its still nice if players can contact the staff to get them ousted instead.