Top Mud Sites Forum - View Single Post

Spazmatic · 03-21-2006, 07:51 PM

Maybe I'm reading Valg's post a little differently here, but it seems what he's saying is quite reasonable. An overly simplistic system is often too easily gamed or too freakin' boring. On the other hand, an overly obfuscated system often fails to reward behavior that should be rewarded.

This doesn't mean a system is wrong just because it can be scripted - anything can, theoretically, be scripted. So, I don't think your implication is fair with regard to Valg's post at all.

I'm going to point further down, into my response to Fern, for my response to this example.

That is all very good (and from a design perspective, perhaps all that is required). Also, let me note right now, I'm very happy that another mud has added RPXP to its system, and I imagine your system is very good. I am not trying to criticize it, and I don't think Valg is either - I just want to know how good it is, since it would take hours of experimentation to discover its quality on my own. If it can avoid certain, traditional problem areas, then why not less other developers know so we'll come play and discover?

Also, you noted appreciation of Valg's willingness to take the discussion off-board. However, I think, more generally, people would like to know whether it passes the two tests he cited (and any similar types of abuses). If your system can discriminate basic scripts from real roleplaying, there would probably be quite a few people knocking on your door to have an off-thread off-forum chat (though, admittedly, any sensible admin would probably go into hiding at that point, to avoid the time-sucking vortex that would result).

Anyways, to the point at hand. At least in the cited post above, you only address coverage, while Valg and myself focus on discrimination. That is, you're trying to make sure roleplaying is encouraged and rewarded by your system (at least in the above cited post), and we're wondering how well it keeps out the naughty types.

Yes, coverage is very important. However, it's also very easy. At the most extreme point, I could give experience per minute logged in at a fixed rate, so that socializing is encouraged, as is roleplaying and any other activity you can think of. I'm not necessarily discriminating, say, roleplaying compared to idling, but I guarantee I'm covering it.

At the other far extreme, you could have a system that could identify good roleplaying by comparing it to examples of good roleplaying, but performs the comparison in such a narrow way as to be absurd, and requires all emotes to be the same. That would be, well, idiotic.

This is a tradeoff that's common in a lot of fields. In statistics, a comparable issue is the bias-variance tradeoff (as an example). Finding a good solution somewhere in the middle is hard, so we like to ask.

the_logos' analogy of rewarding low-risk combat is a question of discrimination, however, I don't think it's the proper analogy to bad roleplay. Rewarding bad (or scripted) roleplay would be comparable to rewarding a level 60 character killing level 1 rats. Rewarding bad roleplay equally as well as you reward good roleplay would be comparable to rewarding a level 60 character with the same experience, no matter what he or she kills.

Very few muds use the latter system.

I realize this is a bit of a promotions thread, and we're rather OT. However, such things happen on the web, and I wouldn't have posted if I didn't think I could maybe either clarify things or have things clarified myself. I also respect your need to keep your methods secret, and if you really insist on that, fine.

However, let me reiterate - answering Valg's question about whether or not those two methods work will not compromise your security, and has the potential (if your answer is "no, they don't") to generate a lot of chatter and benefit the community as a whole. I can, for example, tell you how much computational power is required to break a method of encryption without revealing to you the method of encryption. This is a similar request - we'd like to know how good it is from a developer's standpoint, knowing full well that it's probably very good from a player's standpoint already. Consider it an exercise in the possible.

(Plus, seriously, if those two exploits DO work, then your players will be exploiting them like crazy in the near future.)

03-21-2006, 07:51 PM	#19
Spazmatic Member Join Date: Mar 2003 Posts: 103	Maybe I'm reading Valg's post a little differently here, but it seems what he's saying is quite reasonable. An overly simplistic system is often too easily gamed or too freakin' boring. On the other hand, an overly obfuscated system often fails to reward behavior that should be rewarded. This doesn't mean a system is wrong just because it can be scripted - anything can, theoretically, be scripted. So, I don't think your implication is fair with regard to Valg's post at all. I'm going to point further down, into my response to Fern, for my response to this example. That is all very good (and from a design perspective, perhaps all that is required). Also, let me note right now, I'm very happy that another mud has added RPXP to its system, and I imagine your system is very good. I am not trying to criticize it, and I don't think Valg is either - I just want to know how good it is, since it would take hours of experimentation to discover its quality on my own. If it can avoid certain, traditional problem areas, then why not less other developers know so we'll come play and discover? Also, you noted appreciation of Valg's willingness to take the discussion off-board. However, I think, more generally, people would like to know whether it passes the two tests he cited (and any similar types of abuses). If your system can discriminate basic scripts from real roleplaying, there would probably be quite a few people knocking on your door to have an off-thread off-forum chat (though, admittedly, any sensible admin would probably go into hiding at that point, to avoid the time-sucking vortex that would result). Anyways, to the point at hand. At least in the cited post above, you only address coverage, while Valg and myself focus on discrimination. That is, you're trying to make sure roleplaying is encouraged and rewarded by your system (at least in the above cited post), and we're wondering how well it keeps out the naughty types. Yes, coverage is very important. However, it's also very easy. At the most extreme point, I could give experience per minute logged in at a fixed rate, so that socializing is encouraged, as is roleplaying and any other activity you can think of. I'm not necessarily discriminating, say, roleplaying compared to idling, but I guarantee I'm covering it. At the other far extreme, you could have a system that could identify good roleplaying by comparing it to examples of good roleplaying, but performs the comparison in such a narrow way as to be absurd, and requires all emotes to be the same. That would be, well, idiotic. This is a tradeoff that's common in a lot of fields. In statistics, a comparable issue is the bias-variance tradeoff (as an example). Finding a good solution somewhere in the middle is hard, so we like to ask. the_logos' analogy of rewarding low-risk combat is a question of discrimination, however, I don't think it's the proper analogy to bad roleplay. Rewarding bad (or scripted) roleplay would be comparable to rewarding a level 60 character killing level 1 rats. Rewarding bad roleplay equally as well as you reward good roleplay would be comparable to rewarding a level 60 character with the same experience, no matter what he or she kills. Very few muds use the latter system. I realize this is a bit of a promotions thread, and we're rather OT. However, such things happen on the web, and I wouldn't have posted if I didn't think I could maybe either clarify things or have things clarified myself. I also respect your need to keep your methods secret, and if you really insist on that, fine. However, let me reiterate - answering Valg's question about whether or not those two methods work will not compromise your security, and has the potential (if your answer is "no, they don't") to generate a lot of chatter and benefit the community as a whole. I can, for example, tell you how much computational power is required to break a method of encryption without revealing to you the method of encryption. This is a similar request - we'd like to know how good it is from a developer's standpoint, knowing full well that it's probably very good from a player's standpoint already. Consider it an exercise in the possible. (Plus, seriously, if those two exploits DO work, then your players will be exploiting them like crazy in the near future.)