[Freeciv-Dev] Re: (PR#4326) Pathfinding

["Gregory Berkolaiko" <Gregory.Berkolaiko@xxxxxxxxxxxx>]

On Sun, 8 Jun 2003, Raimar Falke wrote:

> > But it seems it could be approximated by 
> > (a) using TM_BEST_TIME normally
> > (b) using the time from TM_WORST_TIME for tie-breaking.
> 
> This seems like a good solution.

But still not what Christian wants.  For example,
-------------
  path 1:
    10% 4 turns
    10% 5 turns
    80% 6 turns

  (true) average time 5.7
  best time 4
  worst time 6  

------------
  path 2:
    100% 5 turns
  
  time 5

------------

A more practical example of the above is (R-road, G-grass, foot soldier 
moving)
1.  RRGRRGRRGRRG
  (true) average time 6.66666
  best time 4
  worst time 8
  capped time (old warmap mode) 5.33333

2.  GGGGG
  time 5


Unfortunately, I really don't see a way of computing the real average time 
in PF.  

Ditch the randomness seems the only way.

G.