Discussion Index

System Design

1996 Topic Archives

Posted by Coman on 09/12/96

- Having read through what seems like several reams of postings from the last few weeks, I have seen quite a lot of discussion over the new skill trees (and filters), economics, new design ideas, and cheating. From the reading, it often seems as if these are believed to be independent issues when, in fact, they are all bound together.

- One of the ideas that Ptah has been trying to get across is that the system has to be balanced - ALL of the system - from environment to advancement to reality to fun to computer load. Each is a trade off. The only true problems that I have seen on this mud are the overall system design and the cheating.

** before I get thumped for criticizing the design, let me say that this is one of the most creative systems in which I have participated (usually as designer, administrator, referee, or coder), liking or not liking the design isnt the point **

- My primary point here is that the system design not only does little to limit cheating, it actually encourages it. My second point is that there may be some relatively simple fixes.

- Designing a good, consistent, overall system is not an easy task. That is why many designers tend to take reality and modify it with a few rules (e.g. allowing for the possibility of magic). Then, the design is modified again to conform to real time limits such as the computation power available (or sys admins time) and cultural limits (in rp games, the idea of advancement came about because western cultures like to KNOW that they are getting somewhere as opposed to just living out their lives - the best statement of this situation was from Dave Arneson who said that success was measured by living, if you are alive, then you are successful). In fact, systems design of a complex system is the hardest job of all. Coding it is trivial when compared to the difficulty of getting the design right before the coding starts (because coding is much less rewarding, it only SEEMS to be the more difficult task). My sub- point here (any time now, Im going to start numbering this thing like a Department of Defense document :) is that when you make a suggestion, try to keep in mind that it has to integrate with the whole design and strengthen that design.

- Now that Ive made that claim, lets see if I can make some suggestions which will integrate well.

- One of the most consistent themes that shows up sooner or later with all of the others in the communications is cheating. It is a prominent enough problem that I cant believe that there is anyone who hasnt witnessed MANY cases of cheating. It affects all of us. It hurts All of us (even those doing the cheating - if you try to exercise by watching a video, you dont gain anything). In fact, the cheating is widespread enough that I believe that people feel that they Have to cheat to keep up. I am separating cheating here from help. Several people have helped me as I played and did so without giving me hugh amounts of equipment. I will repay these characters in kind for their kindness. But, the existence of the amount of cheating which goes on is due to flaws in the system design philosophy.

- There are some serious design flaws in the current system which allow cheating. There are others which encourage cheating. The first is more difficult to overcome than the second. Some of these flaws are : players maintaining a veritable army of near-clones (one player owning many characters); advancement based solely on xp; advancement of stats based on the equipment carried; and an infinite amount of very powerful equipment.

- One primary flaw in this mud is the tendency away from role-playing for a large number of the players. Role-playing enforces its own form of reality, and reality helps to restrict cheating (in real life, jumping off the bridge means you WILL fall DOWN). Allowing players to create huge numbers of characters means that they have no incentive to role-play. If a player was restricted to only a couple of characters within a certain time limit (personally, I favor 1 character per month), that player would tend to conserve their characters a little more, pay more attention to them. A pkiller wouldnt attack quite so readily if he/she knew that losing a character would mean not playing on the mud for a month or two. As it stands now, pkillers will create new characters, borrow or hand off equipment (a decreed rule, but OFTEN broken), and run up a new character 10 or even 15 levels within a matter of a FEW hours. I have seen these things done. Often. And, no, I wont say how. Those who do it already know. Those who dont know dont need to know; it helps to ruin the game.

--A non-pkiller might work a little harder to shape a single character rather than running 8 characters through the same quest over a period of 2 or 3 days.

--Of course, quests are an entirely different problem, but there might be a solution to that as well.

- Related to this problem with role-playing and numbers of characters is the manner in which characters advance. In a real life role-playing system, character advancement is limited by time simply because it takes so much real time to play the game. In a mud, if the advancement is limited only by experience point levels, then advancement can be made as quickly as the character can kill. The result is that characters team up with other clones (and cheat by multi-playing), or get handed equipment from clones (again, illegal), or team up with friends that they have made (and get handed equipment), and go on killing sprees. It should not be difficult (I can say that with great ease as I have never seen the diku code :) to implement time limits to advancement. Basically, right now, the code checks versus and when these match, the character is advanced. The new code could simply continue this check and also run a simple formula based on level and number of hours played (just as a quick example : (floor(1.75 * n), where n is the level to be reached gives about 3 hours to level 2 and 50 hours to level 7; after level 7, add 20 hours per level or + 20 * (n-7) - this isnt THE answer, just an example). This would solve an ENORMOUS number of problems. It would force more thoughtful play on the part of the players, bring about more role-playing, and restrict or halt all of the problems with an advancement based solely on experience points. It would perform those functions with a very small increase in overhead (of course, EVERY change is always just a small increase in overhead :) since total play time is already tracked.

- Stat advancement might be good for correcting problems with choosing the order of stats when starting a character, but that particular problem could be corrected by suggesting stat orders for new players. Older players will have enough knowledge to determine this for themselves. The real effect of affecting stat advancement by the equipment carried is to promote cheating. It gives players just one more reason to seek free equipment just prior to leveling. This form of advancement also gives older players with multiple characters, friends, and clans an enormous advantage over new players. This advantage is somewhat offset by the help which is offered to strangers at virtually every turn, but the better answer would be to restrain stat advancement to fixed values by level and by class (measurable by the chosen skill set in this case).

- Finally (arent you glad, this is almost over :), there is the problem of infinite equipment. In human moderated role-playing games, the total of equipment is determined by a human who is in complete control of the system. That one-point control also allows for the implementation of an economic system (discussed elsewhere). In the mud, equipment generation is infinite. Moreover, equipment can be duplicated. Each of these implementations has players giving away or sharing large amounts of equipment. I would suggest that it would better serve the mud to limit the equipment found on mobs or purchasable through stores to relatively low level equipment (e.g., cut in half all of the stat mods for the current equipment - a one time fix to the database). Then, award truly unique, interesting, and powerful equipment based on unique quests. Such quests could be imm run or could be coded such that they would appear only once every few months. Such equipment could then be truly unique, would have to be unduplicatable, as interesting as it is powerful, basically strung equipment. Cutting the strength of the modifications would not solve the problem, but it might well reduce it.

- It should be obvious that these solutions work best in pairs or as a complete group. Timed advancement will have the greatest impact on solving problems for the least implementation cost while changing the form of stat advancement brings the greatest amount of change to the system philosophy.

- The one thing Ive learned in 23+ years of wargame and role playing design and play is that some players will do ANYthing that they can to subvert the rules. It is impossible to design a complex system which eliminates all cheating but still allows the players creativity in playing. However, there are elements of every system design which promote and encourage cheating, and attempting to legislate it away has only limited effect. In the end, thou SHALT not doesnt work nearly as well as thou CANT.

- Now that Ive said all of that, let me state that I believe Legend to have some of the most creative designers of any mud anywhere. Ptah has been fortunate to have people of such calibre working on the mud. Having flaws does not make the mud unacceptable. I played in Avalon (a pay mud) for awhile until I watched the sys admins allow one of the gods (played by players who advance to that rank) basically gang- rape my wifes character (a strong term, but quite accurate unfortunately). The god froze her VERY low level character while 2 high level characters who were friends of this god stripped the character of possessions, said some amazingly rude things (you can fill this in as necessary), and would have killed her had not some other high level players intervened (one of whom was killed in the process). The sys admins never had the integrity to reply to my letter when I reported the incident to them. After all, gods can do whatever they want. That type of philosophy makes a mud unplayable (especially when you are being charged real money for the insult). The physical design of Avalon was very good. The philosophy was insupportable.

--And, very often, players will go on playing despite flaws in a design. After all, they have spent time and effort in playing and do not want to throw that away. The point is that players are willing to play so it must be all right with them is not a valid judgment. It does not mean that the system cannot be improved.

...Coman (who is not angry, just an enineer, and, to engineers EVERYthing can be improved :)


From: Bulk Wednesday, September 11, 09:30PM

Although there are valid points in your arguement, regarding cheating and all and having friends around to 'hand' equipment, such flexibilty and room to cheat seems to be required to maintain this mud's flexibility in character design and the path each and every character takes. Granted if stats were given by lvl advancements that it'd be much harder to cheat, but then each character, especially first timers would be stuck having chosen the wrong fighting stats, etc. By leaving it up to the item, there always is enough room to make mid-way corrections, though that may change with the new skill trees. As far as your saying that xperience is the only thing that matters regarding lvl advancement, it maybe nice to complement it with hours played, but to have hour restrictions would only mean ppl running bots or going afk all the time, making this a rather lifeless and dull mud.

Regarding ppl that can advance in a few hours--I personally think, unless they are being manufactured by others, it is a display of skill, or at least a display of your ability to make loyal friends. I haven't seen many that multiplay using their own clones to kill stuff, as grouping xp in this mud is seriously screwed up :), IMHO. But there are times when it seems that a character is passing eq to another of his characters, or does so through a mediator...and all I can say about that matter is that some ppl, regardless of the environment, will cheat, or will try to cheat...

Bulk. Lagged so horribly he decided to write.

From: Diet Thursday, September 12, 09:28AM

I agree with Bulk. Muds are ABOUT getting XP. If level was based on time played then whats the point in killing mobs? DUH! I know you obviously put A LOT of time into your post, but this is the way Legend is and has been. I think skill trees will enhance the way Legend is. If you really dont like the way the game is designed, there are HUNDREDS of other Muds out there. Or, maybe even start your own mud..


From: Ptah Thursday, September 12, 11:57AM

A wonderful and fascinating post on Coman's part. I often feel as fortunate in our players as I do in our staff. :)

There's much meat to respond and discuss here, and really, we ought to take it to the Legend Discussion List, but that's down. :( So I will do my best to respond concisely here.

- issues bound together, balanced system:

Absolutely. I can think of a few more problems than merely "overall system design" and "cheating" however. The issues which Coman identified as problems with the game system design were largely bound up within the current system, whereas I have become convinced over the last few years that the central problem with the design of all muds are these: their static nature and the requirement for maintenance. In most ways, I see everything he listed under system design problems as ones inherent to the static nature of the system, as I will note below.

- the issue of game advancement:

Should I go to Bruno Bettelheim, or to the more mundane Greg Costikyan? :) Either way, the issue of game advancement relates back to the definition of what is a game, what is a toy, what is a competition and what is merely play. There is little question that at least our current playerbase would prefer that the mud remain a game, and not a toy; competitive and not merely play. My personal inclination is to provide space for both types of players to participate, of course. I do not think that we can remove goal-oriented play altogether, but as Costikyan points out, in RPGs where there is no goal, the players make one. But a mud is too static to permit that to any varied degree, really. I m sure that our playerbase would disappear immediately if we failed to supply the goal. As was recently pointed out on the mud newsgroups in regards to a particularly dewy-eyed view of level-less systems--they've been around for a long time, and level systems are still preferred by the majority of gamers. Goals are what make games, and levels are the goal. Yes, alternative goals can be supplied, but something with the elegance and simplicity and ease of implementation of levels--well, that's harder.

- I fully sympathize with the multiple characters issue. Unfortunately, it does lead into a couple of traps. First, it is virtually impossible to enforce any sort of restriction on multiple characters. Even if we went so far as email address verification I personally would still be able to play three or four different characters. Probably more if I really wanted to. There are also various delays that are implied, or at the least, increased workloads, from things such as verifying addresses to character approval times, etc. I strongly suspect that the enforced delay of one new character per month would simply be unsupportable. As it is, I will not play Armageddon, which I regard as one of the best-designed muds on the Internet, merely because it takes three days to get a character approved. If I had to wait a month, I'd find a different mud. :) Now, this is not to say that I don't understand and even agree with most everything you say about the problems with having multiple characters; I tend to simply regard it as part of the territory with muds.

There are some measures we can take to reduce the likelihood of abuse, including installing use of the identd protocol on the mud server, to limit multiple connections from a single account and give us better warning of the worst abuses. Likewise, an automatic notification system for things like repetitive switching between same-site characters, multiple same-site logins, etc, would be a great step forward. But they will still (and always) be merely fingers in the dam, unless you go to complete character approval processes.

- Similarly to the month-wait idea, I think that restricting advancement based on time is not a wise choice, for the simple reason that it greatly restricts the venues for showing superior ability at the game. While Bulk's response of "better able to make friends" points up a drastic flaw, it also emphasizes that people DO measure themselves against others, and without a yardstick, they will go someplace where they still can measure. A mud rid of its competitive element is a mud that tends to atrophy (cf Bartle's remarks on mud playerbase makeup) and although LegendMUD currently suffers from a surfeit of this type of player, it is precisely because of that that we cannot radically alter the system to prevent them from pursuing that goal.

A common solution used on other muds is to replace the single XP scale with dual or triple scales. To reach level 2, you must accumulate X quantity of kills, Y quantity of quests, and Z quantity o rooms explored (with X Y and Z of course being multiple points per achievement). I rarely see a player who is not disgruntled by this, as it confines three separate methods of play (and achievement) into one scale. I'd also note that these scales typically leave out the other forms of mud achievement which deserve recognition, which is to say the social aspect and the roleplay aspect.

We currently offer essentially token rewards in terms of XP for questing/puzzlesolving, and for exploration. Unfortunately, the server really cannot handle the load for a good exploration system (i.e. one that cannot be easily macroed) and the lack of dynamic quests (gosh, there's that static/dynamic thing again) means that rewards for quests must always be merely token ones on the classic XP scale, though they may provide secondary benefits like items that permit faster advancement using the traditional hack n slash method.

One fillip that is often tossed into this is experience for appropriate skill usage--eg healers gain not from killing but from healing. In a classless system, this becomes extremely difficult to implement with any degree of accuracy or fairness, which is really a crying shame. :) In a levelless system, a system where skills increment by usage, a character's viability and ability in a single skill set becomes a very very nice yardstick to measure against, but that requires the entire database and game system to be designed from the ground up with that concept in mind. I don't think we're up for that on a game system that has been in use for two and a half years now. :)

- a brief note--total play time is tracked, yes, but we do actually alter it with the youth spell. :) Age is merely total time played divided by an arbitrary factor.

- infinite equipment... ah, well, I think my thoughts on this have been pretty well expressed in earlier posts. Certainly equipment duplication is a problem, but really it is dwarfed by the infinite resources supplied by the mud. While I dislike dt's as a concept, they are still here mostly because they eat equipment. We would, of course, need a hundred times more dt's to make any sort of dent in the problem. :P

What we really need is to curtail the infinite influx of items. This requires solving the problem of mob competitiveness with players, if resets do not supply their weapons and armor. It also requires solving the problem of a general economy, of course; then again, no economy is really possible without true limitations and a closed system.

Unique quests are simply too much load. Many muds try it, and the quests are always lowest common denominator things. A major LegendMUD quest can take a month to develop, and given that they are often the size and scope of early computer adventure games in and of themselves, it is not surprising that we cannot support the load required to create that many of the quality we require. What we need is a dynamic system where interesting developments happen of themselves. This would entail a major, major, major rewriting of mobile AI, which is not as daunting as it seems, since the theory behind making such a system is well-understood (cf Meier, Crawford, and various recent newsgroup posters). No, the real problem is the fact that it invalidates our current database, since a dynamic system or an artificial life system really tend to preclude the extent of special-cased plotlines that we have supplied. And one is always hesitant to wipe out the game as one knows it...

- I am very surprised at your experience on Avalon, as Dave Austin has always seemed to me to be a professional and capable designer, though I dislike many of the technical decisions he made in that server design (for example, the automatic concatenation of weather and light levels into room decriptions drove me batty given that it was not backed up by a decent landmark system and thus felt like a half-assed job of server-generated descriptions). On the other hand, the scripting and interactivity in Avalon is truly superb, and I have not encountered anything quite like it in muddom. In any case, you have to pay for Avalon, so skip it for that reason alone. ;)

- not only will players go on playing despite the flaws in a system--they will go on playing BECAUSE of them. Rectifying some of these flaws in the past has led to players leaving Legend because they simply dislike change, because they felt that they could not compete, or because the flaw was their sole crutch for competition. In a certain way this is healthy for the playerbase as a whole, but too much shock to the cultural fabric of a mud can be quite devastating. It is a sad fact that the current state of player tactics (by which I mean the extreme emphasis on certain repetitive methods of exploiting flaws large and small, the willingness to give up the solutions for puzzles, etc etc and ad nauseum) is by and large the legacy of a set of players who are mostly all gone, and who never took it nearly as far as their descendants do. Now, I said it before, and I will say it again--we are nonetheless extremnely fortunate in our players, in that they have an awareness of these issues that is clearly evidenced in the popularity of these threads on the board. It doesn't take much to see the difference between the responsiveness and thoughtfulness and sheer concern for the game as a whole on Legend compared to other muds--or even to the "if you put in skill trees, i am leaving" attitude that has sadly but unsurprisingly manifested upon occasion). To put it more briefly, I am very glad posts like yours and threads like these show up on the boards, because it shows that players here do care, which makes the imms all themore willing to keep wrestling with the thorny problems. With every generation of new players, the problems become more apparent, because exploitation of the flaws is more ingrained, and we've reached a point where some radical ideas and redesigns are going to required in order to make the game viable in its current form for much longer than a year or so.



1996 Topic Index