Secret shortcut on Korinthos map?

[GliderRecon]

Playing Springing The Trap, I've found that vehicles can sneak through the eastern woods by using a thin strip of 'passable' woods at the very edge of the map. I assume this isn't intended; perhaps a small map-making error?

[MarkShot]

Is this what people mean when they talk about gamey edge hugging tactics?

[Arjuna]

ORIGINAL: GliderRecon

Playing Springing The Trap, I've found that vehicles can sneak through the eastern woods by using a thin strip of 'passable' woods at the very edge of the map. I assume this isn't intended; perhaps a small map-making error?

GliderRecon,

Nice one. Do you have a saved game from just before that incident? If so email it to me at support[at]panthergames[dot]com. Otherwise, can you do a more zoomed out screen dump highlighting the area?

[GliderRecon]

No savegame, but here's a bigger shot. The yellow strip on the right edge (size slightly exagerated for ease of identification) is passable to all vehicles.

I discovered this when I gave a force positioned by the base icon south of the woods an order to attack the canal bridge and left the route unspecified. The AI decided it was going to go in from the north and proceeded to guide units up the edge onto an east-west woodland road.

I suspected a tiny gap between the woods polygon and the map edge (easy done) but checking in Map Maker all seemed correct.

[Arjuna]

Very interesting. I'll check it out next week when I clock back on duty.
 
TT2956 - Maps - Korinthos - Check out ability of mot units to move through woods on eastern edge

[GliderRecon]

Interestingly, it looks like the issue affects most maps. Whatever the terrain on the extreme right edge of a map (woods, 'impassable', or whatever) vehicles can use it. Everything appears normal on other edges.

[Edit - After a few more tests, it appears not every map has the right edge corridor. Elasson, for example, seems fine.]

[simovitch]

ORIGINAL: GliderRecon

I suspected a tiny gap between the woods polygon and the map edge (easy done) but checking in Map Maker all seemed correct.

A I recall, this is a known bug resulting from maps cropped from a larger "master" map. It seemed to be related to the contours getting clipped at the edges, creating an invisible clear path along the edge.

[GliderRecon]

ORIGINAL: simovitch
A I recall, this is a known bug resulting from maps cropped from a larger "master" map. It seemed to be related to the contours getting clipped at the edges, creating an invisible clear path along the edge.

That might be part of the explanation but I don't think it totally explains the phenomenon. If you create a flat map in Map Maker and add forests to all map edges, you still get the corridors
on right-hand and bottom edges.

[simovitch]

Yea, OK I agree. I ran some tests with a similar map having woods around the edges. Here's my take on the bug (just my observation).

CotA movement algorithms will "transition" movement rates between different terrain types. i.e. there is a 100m wide swath between "woods" and "clear" that shows movement rates as an average between the two terrain types. Cool feature.

My tests show that on the right and bottom edge of a map, this identical 100m swath appears in the woods, like it thinks the area beyond the edge is clear terrain. When I right click for the terrain popup along these edges, the % movement rates are the same as clicking in the map near a woods/clear terrain edge.

[GliderRecon]

No idea if this is helpful or not, but the corridors dont exist in places where woods polygons overlap. A 'double-thickness' of woodland keeps vehicles out.

[simovitch]

GliderRecon,

Good find, yes that helps. But did you notice that the woods overlap fixes the corridor on the bottom edge but not the right edge?

[GliderRecon]

ORIGINAL: simovitch
...But did you notice that the woods overlap fixes the corridor on the bottom edge but not the right edge?

On my test map (flat and clear apart from the edge forests) overlap seems to eliminate the corridors on both bottom and right edges.

[simovitch]

Hmm. I still see it on the "show movement table" in MM even though the overlap extends to the right edge.



(edit) This just in: A triple overlap removes the corridor from the right edge!

[GliderRecon]

I think the reason we're seeing slightly different results is that our test maps are of different sizes. I've just created a new test map (clear, flat, but slightly smaller than my first one) and there's a corridor along the bottom but there's no corridor on the right hand side (the forest polygon on that side works exactly as it should.) If map dimensions are indeed a factor then this might explain why Elasson is mysteriously free of corridors.

[Arjuna]

OK I have worked out why it occurs, but I need Paul's help on how to fix it.

The reason this is occuring is that on the right and bottom edges we have a partial grid - only 76 metres of the normal 100m move grid on the right edge and only 40 on the bottom. The code that sets the percentage presence is not factoring this in. Hence even though in reality the woods occupies all of the last grid, it's determined that 76 out of a 100 is not a 100%...silly really.

The fix is to factor in this for the partial grids. Alas it starts to get into low level bitwise maths here and I'm out of my depth. So I have emailed Paul for help.