[MUD-Dev2] [DESIGN] Removing the almighty experience point...

Aurel Mihai aurel.gets.mail at gmail.com
Fri Aug 31 10:59:24 CEST 2007


John and Damion both brought up a lot of good points a short while
ago.. Levels and experience in gaming are unrealistic because there
really aren't any numbers in nature. There are patterns and there are
fuzzy borders and we can count individual objects, but that's about
it. The only reward that I get in real life for killing an orc or a
bunny is the personal experience of having done it and a memory of how
I did it that particular time. That says nothing about numbers and it
says nothing about how I'll go about doing a similar task in the
future, strictly speaking. If I want to _practice_ a skill, such as
the killing of orcs or bunnies, I not only have to repeat the task
many times, but I have to actively alter the way I perform that task.
Practice doesn't make perfect, it makes permanent. To get perfect, I'd
have to actively work at it. In the end, my proficiency at a
particular task isn't a number or a rating, but rather a combination
of muscle memory and tactics that I've learned.

Translating this into computer gaming is tough. Numbers work great
because computers easily work with numbers so algorithms can be
developed and the entire process can be automated instead of a human
GM watching over every newbie fighting a rat. And as Damion mentioned,
there are plenty of ways in which developers have managed to use
experience and levels to their advantage. RPGs, even if they've lost
the strict RP aspect, remain a fun experience. Question is, do we want
more fun or more realism? Are they the same thing? Can we combine the
two better.. perhaps by inventing a new system that does away with all
the unrealistic numbers?

It's a pretty tall order. A good start might be the ELO system
mentioned earlier, not just on mobs and characters, but on everything
in a game: skills, items (difficulty class..), mobs, characters, etc.
On day one everything can start at 1000. New characters can also start
at 1000 for everything (+/- some standard deviation).

Here's a scenario. I walk up to a locked door a year after the game
has begun with my newly created character. The lock has a difficulty
rating of 1100 and a strength rating of 2500. The door itself has a
strength rating of 1500. How is this decided? Well, the lock has been
successfully picked by most players, but it hasn't been broken by many
at all. However, the door has been bashed in more often than the door,
though not as often as the lock has been picked. I don't see any of
these ratings, but I do get a nice description of the door as being a
solid oak door that's seen better days with a crude, but strong lock
made of iron. This can be visual or text. I decide to try bashing the
door in... lightly. The door barely rattles in its hinges. Its
strength rating goes up to 1510. (Or there could be a ceiling at
1500..) My skill at bashing in doors goes to 995 because I feel
slightly demoralized. These numbers are all internal. To me, the
character, I just feel a little bummed that I couldn't break down the
door and I don't even need a message to that effect. It's a natural
feeling and I'll be less likely to break down future doors. Sorry for
being so verbose, but I hope this conveys what I'm trying to get
across adequately. So anyway I go to pick the lock with a nail I find
on the ground. The nail gives me +200 to my lockpicking rating
(internal number) so it's a relatively easy conquest, though one that
can still be lost. By the way, why is the lockpicking rating on the
lock at 1100 rather than 1000? Because without some sort of tool, it's
going to be very hard to pick even a simple lock. Anyway you get the
idea at this point. Every single act in the game can have a rating
assigned to it.

I suppose this could bring up two questions: how is this different
from experience/levels and why is it better. It's different because
the user doesn't see the numbers. It's different because there's
always the potential for loss and each action ought to be reviewed
more or less carefully so that the situation is just right before it's
attempted. On the other hand, it doesn't stifle risky behavior since
you could just as easily go nuts and get lucky every once in a great
while. It's different because it's intuitive and situational, but
requires some thought and even *gasp* a tiny bit of roleplay in the
form of reading descriptions/messages or just carefully observing
one's surroundings rather than mindlessly mashing buttons and clicking
on enemies to dispatch them. By situational, I mean that each skill is
its own rating. Killing orcs doesn't help me learn how to pick locks.

As for why it's better, one of the bigger criticisms of the ELO system
in chess and other such games is that it creates more risk for
advanced players when playing lower ranked players. A loss at a high
level means a greater drop in rank. It tends to even out ranks in that
way so that if two players compete many times, they'll probably end up
at the same level. Does this sound like something that could be useful
in RPGs? A system that promotes challenging yourself and discourages
farming skills... You might ask, how will anyone ever advance? Real
player skills, picking the right battles, using the right tools, and a
little bit of luck. Just like real life. :) Because it relies on
numbers in the background, it can also be used to create an algorithm
and the simple math involved in calculating rank will be no more
difficult than whatever system is currently in place in any given
game.

Conclusion: ELO system for all aspects of game. No levels, no
experience, no numbers that the user can see. Highly intuitive (can't
pick lock with my finger, but there's a shiny, bent nail on the
ground..). Slightly tricky without being a burden for those that don't
like puzzles (just bring overwhelming force). Unlimited potential for
advancement for those that want to be the very best. Great potential
for everyone else because of a relatively level playing field. Thanks
for reading all that!



More information about the mud-dev2-archive mailing list