Challenges of Jungle/Marine Warfare in Tanoa

[CaptainAzimuth 714]

Need to make a distinction between bushes and ground covering, ie.the high grass and tweeds which blind the player when prone and sometimes kneeling. AI don't seem to be viewblocked by the grass even when prone. Try playing campaign with terrain set to standard or higher and use the ground cover for stealth..notice the ai blasting you while they probably shouldn't. Now set the details level to low and feel the folly of your previous endeavor.

Jungle will have even more ground cover so it's pretty important. I know they lower the model into the ground somewhat for long distance viewing but I don't think they do for mid distance or cqb. The game looks awfully barren with low terrain setting but can't help but feel I'm at a silly disadvantage when going higher. Unless of course, hiding behind invisible trees is your thing.

Ahh, i see what you mean now. My Terrain is on Ultra, but i rarely have the issue of AI shooting me through terrain so to speak, through ground clutter. Perhaps a solution could be have an extra half foot layer of blocking view for the AI? But i don't know how logical that is. The only time i had a problem like that was when i ran into a particular player, who was highly skilled in 3rd person shooting over terrain. Literally, i couldn't kill the guy because every time i flanked, he could shoot over the terrain where i was coming from. He was prone the entire time, and i ended up getting shot in the face before my chest was over the hill. But other than that, most of my deaths to AI come from being impatient and poorly planning my advances, because well, i was impatient, just wanted to finish a certain part of the campaign. Usually, if an AI is prone for awhile, ill shoot in the general direction, or pop smoke, and get a flank.

[mistyronin 1181]

My Terrain is on Ultra, but i rarely have the issue of AI shooting me through terrain so to speak

Neither do I. Until now I'm only been shot if seen by the enemy on the open. Then I return fire in the area where the bullets come from or if its too intense, try to use support artillery or CAS.

[CaptainAzimuth 714]

Welp, there's always Dev Branch testing, so i guess when ever that happens, cough, next year, than we'll find out. =D

[froggyluv 2136]

@Semiconducter:

So you like the little 'stand, now sit down, now turn around, now crouch' a rifle squad that is about to encounter an incoming tank while in the middle of town -do right now? Then you and I are very different gamers I suppose, frankly the idea that they could quickly garner that they have no counter and will proceed either indoors or very hard cover and stay outta sight seems like progress to me.

AI should play out as close to reality as possible in relative terms -should real life Infantry encounter a tank in an urban (or an environment) and have no viable defense/counter attack -they would hunker down and wait for support -THAT is the mission makers responsibility. Send your teams out and they should respond in realistic, interesting and unique ways and the mission maker should also support them with the proper, well, support.

Currently, show me a mission that I cant just pull up and wreak havok against idiotic infantry that just dance around? Can they reasonaly be scripted to run indoors all the time or get to nearby forest? To some level yes but nothing remotely close to a level we can count on nor enjoy. This sort of progress needs to happen down at core levels. What your afraid of is also what we have all (the AI Reformist Committe) been calling for, for a loooong time:

OVERIDE COMMANDS!!!

..as well as _this disableAI "FINALLYGOTABRAIN"

should cover it.

[semiconductor 309]

frankly the idea that they could quickly garner that they have no counter

Yeah, that would be cool, I agree, but it's something that easy to say but hard to do. The problem is, a computer can't "garner", it can only blindly follow a routine that someone created for it. As for now, no computer program can interpret the world it experiencing, inspect and compare objects, observe events make a logical connection between them, and then, based on accumulated knowledge and experience, make its own judgement and arrive to a decision. All that is a requirement for being able to tell that "This group of objects is indeed an enemy armor platoon." without external help. Hell, even the majority of the humans (the crown of evolution, mind you) aren't able to tell that without additional learning. If BI would somehow manage to create an AI that will be able to experience the world it exists in with efficiency of a smart monkey, it would not just be a progress for some $50 computer game — it would be a historical breakthrough for a mankind. :D

What's left? A scripted solution, a bunch of "if...elses". Here's another problem: that solution should somehow be relatively universal. Since AI can't think and decide on its own, that solution should have a hard-coded logic that will cover the major part of possible tank platoon combinations and conditions, squad abilities and armament, town layouts and so on. I'm guessing it would be extremely big and extremely slow. If BI makes that logic a bit simpler then it becomes less universal thus causing dumb AI behaviour that will, in turn, cause mission maker to deal with situations created by somewhat-smart-but-not-smart-enough-AI. The AI would actually feel even more frustrating because it would quite often do something that mission maker does not expect from it.

IMO, BI should concentrate on small building blocks of AI behaviour from which mission makers would be able to build a behaviour they want to see on certain mission. Something like more advanced Move waypoints, maybe introduce new "SEIZE BUILDING" waypoint but not something global. For a sandbox tools sometimes are more important than premade complex solutions. :)

Edited June 27, 2015 by Semiconductor

[chortles 263]

Yeah, that would be cool, I agree, but it's something that easy to say but hard to do. The problem is, a computer can't "garner", it can only blindly follow a routine that someone created for it. As for now, no computer program can interpret the world it experiencing, inspect and compare objects, observe events make a logical connection between them, and then, based on accumulated knowledge and experience, make its own judgement and arrive to a decision. All that is a requirement for being able to tell that "This group of objects is indeed an enemy armor platoon." without external help. Hell, even the majority of the humans (the crown of evolution, mind you) aren't able to tell that without additional learning. If BI would somehow manage to create an AI that will be able to experience the world it exists in with efficiency of a smart monkey, it would not just be a progress for some $50 computer game — it would be a historical breakthrough for a mankind. :D

And even bigger studios with more backing have yet to perfect the concept... not least because it's questionable how worthwhile the attempt would be for almost any computer game:

I’ve talked with plenty of people over the years about “good game AI.†I’ve had developers and critics alike tell me that good AI is impossible, because, to them, good AI is what is technically known as “strong AI,†or AI that is essentially as smart, if not smarter than a human. Personally, I don’t think that’s good game AI. Good game AI is that which makes a game enjoyable to play.

A strong AI would be more like another player, with the same goals a player has: it wants to win. Good game AI doesn’t want to win, it wants to make you earn your win. This is why the alien is not some realistic simulation of movie’s greatest monster, but an impression of it instead.

[froggyluv 2136]

All that is a requirement for being able to tell that "This group of objects is indeed an enemy armor platoon." without external help. Hell, even the majority of the humans (the crown of evolution, mind you) aren't able to tell that without additional learning. If BI would somehow manage to create an AI that will be able to experience the world it exists in with efficiency of a smart monkey, it would not just be a progress for some $50 computer game — it would be a historical breakthrough for a mankind. :D

A tad overdramatic eh?

Lets say a simple scenario in which 1 tank is rolling into a town defended by an AI squad without an AT soldier or he hasn't sufficient ammo. As soon as theres knowsabout, the check squad is done and their commander calls out:

A. Looks for closest building to fit his squad -check nearest building/count buildingpositions

B. Orders his squad to flat out sprint to a high density cover area and pray they dont get mowed down -requires the now promised AI Sprint function :p

C. Status Quo as we have it now - order a few of his men to target said armor with impotent small arms while the other guys creep along at strange angles not seeming to care

Ill take A and B anyday. These aren't massive Clive Linden inspired delusions, but doable AI enhancements that this game sorely needs. Believe it or not Im acually huge fan of this games AI and progress -just stopped using AI mods for the campaign and having very good time of it. As regards to the Alien reference -pretty sure 99% of fans would much rather enjoy flushing hiding AI out of buildings with Armor than just fappin away at the standard turkey shoot.

[chortles 263]

As regards to the Alien reference -pretty sure 99% of fans would much rather enjoy flushing hiding AI out of buildings with Armor than just fappin away at the standard turkey shoot.

Apparently "the standard turkey shoot" is because when AAA shooter devs implemented flanking behavior the very QA testers who'd professed to want "tactical" AI then complained that the AI had teleported... presumable making the AI programmers' work look like a waste to their project leads. :rolleyes:

I've linked this elsewhere, but A may be dependent on "correct" configuring of the terrain objects/buildings (just like "AI not spotting you through bushes") and oukej seems to have thrown in the towel on that WRT Altis and Stratis while promising that Tanoa will be better at this.

[semiconductor 309]

Lets say a simple scenario in which 1 tank is rolling into a town defended by an AI squad without an AT soldier or he hasn't sufficient ammo. As soon as theres knowsabout, the check squad is done and their commander calls out:

A. Looks for closest building to fit his squad -check nearest building/count buildingpositions

B. Orders his squad to flat out sprint to a high density cover area and pray they dont get mowed down -requires the now promised AI Sprint function :p

Again, yeah, it sounds simple for us humans, who are able think and act on their own (at least theoretically :rolleyes:) but an AI needs an extremely precise routine telling it what exactly it have to do. For example, let's try to convert our one-tank one-squad situation into some sort of pseudo-algorithm. Here's some checks that I can think of:

1. Is I_Tank_F object is a tank? Okay, that can be done by config checking.

2. Is the AI squad's location is a town? That requires hardcoded town areas (I think something like this already exists).

3. Should AI defend a town at all? What if it's just passing through on their way to destnation? Its waypoints should be checked.

4. Is certain I_Tank_F is rolling into town? Again, we need to check its waypoints. (It's a hack - IRL SL can't access enemy waypoints).

5. Is distance between tank and squad short enough to justify hiding? What if tank entered big town from the opposite side?

7. Is there a buiding that is

• in town area
  
• big enough for squad to fit in
  
• close enough to justify leaving a position and prevent running to another side of the town
  
• will provide sufficient cover from tank (damaged and partly ruined buildings won't do that)
  
• can be reached without appearing in tank's view
  

7a. What if there is no such building? Should a squad leave town of it's better to attempt to hide in unsuitable building?

8. Where exactly it should retreat to? To a nearset cover? What if nearest cover is across an open field that is clearly observable by tank commander?

9. What if tank have left the town? What if tank have left the town while squad was on its way to cover? Should it return to its original position or remain in cover?

10. How exactly various exceptional situations should be handled? What if there are incorrect value in the code? What if some unit stuck in textures? What if pathfinding won't work properly?

Not to mention that navigation and firing inside building is a whole another problem for an AI.

Note that all those checks are required for an very simplified situation. There is no other enemies, there is no allied squads, no support, no means for resupplying, no changing environment (building won't get hit with GBU-12 before squad reaches it) so the major part of required logic can be omitted. But even in this situation AI will often make a fool of itself by trying too hard to act intelligent. And to make the situation worse, all those check should be executed at real-time for at least several squads, eating precious FPS.

Don't get me wrong, I'm not opposing the idea of smarter AI but I think it's almost impossible to make an AI smart enough for a very complex and fast-paced sandbox. I'm all in for improving AI in terms of following orders (issued by players or mission/addon makers) but I don't believe in AI's ability to make a meaningful decisions on its own. So I'd rather have a somewhat dumb but obedient AI than sorry-that-we-killed-ourselves-while-trying-to-hide-at-enemy-base-from-that-tank-30km-away-kind of AI. :D

Edited June 27, 2015 by Semiconductor
typos

[froggyluv 2136]

I hear your points and agree giving AI playbooks is no easy task but I think your also over-analyzing. In Arma 2 I used to run the exact same tests in Utes, riflesquad without AT soldier in small village with a column of tanks approaching -the infantry squad was pretty darn good at getting out of sight and off the road.

1. Agreed

2. Unknown if they have that knowledge but thats another request -zoned/gridded defined areas

3. Well that depends on it's original waypoint, Sentry, Hold etc then yes. If they were merely passing town then the incoming or close proximity threat should buffer zone push them out of their original path to some extent. This is supposedly already in game as far back as Arm2 when Dev's stated AI will take least dangerous paths to destination if possible -of course, this has wide amount of parameters.

4. Distance, orientation, closing distance, and admittedly some amount of hack knoweldge would help - qualified by squadleaders skill level or something among those lines.

5. Well detected by West/East is already the current alarm, so that would need to be expanded with parameters from above. AI already "react", they just have no further hardcoded orders after their "cover dance'

6. What happened to me!?!

7. Agree this one would be tough but not impossible. Hack scan of closest buildings and pick the one with most available positions that can be reached in reasonably fast time via much needed sprint. I don't know if you saw the video I already posted of this in AI thread using Drongo's Command Enhancement Tool, but I was able to get a giant riflesquad indoors in mere seconds while AI solders just stood and died. Of course Im human ( regardless of wife's objections), but this is doable. Also again, Id rather see the AI try something, anything, than just stand and take it. This looks the most foolish to me and immedietly breaks immersion.

7a. If no such building, add flat out sprint to list of current options ie stand, dance and die. At least we'd have a variety of foolishness.

8. See above. If zone/grid's are implemented and they are within reasonable distance, than yes, they should hoof it to the forest etc.. depending on threats distance and impending nature. If you had a choice between stand and get shot or run like a fool in clear view of enemy -most would run regardless. More than anything, it shows the AI wants to live if nothing else.

9. Good point but as the AI already have a problem with returning to "Clear!", I'd say this goes hand in hand with that

10. Same as they do now -with years of calibration and recalibration. Firing from buildings is indeed a whole nother matter...

Again, there is nothing more foolish looking than acting like a threat ie Tank, directly in front of you, is not there while taking aim out to sea whilst crouching, standing up, re-aiming, and laying down -it's awful. All progress starts with two steps back and when it comes to AI maybe 10 tens back until one goes forward. Ever wonder why more and more of the older games that actually featured great AI have downgraded to 'node whack-a-mole' shooters or just linear scripted BF nonsense? Because people accept it. If everyone called out for better and more interesting AI, it would call the industry's hand and they would allocate a much more substantial slice of the budget for it at the cost of a mere God Ray. I often laugh at other early access game forums in which people decry "I hope they don't make the Ai too good......."...............

Edit: Again, I also think we are both visualizing two different scenarios. Your thinking how this could adversely effect official type campaigns/missions or tightly defined ones in which you need AI to 'not fuck around' while Im thinking both just drop and play editor scenarios or maybe spawned groups in something like Alive. Simple solution would be to define all these new behaviors under new banners that can simply be turned off or back to default with disableAI "SURVIVAL"; "AGGRESSION"; "FORTIFY" etc...

@Chortles: Thats unfortunate. Akin to why create extremely complex survival moves when you can just make an important NPC invulnerable to fire? *cough* *Cough*

Edited June 27, 2015 by froggyluv

[bad benson 1733]

@semiconductor: i also think this sounds a little overly dramatic. you make it sound like the AI currently thinks tanks are infantry and you'd have to write a whole new system ontop of the current one to make simple survival moves happen. squad leaders (atleast) already analyse as far as distinguishing if a tank is around or not. how else would they order their squad AT guy to fire upon it?

i think AI overall needs a responsive switch between aggressive and defensive behavior. so what if there are other threats around or a friendly team with AT and ammo for it? that still makes the close tank the biggest threat until someone (who cares if it's even blue on blue/red on red) destroys it.

defensive mode would mean only fire when fired upon which should already be very easily detectable considering there is engine side suppression detection in place. considering the AI is robotic and non-intelligent in every way indoors (choo! choo! movement on rails) i'd say making them go there is not a good idea though (atleast until we get something htat qualifies as indoors AI). and probably the most potentially mission breaking thing.

at the very least going prone and/or towards cover is needed though. not only in that situation. the AI overall lacks these very simple self preservation maneuvers. it's how they should react in general when confronted with superior firepower (like being suppressed the shit out of) imho. arma 3 AI seem more agressive in genreal to me, could be subjective though. it would make firefights more interesting overall if you'd be able to push back forces. have them reset a few hundred meters back and approach again or flank.

Edited June 27, 2015 by Bad Benson

[gibonez 18]

Few things.

1. Jungle warfare will hopefully make it hard to traverse the land on vehicles and heavy Main Battle Tanks will be restricted in their use and role.

Movement via foot and helicopter will hopefully be vital. This however depends entirely on how built up Tanoa is if it is just a lame resort island then ignore what I said. If it is a dense jungle map with few civilian towns and looks like vietnam then yes these should apply.

2. Towable weapons - Air carrying weapons - These two things should be vital in a Jungle map. Due to the rough terrain getting the weapons set up in remote areas should require tow mechanics and the current airlift mechanics in order to transport and emplace guns, missiles, anti air and other crew served weapons.

3. Artillery - Due to the heavy brush Artillery must play a vital role in all aspects of gameplay as it is one of the most effective means of suppressing the enemy in this type of environment.

https://www.youtube.com/watch?v=SscN-aBYHNI

4. AI - Not entirely sure how but they will have to make it so AI cannot see through thick brush.

[CaptainAzimuth 714]

Few things.

1. Jungle warfare will hopefully make it hard to traverse the land on vehicles and heavy Main Battle Tanks will be restricted in their use and role.

Movement via foot and helicopter will hopefully be vital. This however depends entirely on how built up Tanoa is if it is just a lame resort island then ignore what I said. If it is a dense jungle map with few civilian towns and looks like vietnam then yes these should apply.

2. Towable weapons - Air carrying weapons - These two things should be vital in a Jungle map. Due to the rough terrain getting the weapons set up in remote areas should require tow mechanics and the current airlift mechanics in order to transport and emplace guns, missiles, anti air and other crew served weapons.

3. Artillery - Due to the heavy brush Artillery must play a vital role in all aspects of gameplay as it is one of the most effective means of suppressing the enemy in this type of environment.

https://www.youtube.com/watch?v=SscN-aBYHNI

4. AI - Not entirely sure how but they will have to make it so AI cannot see through thick brush.

AI can only see where you went through said brush. Meaning they will engage you through that bush, if they saw you go through it prior to engagement. I mean, it's only a bush. If i saw a guy run through a bush, i would unload in the bush... in fact, i've done this many times and got kills. =D

[Sterlingarcherz101 15]

AI can only see where you went through said brush. Meaning they will engage you through that bush, if they saw you go through it prior to engagement. I mean, it's only a bush. If i saw a guy run through a bush, i would unload in the bush... in fact, i've done this many times and got kills. =D

Thats all good. It's more the ground vegetation I'd be concerned about. On altis crawling in it prone you are at a disadvantage vs AI. It's ofp landscape as far as AI is concerned you on the otherhand forced to go to your knee or stand up . This is a pain on altis. in a jungle setting, a nightmare.

Hopefully the AI is also forced to go to its knees to shoot in such situations .

---------- Post added at 20:07 ---------- Previous post was at 20:05 ----------

AI can only see where you went through said brush. Meaning they will engage you through that bush, if they saw you go through it prior to engagement. I mean, it's only a bush. If i saw a guy run through a bush, i would unload in the bush... in fact, i've done this many times and got kills. =D

Thats all good. It's more the ground vegetation I'd be concerned about. On altis crawling in it prone you are at a disadvantage vs AI. It's ofp landscape as far as AI is concerned it doesn't need to shoot blind. It can see you clear as day.you on the otherhand forced to go to your knee or stand up . This is a pain on altis. in a jungle setting, a nightmare.

Hopefully the AI is also forced to go to its knees to shoot in such situations .

[chortles 263]

Double post is double post... though I wonder if the same remark(s) apply, re: correct configuration of terrain foliage/objects being key?

[CaptainAzimuth 714]

Thats all good. It's more the ground vegetation I'd be concerned about. On altis crawling in it prone you are at a disadvantage vs AI. It's ofp landscape as far as AI is concerned you on the otherhand forced to go to your knee or stand up . This is a pain on altis. in a jungle setting, a nightmare.

Hopefully the AI is also forced to go to its knees to shoot in such situations .

---------- Post added at 20:07 ---------- Previous post was at 20:05 ----------

Thats all good. It's more the ground vegetation I'd be concerned about. On altis crawling in it prone you are at a disadvantage vs AI. It's ofp landscape as far as AI is concerned it doesn't need to shoot blind. It can see you clear as day.you on the otherhand forced to go to your knee or stand up . This is a pain on altis. in a jungle setting, a nightmare.

Hopefully the AI is also forced to go to its knees to shoot in such situations .

Hehe, Ninja's by Chortles, double post.

Yeah, i guess grass/terrain and such is an un-solved thing... Then again, there could be ways of solving it. For example, AI view limited by such and such while prone if 3rdP = False, and AI view is fine if 3rdP = True. Seeing as some players would take advantage of the AI not being able to see prone, yet having 3rd Person themselves, it's smart to have it as a setting based factor.

Just learned something... when suppressing AI through heavy brush, they will immediately do the same to you. So i suggest picking a nice thick tree to stand by when suppressing fire, they don't take too kindly to bullets flying in there general direction, and insist on participating in a random shoot, more rounds down range stand off. This actually surprised me. It's great though, makes the AI feel like another pissed off person shooting back just going off of sounds, and tracers.

Edited June 30, 2015 by DarkSideSixOfficial

[oldy41 61]

I have just set up a very simple scenario with a small fire fight in a forrest, and must conclude:

With the AI in its current state I am very concerned about this jungle warfare thing.

One of the problems is AI spotting through foliage, which is definitely way beyond player capabilities.

Another issue is the bad approximation of concealment by grass and ground clutter. On small engagement distances this gets even more noticeable.

So with current AI I am afraid jungle warfare would probably feel awkward.

On the other hand, if this means BI intends to improve AI for the expansion: Be my guest!

[eggbeast 3684]

well in A2 we play jungle warfare against 150 AI in each zone, without any complaints about AI. you can survive with one or two squad mates or with a full 15 man platoon, supporting tank, APC, cobra and phantom...

in the jungle the key is stealth. no more running about everywhere, whole team in walk mode, having a guy 15m ahead on point, and no-one carries a rocket launcher (u have them in your nearby support vehicle) - this is because the enemy goes nuts when they see a guy with a rocket.

lastly, you move from tree to tree to maximise concealment and cover. everyone carries at least 10 grenades, including WP, which doubles for calling the cobra/phantom strikes.

we also have enemy AI set to max intelligence/skill and min accuracy.

we have absolutely zero complaints about AI. and tremendous fun.

we've been jungle fighting for 2 years solid.

hope that's reassuring.

[2nd ranger 282]

hope that's reassuring.

To hear how the AI works on user-made terrains in Arma 2? No.

[bad benson 1733]

@eggbeast: do you guys have viewblock geometry on your clutter? i remember playing around with that in arma 1 for wheat fields. i think the conclusion was that it worked until you got spotted once. then they just knew where you are even if you moved around prone in the ense field.

i woner how that behaves in arma 3

[Chairborne 2594]

Jungle.

I agree with you, AI needs improvement in certain areas to make the experience better.

Spotting, aiming, reaction times should be balanced out-of-the-box, meaning that they don't require any changes by the player to be realistic.

I know that it is possible already to obtain a good level of AI behavior but i think it requires a lot of fine-tuning or even worse community-developed mods (which require further maintenance time and sometimes degrade the game performance).

Maritime

I never imagined myself doing scuba firefights (is the enemy expected to keep scuba divers 24/7 patrolling hundreds of square kilometers of underwater terrain waiting for you? lol).

However what I think is of interest to the scope of Arma is rather moving objects from ship (pls BI gibe amphibious assault ships) to shore (pls BI gibe landing crafts) and perform amphibious landings.

Naval warfare, as in ship-to-ship combat and the likes, seemed a bit cluncky even in a strategic game like Wargame Red Dragon, where the size of the terrain was a lot bigger and the game had certain mechanics like radar/anti-radar systems, CIWS, anti-ship capabilities, anything up to frigate-size vessels, in Arma this is amplified by the fact that the size is a lot smaller and they require a lot more detail.

Maybe they would fit well as fire support for ground troops, like a huge powerful piece of artillery at sea. I think that would be nice.

[oldy41 61]

...

in the jungle the key is stealth. no more running about everywhere, whole team in walk mode, having a guy 15m ahead on point, and no-one carries a rocket launcher (u have them in your nearby support vehicle) - this is because the enemy goes nuts when they see a guy with a rocket.

lastly, you move from tree to tree to maximise concealment and cover. everyone carries at least 10 grenades, including WP, which doubles for calling the cobra/phantom strikes.

we also have enemy AI set to max intelligence/skill and min accuracy.

we have absolutely zero complaints about AI. and tremendous fun.

we've been jungle fighting for 2 years solid.

hope that's reassuring.

Sorry, no, its not completely reassuring me.

I've been playing the whole series beginning with the first OFP demo on tons of terrains.

Imho the A3 AI provides the best experience by far, and using AI mods like bcombat even improves that.

But then lying in deep clutter, which blocks your view of the whole world beyond 3 or 5 meters, not moving a single virtual muscle, not hearing anything besides crickets, not even blowing your virtual nose, ...

...and then suddenly hearing _your own avatars_ voice indicating an "enemy, 75 meters straight ahead" gives a clear impression, that the spotting system is not adequate for forrest and brush environments.

Imho nerfing AI accuracy to compensate for their super human spotting capabilities is not satisfying, since this is prone to yield ridiculous situations when suddenly having an AI in plain view.

[eggbeast 3684]

well, as i say, it has worked for us for 2 years now, no complaints and hundreds of old-timers through the server... it can be made to work... hop in and see for yourselves.

[oldy41 61]

well, as i say, it has worked for us for 2 years now, no complaints and hundreds of old-timers through the server... it can be made to work... hop in and see for yourselves.

Thanks for the invitation! I hope I can give it a try and return as an optimist, too! :)

[froggyluv 2136]

Messing around with boat missions in N'ziwasago just to get a jungle locale feel and one thing thats gonna need attention is sandbars -and the lack of the AI's ability to get around them. Player has the same issue but with help from Feint's Push script makes a nice workaround but the AI is literally dead in the water if its too shallow.

for the sake of gameplay, if the boat is in water just let the boat move slower in proportion to how deep/shallow -but please let it move or we're gonna have a lot of frustrated gamers and mission makers.