[Bugs] Ship path behavior

Beta ver 0.95.5.0

The ship should take the direct path (red in example) instead of the detourned path.
End of quote




13,824 views 17 replies
Reply #1 Top

Not a bug.  Non explorer ships tend to avoid fog of war areas.

Reply #2 Top

It's the role of the scout to discover new areas, and look at the second image, there is fog of war everywhere.

And, it's the same for an already discovered area, and even if the full area is completely visible (all area discovered and monitored by sensor ships).

Reply #3 Top

Ticket:

#NYZ-135-91466

Reply #4 Top

No, he's right.  The ships always take the long route.

 

Reply #5 Top

I had a mix of auto route choices last night, altho majority were the long way round had several constructors that took more direct routes - including through 'uncharted' space, must be typical engineers!

 

Edit: would be nice to be able to select way points as suggested by someone a while back!

Reply #6 Top

I would just like to point out that it is not "the long way" 

the ship will travel the same amount of hexes left -right and the same amount on an angle in either example. In this case you can tell pythagaras to shut it

the only difference is that in the one example shown the ship moves something like 30 tiles left then 30 tiles up & left. Wheras the suggested example would have the ship move 1 tile left and one tile up & left 30 times

Reply #7 Top

When I set a destination to a ship, it should take the direct path, so it discovers a new area between the origin and the destination (or it takes the direct path to avoid an enemy area, etc.).
If I need to set a smaller path, per example, 8 tiles, then 8 tiles, then 8 tiles, etc., instead of directly select the destination located 100 tiles further, it becomes rapidly really annoying.

In addition, I know it can do it, just look at this screen, the path is just as it should be, direct to the destination. :)

Reply #8 Top

This bug is especiially frustrating when you set up a line of slipstream and the ship zigs and zags out of it. Has this bug been discussed anywhere else? 

Ticket:

#NYZ-135-91466 resolvesd?

Reply #9 Top

Okay first off the ships are taking direct routes to where you are trying to get the red line doesn't go through tiles it skips tiles, ships can not skip tiles, the yellow lines go from tile to tile in the most direct path.  Ships can only move where edges of tiles connect, not where corners of tiles connect.

Reply #10 Top

It's the difference between the pathing algorithm calculating a precise approximation of the direct path or a rough approximation.
The length of the paths is the same but the differences lie in the performance, a precise path probably uses more processing power or memory space or both.
So it's just a question of the optimisation of the pathing algorithm.

/just the humble opinion of an informatics student

Reply #11 Top

Seilore, I am not sure what to make of your reply. It seems like interesting information, but doesnt answer my question. Trying to figure what use I can put your information....

 

Has that ticket number been resolved?

 

 Is there any work being done on it?

 

I understand its a huge game with much ongoing work being done to it, I am just trying to see where things are.

 

Saladman

 

 

Reply #12 Top

Saladman, 

We are always looking to make improvements to the pathing.   Two things -- first the paths are similar in length.  That is, if you count the tiles in screens above, the direct path tiles vs. the one where it seems to swing wide are the same.  The reason that it choose on path vs. the other has more to do with what obstacles the ship thinks may be blocking the destination.    That's not to say there aren't bugs in the system, but that's how's working now.  

Reply #13 Top

Is there any plans to have the path consider slipstreams? I have set up lines of starbases with slipstreams, the ships sometimes path on the slipstream route, sometimes not......

 

Saladman

Reply #14 Top

In order to do that, you would probably need some sort of path-cost database for each tile to determine fastest path.  That would have to be maintained dynamically as starbases appeared, upgraded, or disappeared, even in mid turn.  Even if there is some more clever way of doing it, it will likely be a lot of computing overhead when you pile up factions and starbases.  I would like to see the pathing get more direct, but I do appreciate the improvements that have been made so far.   I think what I would want for your purposes (and mine) is to be able to set rally points along my superspeed space highway and give a ship a list of rally points to follow in the "go to" command.  This kind of thing could also be used to set up patrol routes, even looping ones, which is a different feature, but another thing I would love to have.

Reply #15 Top

Quoting pshaw, reply 12

We are always looking to make improvements to the pathing. Two things -- first the paths are similar in length. That is, if you count the tiles in screens above, the direct path tiles vs. the one where it seems to swing wide are the same. The reason that it choose on path vs. the other has more to do with what obstacles the ship thinks may be blocking the destination.
End of pshaw's quote

 

For a while this pissed me off (well it did but les me say:) bugged me.  pshaw ti 100% correct,  i did a lot of testing (FOW on and off) and in almost all cases the paths were the same or within1.  By paths i mean the one the AI selected and my counting the most direct / shortest path.  The path finder does a good job of avoiding things that would slow you down, like nebula, asteroids, .....    In some but few by my testing it chose to detour when it really did not need to but it did not result in many extra required moves.

 

Bottom Line IMHO SD did a good job with path finding (not that it could not be even better)

Reply #17 Top

Quoting Saladman, reply 16

waypoints, yes that would do it. :-}
End of Saladman's quote

i vote yes to this