[MUD-Dev] UI Issues: Anti-scripting techniques

[Brian Price]

> From:          Adam Wiggins <nightfall at user2.inficad.com>
> [Travis:]
> > Brian Price <blprice at bedford.net> wrote:
<snip practice concept>
> 
> Yup.  This is something that gives a lot of people trouble when they first
> play Arctic.  They assume that by typing sneak or hide over and over, they
> can raise those skills.  This is not true at all, of course.  The best way to
> learn hide is to find a heavily populated area and try to hide for a long time.
> The best way to learn to sneak is to sneak around town, since lots of perceptive
> people are going to have a chance to notice you.  Early on you may never manage
> to sneak or hide very well at all, but you still learn from your failures.
> Later on it becomes damn near impossible to learn your skill by just doing it
> around town, but by successfully hiding against very 'high level' mobiles,
> especially aggressive ones, you can still learn.

I'd be very interested to know how much code complexity and data 
storage requirements the above added to Artic.  It seems to me that 
the above is a relatively heavyweight solution, albiet quite a 
realistic one given the practice premise.

<snip lockpicking example and resultant spam-improvement problem definition>

> >     locking it after each success.  In many places, this would be prevented
> >     if there were patrols in the area or alarms attached to the lock --
> >     realistically, one shouldn't be able to simply sit in the middle of a
> >     military base and work on picking a lock without having to worry about
> >     being discovered.  This leads into the next point...
> > 2.  One shouldn't be allowed to indefinitely repeat the exact same attempt
> >     and keep learning from it.  In the example with the lock, there's only
> >     so much one can learn about lockpicking from one lock -- after one or
> >     two successful attempts, the character should have to move on to another
> >     lock to learn more.

Ok, I'll go way out on a limb here, I don't see any problem 
whatsoever with someone trying to pick a lock multiple times and
eventually succeeding given the following:
    Random events may limit the time they have to pick the lock.
    Failure may have repercussions such as triggering an alarm and/or
    breaking a tool. No experience is gained from the task. No skill
    improvement results from the attempts.

> Arctic keeps track of what your character does.  Learning to lockpick on
> Arctic involves cruising around to every town or fort in the game (there are
> quite a few of both) to try your hand at picking their locks.  This is,
> actually, just an extension of their combat regulartor, which keeps track of
> what NPCs you kill frequently.  You learn very little from fighting against
> NPCs you fight all the time, which prompts players to explore, just in the hopes
> of finding someone new to fight, and thereby getting a chance to learn their
> weapons skills and such.  I rather boggled at how they managed to store
> all this info for their playerbase which hangs out around 15,000 when they
> came up 6 years ago (given technology at the time), but now it seems much less

It sounds as if the Artic regulator approach was required to solve a
game design flaw, namely the lack of an ecology.  Given that such a
mechanism was already 'required' extending it to solve a flaw in skill 
improvement would make sense.

> impressive.  At any rate, it works very well.  Once you realize how things
> work, you have a lot of motivation to explore.  While playing a thief I
> don't remember anything quite as gratifying as locating a new trap that I
> hadn't tried to disarm before, particularly if it was a tough one.

I had a rather long and involved response to this but I'll sum it up 
simply;  in reality very little skill improvement occurs during the 
normal exercise of a skill.  Rapid improvement comes about through 
training.  A fully realistic model would undertake to merge both 
cases.  However, given the long time periods involved for gradual 
improvement through normal exercise of a skill, in the design of a 
playable and implementable game, I'd much prefer to sidestep the 
gradual improvement altogether.

The on-topic relevance is that there is no possibility of 
spamming/automating skill improvement with the training model.
I rather like solutions which require 0 code and 0 data.

> >     The Arduin paper RPG has a wonderful concept which is related to this.
> >     In it, once a character has succeeded or failed at a particular task,
> >     that character doesn't need to roll for it again unless circumstances
> >     change in a way that will lower or raise (respectively) his/her chance
> >     of success.  For example, if a thief fails to pick a lock, he/she will
> >     fail on any subsequent attempt to pick that lock until circumstances
> >     change in his/her favor (e.g., the thief gets better at picking locks,
> >     obtains a better set of lockpicks, or can spend significantly more time
> >     trying).  In general, this only applies to long-term actions, not to
> >     "instant" actions -- for example, it doesn't apply to hit rolls in
> >     combat.
> 
> Very nice - player friendly as well as being a good way to limit scripting.

Yes, but at what cost?  Unfortunately muds are resource limited, some 
muds have to be rebooted every few days to be playable, others 
require greater and greater payments to site providers as they 
outstrip the memory space allotted.  The more data that is tied up in 
skill tracking, the more code that is tied up in patching game flaws, 
the less that is available for objects, mobs, areas, and code that 
makes the game itself better.

Beyond that, I'm not at all certain that such a system is realistic.  
Just because I pick a lock once doesn't mean that the next time I try 
I don't break a pick or fail due to a case of the jitters
. 
> > 3.  Difficulty should play a role -- if something is so easy that it can
> >     be done automatically, or nearly so, the character shouldn't learn
> >     anything by doing it.  To give a few examples, a skilled programmer
> >     won't learn anything by writing yet another implementation of a
> >     bubble sort, calculus teacher isn't going to learn anything about
> >     math while balancing a checkbook, and Bobby Fischer isn't going to
> >     learn anything about chess by playing me.  :-)

The following is all off-topic but I can't let the above pass without 
comment:

A skilled programmer will learn by embarking on a course of 
self-improvement, experimentation, advanced courses, etc.  I doubt 
any learn much about the art of programming from hammering yet 
another customer spec into submission.  A calculus teacher is very 
unlikely to learn much of anything from teaching calculus past his 
probationary period, if said teacher is to advance in his knowledge 
of calculus, he or she would have to attend fewer union meetings and 
embark upon a course of study and/or experimentation.  Where chess is 
concerned, you don't learn much playing the game once you've mastered 
the basics.  You learn by analyzing past games, studying the moves of 
the masters, and experimenting with alternatives to played games at 
various points.

> 
> And this is what we use.  Thus, fighting the same sorts of things over and
> over should eventually cease to help you, since your skill will advance past
> theirs.  Unless they learn as quickly as you...
> 
> > 4.  Lastly, characters should start out reasonably competent in their
> >     skills and the players should know that.  There's much less
> >     incentive for a player to raise a skill which has a 40% chance of
> >     success than one that has a 1% chance of success, and the time spent
> >     also seems less rewarding -- moving from 1% to 10% seems like a huge
> >     difference, but 40% to 49% is less exciting.
> 
> IMO on a scale of 1 to 100, 0 to 20 should mean you suck and actually
> fail most of the time, 20 through 50 should mean you're competent but not
> anything special, 50 through 80 means you're really good, and 80+ means
> you are a true master.  Arctic suffers from this quite a bit.  Anything
> up to an 'average' rating usually means you totally blow, and can barely
> do it at all.  This is annoying because it takes quite a while to get
> any skill up to average, even if you know what you're doing.  I feel that
> it shouldn't take long to get a skill usable (ie, being able to swim or ride
> a bike) but getting from there to being a master (swiming the english channel
> or riding in the tour de france) should be a serious endevour.

I'd much rather see: basic skill level - you no longer require 
training wheels.  Medium skill level - you can reliably shift gears 
and have a good idea when to be in what gear.  Advanced skill level - 
you can pace yourself to cover as much ground as possible in a given 
period of time.  Master skill level - you can specify and make use of 
custom designed bicycles which enhance your performance.

An idea that Traveller advanced was that many skills were 
performable at a minimal level of competence without any training.  A 
bicycle with training wheels makes a fair example.
 
> >     On SWmud, we made a mistake that relates to this -- we show character's
> >     skill levels as a percentage of how much they can learn, as opposed to
> >     a chance of success.  Several things factor into chance of success on
> >     SWmud -- a base chance of success, score in a related ability, whether
> >     the guild which gets the skill is your primary guild, and, lastly, a
> >     learned ability with the skill.  Hence, a character with "0%" listed
> >     for a skill may actually have a 40% chance of success -- and a character
> >     with "100%" listed might only have an 80% chance of success.  "0%" means
> >     only that the character has no learned ability with the skill, and
> > "100%"
> >     means only that the character has reached the maximum score in the
> >     learned portion of the skill.
> 
> Yup.  We ommit the % since it's misleading in this way.  A rating of 30
> is essentially average; most people think of 50 or even 75 as being average
> when you rate it by percentage.

I'm against numeric displays of anytype except where they represent 
something that could be realistically measured on some scale within 
the game.  In most cases I believe they lead directly to power gaming 
attempts and indirectly to scripting/spamming.  If you simply must 
use a gradual improvement % based skill system, I believe that using 
incompetent for 01-10, barely competent for 11-20, etc. would give 
the necessary feedback to the player while doing much to hide any 
gains from spamming or scripting, reducing the incentive as it were.
 
> > >I believe that by carefully examining the game constructs used in a
> > >mud design, one can eliminate much, if not all, of the incentive and
> > >opportunity for scripting.  In my opinion, this results in a better
> > >game.  After all, ask yourself this question: "Why is the player
> > >scripting this process?".  In many cases the answer is exceedingly
> > >obvious: "Because it is mind-numbingly boring and adds nothing to
> > >gameplay."   Granted, pk/combat scripting is another issue and will
> > >likely still require solutions such as presented in the first and
> > >second attack discussions.
> > 
> > I'd submit that something can add to the *game* without adding to
> > *gameplay*.
> > Also, boredom doesn't seem to be a part of it, from my point of view -- I
> > think it would be much more interesting to actually *play* the game instead
> > of mindlessly focusing on simply raising one or two skills as fast as one
> > can.  IMHO, many scripters are simply looking for an advantage -- a way to
> > advance more quickly than the other players can by "playing" either faster
> > or more often.
> 
> Well, speaking for myself, I have scripted many things on muds.  All of
> them were things that I thought were mind-numbingly boring, but that I had
> to do in order to continue my advancement.  IMO a good desing rules out the
> need or desire for scripting.  If there are a given set of actions that
> I need to repeat on a regular basis without any variation, it should be
> automated.  If it is not, I will automate it myself.  It's that simple.
> For example: if creating a healing potion means getting a flask and two herbs
> from my pack, mixing the herbs and enchanting the potion, without variation -
> I will eventually automate it.  (If I have to do it enough.)  If it means
> analyzing the person's wound, deciding which herbs I need, going and finding
> them, grinding them up just so, making a potion for internal wounds or a
> salve for external, applying the concoction in different ways to the wounded
> area, watching the effects on the flesh after the concoction is applied..
> then it's difficult and unproductive for me to automate it, plus I won't want
> to, so I won't.

I fully agree, you've grasped my meaning despite my words :)

Brian Price
                   Brian Price
               <blprice at bedford.net>