Pathfinding

[GoodGuy]

I've got a few questions regarding pathfinding. There are many cases where the pathfinding can be a real pain in the player's butt, since it does some strange things. I'll post a couple of examples to show what I'm talking about. I figured some issues might be caused by flaws in the scenario design, but most are caused by irregularities in the pathfinding process/rules.

Example: Kalpaki - Charge of the Centaurs.

You can see a major spawnpoint on the top right corner. The spawning Inf Bns don't have any problems with moving to the west or to the south, they'll use the tracks nearby. But the motorized units have major problems (partially because of the design of the woods/track nearby), and it can take days to reach Pogoniani (the first village in the west) or Palaiopirgos (the second village to the south, south river bank), if you don't watch every step of those motorized units - you even might have to add a sh*tload of waypoints (pls check the 2nd screenshot) to get 'em to the target location:

Ok, the "quickest" path won't get us there (this path could take 1-2 days at least, even longer with tanks):

[GoodGuy]

Hmm, ok, the units can't cross the river near the woods....ok, but quick/undefined path to the south gives another bad result, as the columns keep leaving the track, even AFTER passing the river on its right side. The only solution is a set of waypoints. But even a move-task with like 8 or 10 waypoints won't work. The only solution is to use an attack-task, in order to keep the column right on the track, before reaching the reorg.-point.

[GoodGuy]

Another example: I want to move an Arty unit. Let's say I'd play at fast speed and with order delay, this would really cause trouble, if I'd issue the order without checking the result, as the quickest path won't keep the unit on the road-network:

[GoodGuy]

That's interesting, "undefined" path would do it here.

[GoodGuy]

Another example: Inf troops moving next to a track, but refusing to choose the track as quick route, although the movement modifiers for tracks and roads, escpecially with regard to fatigue, create a way more friendly environment than the "broken" terrain.

[GoodGuy]

Once again, selecting undefined brings up a better solution (for Inf units, at least):

[GoodGuy]

Next example: the "river-issue":
I tried to get the HQ, shown on the next screenshot, to the ferry. While I'm happy that crossing rivers using ferries takes way less time in COTA, it's a real pain trying to get the troops there, in this case. The HQ won't cross the small river river delta for one reason: There are 2 motorized units (AT unit and another one), that aren't able to wade through the river.

-snippet- I don't know how deep it is here, but I'd be happy if you could review some of the unit abilities for the next installment, Germany had tanks later on, which could be equipped with snorkels, in order to wade through rivers (around 2 m deep, if I remember correctly). -/snippet-

Anyway, I made another attempt, after detaching the cumbersome mot. units, and even then it took a few attempts to get 'em to use the ferry. I had to detach 2 units to get things moving, it kinda felt like I had to cajole the Bn into crossing the Delta and using the ferry, plus I had to "insist" (by issuing the "short" route order several times) that things get done.

This can be a real pain if playing with painfully realistic delay. !!!

[GoodGuy]

Another issue: The Arty got order (from 23 HQ) to move right into a scratch force of enemies trying to regain control of the subjective (south of Limni/Gormos River). There's almost no way to avoid that the arty stumbles over these enemy units, if you're playing with painfully realistic order delay, since it will stay on this "road of ruin". The only solution is to detach it and hope a "rest" order will save it.
I had hoped that the "arty-code" had been reviewed, since bases seem to be positioned correctly/carefully.

[GoodGuy]

Last screenshot: Sorry for the fuzzy image, I didn't have time to reinstall all my DTP tools (new puter), so pls bear with my "MSPaint-ology" for now.

The "Green 1"-area should be re-designed and included in a patch, since - as described above - it's causing major troubles.

Another example for strange pathfinding: The highlighted mortar unit insists on moving through broken terrain, then across 2 rivers and finally through a large wood (green 2), although it could access a track nearby, which leads to a road providing the quickest way (red line). The white lines highlight tracks in that area.

Can we expect this issue being addressed prior to the release of the next installment (means FIXED in the next installment)?
The pathfinding wasn't that much of an issue in HttR, but with modifiers for terrains like woods or broken being changed significantly in COTA, it starts to be an issue that takes away some of the fun. Also, can we expect arty-movement to be reviewed as well?

I understand that engines have to be "recycled" (means used for several installments of a series or franchise), especially in this niche-genre, but correcting these issues would make me (and I guess others) really happy. I still appreciate that COTA is a major step forward in the process of creating a perfect wargame.

[Arjuna]

GoodGuy,
 
It's good to see you getting stuck in here. Ideally it would be good to have saved games to check out the exact reasons for the behaviour in question. But I'm supposed to be still resting and so I won't be looking at these for another week. But in the meantime, here are some comments:
 

• Tracks provide no benefit to motorised units in the Greek scenarios. Right click on them to see the motorised movement effect.
• A force moving in non-road column formation will end up with subGroups moving off the main line of march. So even if the main line of march is along a road they will be off the road and slowing everyone down.
• Ferry's/Fords take a long...time to cross. It's often quicker to go around them. If you want to force them to use the ferry, use a Shortest path.

OK got to go. My darling wife thinks I shouldn't be here. [;)]
 
 

[GoodGuy]

ORIGINAL: Arjuna

• Tracks provide no benefit to motorised units in the Greek scenarios. Right click on them to see the motorised movement effect.

Sure, I understand. But the spawn-point (top right corner, map: Charge of the Centaurs) can be a real pain. Also, I've replayed the map several times - sometimes the pathfinding works fine - while it fails to come up with correct solutions on other occasions.
Also, regarding the non-existent benefit, that's why I use to keep away my motorized units from such areas, it's just a pain that the START-point of the tank/mot. units is in the worst terrain half of the time. (note: these units use to spawn at different spots each time you fire up that mission).
Furthermore, the tracks provide up to 70% as movement effect for the INF troops, while the broken terrain provides around 28-32 in that area of the map. (check the screenshot, the mortar unit -not motorized- refrains from using the tracks). I'd say troops on foot should favourize tracks, no? :p

• A force moving in non-road column formation will end up with subGroups moving off the main line of march. So even if the main line of march is along a road they will be off the road and slowing everyone down.

Oh yes, that's the theory. Using "undefined" for a move task, or "in-situ" for an attack-task, can deliver decent results, tho. It feels like the pathfinding starts to bring up failures caused by a mix of game speed ("fast"), plus you have tired troops (especially with the "no basics"-issue [:D][;)]), and CPU-load. Well, I'm not a programmer, but it feels like that.
Also, it seems like even when ordering columns units will be off the road, on some occasions. I've described that in another thread, dunno whether it deals with level of fatigue, or flaws in the pathfinding (it used to be pretty common in HttR). Maybe you could shed light on this.

• Ferry's/Fords take a long...time to cross. It's often quicker to go around them. If you want to force them to use the ferry, use a Shortest path.

That's what i use to do, I've started to do that in HttR already.
It seems you didn't understand what I was heading for, though. Lemme try again:
I'm talking about the difficulties to PASS the SMALL river DELTA with INF UNITS, prior to using the ferry nearby, although I was using the SHORTEST ROUTE.

The "quick" path can be viewed on one of the screenshots, it advised to ship around a massive river network, and then pass through a cluster of several enemy Bns. So, that wasn't an option [:D]
I've got the impression that crossing ferries takes WAY LESS time now, though (several hrs only). So i'll keep bypassing that way, to trash the AI [8D].[/quote]

Ideally it would be good to have saved games to check out the exact reasons for the behaviour in question.

I could come up with several hundreds of savegames for HttR, showing flaws in pathfinding, and I'm convinced that I could come up with at least 25% of that amount for COTA (IF i would have saved a game each time I encountered a flaw :p). No offense, but I'd rather say that this is a job for BetaTesters, not for customers. [:)]
The pathfinding might have attracted my attention, because I use to play on "fast" for like 70% of a map's mission time, simply because it offers a quick review of how orders are being executed.

Last but not least: I'm not trying to cause trouble or anything. But I'd really love to see these kinds of issues being fixed in the next installment (since I plan to buy that one too), as this is simply the best game in this genre, and a fix would make it even better.

Also, although I appreciate ppl's replies (eg. from MarkShot or from you), coming up with excellent tips for new players, or for those players (like me) who didn't read (due to lazyness) some interesting parts in the manuals, I'd really appreciate if you could check if you can reproduce (that should be easy, no savegames needed imho, especially for the arty-movement) any of the flaws displayed on the screenshots, along with some more detailed comments from you why paths are being processed like that at times, and if some flaws will be addressed in the future. Thanks in advance.

[MarkShot]

Goodguy,

Lifted from the COTA Readme which is viewable as the final step when you run the game installer. Which, by the way, I wrote.

- How to Report a Bug -

A bug is something that causes the game to crash or not function in a manner consistent which the documentation provided with the game. This is opposed to an enhancement request which is a change to what is modeled by the game or introduces user interface features or game play mechanics beyond what is covered in the documentation.

Of course, we are always happy to hear from our customers. We have endeavored to make COTA the best and most complete product possible within the constraints of our resources. For the most part, we will note your enhancement requests and where possible consider them for inclusion in future games.

In terms of bugs, to the extent that we can, we will attempt to diagnose reported problems and resolve them with future patches to the game. As with everything else, this is limited by our available resources and the severity of the problem.

Bugs should be reported on our technical support forum. The more detailed information which you can provide, the better chance we will have to find and fix it. Screenshots of the particular problem before and after are useful. Detailed step by step descriptions of how to reproduce the problem are useful. Most useful of all are save games that illustrate the problem. In some cases, we need save games from both before and after the problem occurs. If you think that you have found a bug, it would be best not to delete any save games until you have contacted us.

Detailed screen shots and descriptive text such as you have provided here are useful. However, save games illustrating before and after states of the game can be even more useful, since within their data structures may be clues as to the exact reason that something happened or failed to happen.

The beta testers and data design team members are volunteers who believe in the engine enough that we are honored that Dave permits us to be involved with its evolution. We are not a professional software quality assurance department for some large commercial system developer; although some of us do, in fact, find employment in the systems industry. Thus, we play the game, try various things, pay careful attention, and try to document very carefully any issues we come across.

Do we or can we catch everything? Of course, "no". There is only a handful of us and we each have different styles of play. We cannot possibly mimic all the interactions which the general public may conceive of. For example, I always pause when issuing orders, except for when I move bombard targeting coordinates. Clearly, you have a different playing style from me.

I personally felt that the pathing in the game was ready for "prime time" consumption. Despite that, being the cautious person that I am, I will often use the pathing tools to check what the game will do and I will place waypoints sometimes to be extra sure.

The only two issues I have generally had with pathing are:

(1) Sometimes with a large force assigned a particular route, some smaller groupings of units may choose to take an undesireable path, since you usually create the path from the HQ directly under your command.

(2) Occassionally, a map has many small streams or rivers, the AI will look for a certain footprint when executing a particular order (like FUPing). This may mean for a unit to reorg 1km South that it may need to travel 7km across a bridge and be on the opposite bank. Less than optimal.

Ah, but what's the point? (I seem to be drifting.)

If the number of beta testers and data design team members are few, then the number of Panther Games staff is even fewer. The COTA code in my opinion is very good and has a very low defect rate. Yet, I would not install it as the guidance system on a manned mission to Mars, because we for sure have missed stuff.

If as a player and customer, you find problems and it is important to you. Then, make your best effort to report, document, and gather any useful information towards fixing the problem. Also, please consider the real situation of the developer(s) and testers when you post and state your expectations. If you try to be reasonable in your expectations, then you will have done the most you can to improve the game and you are less likely to be dissappointed.

I hope that help make things clear about what testers do and how customers can realistically help assist that process. Thank you for your efforts.

[Arjuna]

GoodGuy,
 
As I said I do appreciate your feedback and I will have a look at these more closely once I finish my leave ( which I am supposed to be on ). So l will get back to you then.[:)]

[GoodGuy]

Hiya MarkShot....

Well, maybe you're getting me wrong. I didn't intend to transform you and your fellow testers into scapegoats, really.
I'm well aware of the fact that Panther Games is a veeeery small developer, grain of sand on the large beach of developers, figuratively.
So, I didn't even think of reporting this as bug.....

a) as I suspect this to be a rather complex issue...........
b) since human resources might be too scarce

.....so that this won't/can't be addressed in a patch.

Also, I wasn't sure whether this behaviour is being caused by new additions/rules (where I failed to read the corresponding part in the manual (as usual:lazyness)) or not, so it was meant as a call to enlighten me. :p

I wouldn't even think that this would be addressed this time, anyways, the "basics"-issue bothers me way more. I was just collecting examples, and was asking if someone could shed light on this, or, if it's a problem within the engine, if this will be subject to review for the next installment.

I'm convinced that the general pathfinding (which works like a charm many times) will bring up similar examples (like in the screenshots) after playing the game for a few minutes, so no need to pester panther with savegames, imho.

I'm well aware of the fact that savegames might deliver some good insights into situations where things go wrong, as I've been credited Tester myself, that's why I sent my savegame covering the "basics"-issue.
If Dave insists on getting a savegame for this here, I'll send one though, but the behaviours described above can be reproduced quickly, if you just play the game for a few mins, since it's been in there since HttR (I never played RDOA). Plus, I'd say it's a product of both, having new terrain types along with a certain way of computing/proposing paths that can be experienced since HttR.

I know that a group of volunteers can't deliver the results a fully fledged QA-division would produce, and I guess noone expects that.
The behaviours listed above are obvious, and aren't phenomena that occur every 2 yrs, only, nor are they hardly reproducible, so I don't think you need a huge division for those ones, that's why I was pointing at you, the fellow testers. Again, no offense intended.

I was just wondering why noone (I think there's only one question/thread dealing with the "safest" route) ever asked about pathfinding (erm byte me, I didn't use the search function ), so I felt like doing it.

EDIT: Yeah yeah Dave, I heard ya. I didn't mean to pester/put pressure on ya. Really [:)] And would you pls go away and enjoy yer holidays? [8D]

[philsublime]

Maybe a "move beahavior" option: "use roads only" would do the trick? I had the some problem as well sometimes.

[simovitch]

Perhaps a little known, but significant enhancement to the CotA movement algorithms is beeing overlooked here. The slope of the ground now has an effect on movement rate for both infantry and mech [8D] (see page 31 of the reference manual.)
 
Although it may look like the pathfinding routines have gone awry, They may be taking into account the steep slope of the ground, sometimes obscured by the underlying terrain elements.
 
The "spawning" locations in the rough terrain for mech units in "Charge of the Centaurs" has already been identified by the beta-testers. This condition occurs in 1 of the 3 variable scenario start options for the Italians, and the scenario will likely be amended and released through the official website or in a later patch.

[jungelsj_slith]

It does seem like some of these are an obvious bug. I would much appreciate this being fixed and not categorized as an 'update to the next project.' I'll try to keep an eye out for these and make sure to create a save game. Thanks for your input, arjuna.