[MUD-Dev] Chances of success (was d20 system)

[Daniel.Harman at barclayscapital.com]

From: Travis Casey [mailto:efindel at earthlink.net]
> On Wednesday 15 August 2001 05:42 pm, Bobby Martin wrote:

>> For example, in Cosm we decided we want the chance of success on
>> a skill to be proportional to the ratio (skill level)/(skill
>> level + difficulty).  This way an increasing skill increases the
>> chance of success, but there is never either a certainty of
>> success or certainty of failure.
 
> Yep.  I suggested exactly that on this list a few years ago.  See:

>   http://www.kanga.nu/archives/MUD-Dev-L/1997Q4/msg00045.php
 
> Shawn Halpenny replied back then that he was already using a
> similar system.

So have any problems come to light with this system after 3 years
usage? I like it although I'm not completely convinced as to how
easy it would be for a person with 50 skill to beat someone with 100
skill. Especially if the fight is longish which would start to
correct any statistical abberation.  For me marginally long fights
are a necessity as quick fights are more likely to preclude tactics
and make healing spells etc. of negligable use in combat.

I do have one issue with the system proposed in the linked post
however, and that is with having the arbitrary hit point cap of
120. I've not thought about it too much, but it seems to me that by
putting a limit in you are setting yourself up for difficulty in the
future. Of course if you don't have it, then as people level, the
chances of David get vanishingly slim against Goliath. This goes
against my instincts though in that I think continual character
improvement is key to this type of game. Of course better armor will
factor into the original hp limited equation too I guess.  Balancing
these multidimensional curves is truly a nightmare. Just because
computers allow you to make complex systems doesn't make us any
better at concieving and understanding the full ramifications of our
design decisions ;)

Another thing I've been thinking off is how to efficiently model dmg
curves for a weapon if I don't want to stick with a linear
probability or standard bell curve. I think it would be nice to
literally drag points on a graph.  You could then have weapons with
a mean dmg of 20 but a potential dmg of 200 every 1 in 200 blows. In
fact if you had weapons that had negative points on the graph, they
could even dmg the user at times (well its a dangerous weapon /shrug
- hell I've smacked myself with a pair of nun-chucks :). With that
weapon that randomly does 200 dmg, you are also helping out the low
level player who wields it - if he gets lucky there, then he could
indeed take out a high level. This all takes me back to the
difficulty of working out all the compound probabilities of success
- especially as I don't want to sledge hammer the problem by capping
things at given levels.

P.S. As an aside I really like it when people link to previous
discussions as there is great info there some of which could be
further explored.  Otherwise we tend to get the problem whereby the
old hands have no interest in covering old ground again and again
for relatively new people like myself. Whilst the new people just go
round in circles :)

Dan


_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev