[gurps] Failing autopilots and RVO

[Zan Lynx]

Knapp wrote:
> Zan Lynx wrote:
>> For the computer, for any decision there is going to be one option that is
>> "best".  Choosing the non-best option may confuse the opponent, but it still
>> isn't best.  Unless the opponent is expecting the "best" tactic  of course.
>>
>> Sorry, but as a programmer on a government contract are you going to write
>> software that uses random choices to obtain 80% effectiveness in the
>> government requirements testing or are you going to write software that
>> achieves 95% using well chosen almost-always-best tactics?
>>
>> --
>> Zan Lynx
> 
> If you can get 95% without the random factor, then go for it. I mean
> if you are write more that the other guy then you win unless the fates
> are against you. The problem is that in areas where people use things
> like the Monti Carlo you CAN'T get it right without the random
> answers. There are 1,000,000,000,000 or more of answers in many cases.
> The random help to pick one that is good because you can not test all
> the answers. This will not give you the best answer but it will give
> you a good answer but it will never give you the same answer. The more
> random answers you can test the better the answer will be but there
> are limits of time, information and computer power. You pick an answer
> that is better than all the others that you have tried but that does
> not make it the best or the only answer and it will be different each
> time you run it unless your random function sucks.

Ok, I see you are wedded to your Monte Carlo idea.  I won't even argue 
against it, it is an appropriate algorithm for many things.  But it is 
still predictable!

It picks a random possible solution and then tests it and tries to move 
toward a local optimum solution from that starting point.  Those 
equations that define optimum solutions are themselves fixed!

So a program using Monte Carlo simulation will quickly pick some answer, 
but certain solutions will obviously have equations that make them 
*better* than other solutions, and the decision making will favor those. 
  As it should be!

An opponent that has the Monte Carlo decision code in hand or has 
observed it over time will know what the program is likely to do. Every 
copy of the program will have the same probabilities.

Ok, there could be several "personalities" defined for the program, but 
each one will have known quirks.  Or the criteria could be adjusted at 
random, but I hardly think that's going to result in an effective program.

I am a software developer, a "hacker" who loves computer stuff.  I 
*have* poked around in the code of several Quake 3 bots, poker and chess 
opponents, so I do know pretty much what I am talking about.

There are definite limits to randomness.  For example, one Q3 bot used 
randomness to select choices from decision trees, a kind of state 
machine thing.  It had some override code in there to keep it from 
jittering in place.  Deciding to charge forward or retreat at random is 
all fine but if you change your mind too often the effect is of standing 
still in one place.

I still think that I am right of course!  And if I was running a 
Transhuman Space/Space/Cyber game the skill programs would be 
predictable and only LAI and AI would be immune.

But everyone can and will disagree with me anyway, I don't mind too much.
-- 
Zan Lynx
zlynx at acm.org

"Knowledge is Power.  Power Corrupts.  Study Hard.  Be Evil."