0.91 Bug - pathfinding tool shows incorrect distance army can travel on current turn

Do you know at what point this becomes a problem? As in, is there a certain distance that causes it? Or perhaps not calculating terrain properly?

A screenshot or two of that would help.

I am seeing that too. One thing that also happens is that when the projected route leads through a city, the tool does not account for the one free movement and shows the possible destination as one tile shorter than you can actually go.

I have seen this as well ... I have only noticed it for road travel, where it over-estimates how far a unit can travel by road, usually by one tile, sometimes by more.  It is very frustrating.

It's some variable involving road travel and possibly also moving through cities where its calculating that the unit can move one tile further than it actually can. Unfortunately from a bug squashing perspective it's not a very reproducible bug, but I'll see what I can do to pin it down further for you.

Yes, it is tricky to document ... you don't realize the game gave an incorrect estimate for how far a unit can move until after it moved, at which point it is too late to take a screenshot of how far it had estimated you could move before you actually moved.

Good point.

I'd love the ability to easily setup scenarios to test stuff, but as it stands now I'd either need to get heavily into modding (which could invalidate the results) or use an existing game, which doesn't always work out since I don't have access to everything.

Sounds like something for the devs to confirm/deny.

I've seen this in a non-road situation also. It occured with a unit I had split a from a stack with Tireless march.

When this happens on the road situation, if you move your cursor around, it will recalculate and give you the correct distance. Very difficult to reproduce. I've been trying to get a good screenshot for a while now.

I suppose one way to catch it is to use a screen recorder to record a game, but ... honestly, my system is not up to snuff for recording games to video.  I don't have much storage space left.

At long last, I have captured Sasquatch indisputably on camera!

Err, I mean Loch Ness.

Err, I mean aliens.

err ... the movement path bug.  Yeah, I know, this one you won't believe.

... and after right clicking on that 3rd tile (which math indicated it couldn't make it to), indeed this is the result:

As you can see, its movement is still set to the 3rd tile, but it only made it to the 2nd tile despite the mouse-hovering path (before issuing the move command) indicating it could make it to the 3rd tile in one turn.

Great job, Chibiabos!

This is my experience, as well. Move your cursor and re-position it and it will show you being one move short.

Measure twice and cut once, seems to be the answer here.

I've noticed the problem frequently occurring if I move one square, and then let it try to estimate.  It fails to do so.

Thinking on it, especially since I started looking to try to document it, it seems to be a weird rounding error when fractional movement points are used (either because the terrain, such as hills or swamp, consumes one-and-one-half-movement points, or roads which moving upon consumes only one-third or so of a movement point).  It isn't simply rounding up, though ... no matter how small of a fraction remains, you seem to actually be able to move at least one tile (so it should always be rounded up) but for some reason, an extra rounding up seems to occur sometimes, seemingly making half a movement point for some reason rounding up to 2.

Old math joke: 1+1=3, for very large values of 1 and very small values of 3.

This is the same bug I was trying to describe back in v0.80: https://forums.elementalgame.com/443143/

Nice job getting a screenshot!

I've also noticed that if I mouse over a spot, sometimes it will give me the false number first, and then if I move the mouse away and then back, it will give me the corrected number.

Is it possible it gives the first one as a quick estimate and the second as an intentionally calculated one?

I've found my earlier theory about fractional movement points inaccurate, as its overestimating how many tiles can be travelled the first turn even on normal plains with a single movement point cost. :/