Path: utzoo!mnetor!perle!kevin
From: kevin@perle.UUCP (Kevin Pickard)
Newsgroups: comp.sources.games.bugs
Subject: How to win at 'bt'
Message-ID: <530@perle.UUCP>
Date: 7 Mar 91 16:01:11 GMT
Organization: Perle Systems Limited  Scarborough, Ontario, Canada
Lines: 45


     To quote from the 'bt' documentation:

>Swamp hexes (' S ') are the armpit of Broken Throne! Troops are very
>unhappy here; the swamp will often gobble one army of the advancing
>force. Swamps have no population, and unless in the player's way, aren't
>worth slogging through. Troops in a swamp are at a 10% DISadvantage,
>both attacking and defending!

     Therefore if you want to win just attack Cities, Towns,
Forests and Mountains from adjoining Swamps.

     Confused?

     I sure was, but it worked, so I  looked  at  the  code.
There  is  a bug in the code where it computes loses in com-
bat.  The probability test is  backwards  resulting  in  all
terrain  biases  being  reversed.  The result is that if you
attack using the worst scenario (ie. from a Swamp to a City)
you  end  up  with a 75% chance of *winning* each individual
army combat.  Not bad  for  a  bunch  of  soldiers  trudging
around in the mud and muck.

     Anyway, the fix is easy.  The following segment of code
is  from  btserver.c (around line 265).  It is shown here as
it should be.  The original code segment  had  the  'unitsa'
and 'unitsb' variables within the if statement reversed.

      while ((unitsa>0) && (unitsb>0)) {
      if (weight>((rand()/32) % 100)) { /* In search of good random #s! */
        unitsb--;
        }
      else {
        unitsa--;
      }
    }

     The above change allows the game to work the way it  is
documented.  Have fun.

--
------------------------------ ~~~~~~~ ---------------------------------------
                              | o   o |     Kevin Pickard
                              |   .   |     UUCP: ...!uunet!mnetor!perle!kevin
--------------------------^^^-----------^^^-----------------------------------