Re: Some crunching of simple contest probabilities

From: Aki Martti <marduk_at_...>
Date: Sat, 01 Nov 2003 13:32:23 -0000


Bryan:
> Well, OK, something to keep in mind if you are freakishly
> obsessed with numbers the way I am :)

Well, maybe not freakishly, but somewhat :)

One thing that you didn't bring up was the probabilities of winning big when equal skill levels are involved. I haven't really been that active on this list, so I'm not sure if this subject has been brought up before... If so, could someone please point me to the appropriate thread? Thanks.

Anyways... if both participants have a skill level of 20, the chances of causing an impairment or a more serious result (to either participant) during a simple contest are a less than 10%, I think.

When both skill levels increase to 10w, the chances of an impairment or a more serious result increase to ~50%. Quite an increase, I say.

Of course, the probabilities lessen again when both skill levels reach 20w and increase again when they reach 10w2. The probabilities follow almost exactly the function sin(x), where x is the skill level and the wave length equals one mastery, and the possible values range from ~10%..~50%. These are not accurate result of any calculations, just some evaluation off the top of my head.

One could just ignore these results as statistical glitches, but 10%->50% is not a glitch in by book, it's something more. One could also speculate that fighting "really" becomes more deadly at a certain level, but becomes less deadly again when both participants become more skilled at defending themselves. I'm not sure about the fluctuation of deadliness when the skill levels keep on increasing, though.

This really hasn't been a problem in our games, but since we're doing math... :) Any ideas on how to explain the above?

Powered by hypermail