Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!howland.reston.ans.net!Germany.EU.net!EU.net!sun4nl!sci.kun.nl!cs.kun.nl!manuelp
From: manuelp@cs.kun.nl (Manuel Polling)
Subject: Re: How to detect human patterns ? (more details)
Message-ID: <D5J5JM.H5s@sci.kun.nl>
Sender: news@sci.kun.nl (News owner)
Nntp-Posting-Host: zeus.cs.kun.nl
Organization: University of Nijmegen, The Netherlands
X-Newsreader: NN version 6.4.19
References: <LAVOIE.95Mar15235611@gemini.genie.uottawa.ca>
Date: Thu, 16 Mar 1995 11:06:58 GMT
Lines: 27

In <LAVOIE.95Mar15235611@gemini.genie.uottawa.ca> lavoie@gemini.genie.uottawa.ca (Lavoie Philippe) writes:

8< [cut out other subjects]
>My idea is that if I can categorize a given player has good, I will try 
>tactic A. If he's aggressive than tactic Y would be better, if
>he's patient tactic B he's ok. If he's suicidal then another one should
>be used. My problem is that categorizing someone has being aggressive, or
>patient is something that can only be done over time. There is a limit
>to the amount of information that I can process since I need to move
>every 1/10th of a second. 
8< [cut out other subjects]

I don't know how fast a ship can be destroyed, but I assume that you can get
get the time-factor be less problematic by not changing your tactics every 
1/10th of a second. A human player will have a strategy of her own and only 
change it if he knows beforehand that it will not work (like acting cautious if
the oponent is fleeing as fast as he/she/it can) or this becomes aparent while
fighting.

How to decide in what way the strategy should be changed is something
else of course


>Phil
--

	Regards, Manuel.
