6839 Pathing

[PeterStep]

M113 are amphibious but curious path choice

Screenshot 2023-02-03 151353.png (513.69 KiB) Viewed 1124 times

in 'Cross Checking' when there is a usable bridge close by. Do you want the save file?

[CapnDarwin]

The unit decided it was faster to swim than deal with the slow down at the bridge. If this was your unit, you can adjust its waypoints to keep it on the road. I will pass this along to William to see if he has any other thoughts for the AI.

[PeterStep]

Looking back at the pathing it is 3rd in line behind the two tank units which get across quickly. The jam does not occur until the two inf units following get onto the bridge.

Screenshot 2023-02-03 170258.png (328.18 KiB) Viewed 1104 times

Waypoint 1 reached at 0340 via river and WP 2 at 0358 or if I pull the waypoint one hex SE it routes via road and WP1 is 0323 and WP2 0353. So it opts for a tactically risky punt down the river in open view rather than a covered run via the forest road to get to WP2 faster. I can see the logic, but....and yes I should have noticed it earlier.

[PeterStep]

When I said faster I meant slower!

[CapnDarwin]

I have passed this on to William to review.

[PeterStep]

Did a bit of playtesting today - it is possible to get the mech inf unit to move hasty in the river about 5km and come ashore right on the southern VP location. Takes and returns some fire on the way but losses from tank fire 2 hexes away on land much less than I would expect. Looks like an unexpected consequence of amphib capable units on large rivers.

[pullg]

With all that in mind, is there some movement logic that needs modification? Amphibious crossings, even with amphibious vehicles, aren't spur of the moment operations. That goes double for big European rivers flowing through urban or developed areas. River banks there are often steep (or lined with erosion-prevention measures), so engineering preparation is often needed for entry and exit, and river currents of 2-3 kph are a significant factor for APCs and scout vehicles that manage 5-6 kph in water at best. That's not even considering the vulnerability of crossing vehicles to enemy fire or simply capsizing. It better be a pretty significant traffic jam to make a trip on the river more attractive!

[WildCatNL]

Did a bit of playtesting today - it is possible to get the mech inf unit to move hasty in the river about 5km and come ashore right on the southern VP location. Takes and returns some fire on the way but losses from tank fire 2 hexes away on land much less than I would expect. Looks like an unexpected consequence of amphib capable units on large rivers.

Peter, would you have that test scenario available for me? It would save me the time to set-up the same case.

With all that in mind, is there some movement logic that needs modification? Amphibious crossings, even with amphibious vehicles, aren't spur of the moment operations. That goes double for big European rivers flowing through urban or developed areas. River banks there are often steep (or lined with erosion-prevention measures), so engineering preparation is often needed for entry and exit, and river currents of 2-3 kph are a significant factor for APCs and scout vehicles that manage 5-6 kph in water at best. That's not even considering the vulnerability of crossing vehicles to enemy fire or simply capsizing. It better be a pretty significant traffic jam to make a trip on the river more attractive!

That's why we want to check this.
The game should model penalties for entering the water (i.e. preparing and checking the vehicle to be amphibious), and has slower movement in water.
We've taken some liberty with modeling river banks (as part of the map creation) for two reasons: (1) information on river bank composition and steepness is really hard to obtain, in contrast to elevation data and satellite imagery, and (2) we assume the 200+m river bank has one or more nice 'fordable' exits.

What also plays a role here is that units also take penalties crossing bridges - they'd have to line up in column formation, cross the bridge section in turn at reduced speed, etc.

This feels like an oddity in the weights being used to drive the pathfinding, not so much the movement logic. We'll find out.

[PeterStep]

William, save files attached as requested.

Cross Checking.sav.zip

(477.4 KiB) Downloaded 15 times

Had not saved he original playtesting so recreated with similar parameters - mech inf still reaches VP via river route

[WildCatNL]

Peter, thanks, really appreciate the extra effort to recreate it the situation.

[PeterStep]

William, don't know whether you have had chance to look at this but started playing it from scratch again and there are more strange pathing choices appearing.

Screenshot 2023-03-01 104453.png (656.54 KiB) Viewed 934 times

[CapnDarwin]

Peter,

I let William know about this. Thanks for posting. We will evaluate and hopefully fix this issue.