Black Shark's AI

[bratwurst]

If you were part of the BFC community when I was, almost 10 years ago ;) , you would have seen that those grognards (the real ones, the ASL veterans, of which not so many are left over there :) ) preferred handmade, historically accurate scenarios. I can remember discussions about the following in quick battles:

 

1. "Cherry picking" desparate units rather than using the accurate TO&Es to make a very powerful, but historically inaccurate force composition.

 

2. Point costs for QBs without regard to historical rarity. You could load on Tiger Is and field them en masse even thought this was actually a fairly rare tank with only about 1300 made. You could do the same with even more obscure and rare units. I can remember the uber-Hetzer:

 

http://en.wikipedia.org/wiki/Hetzer

 

Very powerful as modeled in the game due to its sloped armor and very cheap because of its lack of turret. However, it wasn't nearly as common as it seemed to be in QBs. There were similar arguments made about the UK Fireflies.

 

3. Unrealistic engagements such as the meeting engagement. Allies should predominately be on the attack, Axis defending.

 

4. Unrealistic experience levels. Again, see the part about point cost without regard to rarity. In reality you would be fielding Conscripts, Green, and Regulars. Veterans would be rare, reserved for airborne units perhaps, Crack even rarer, Elite rarer still. Besides point costs, nothing prevented you from using Vets all the time in QBs.

 

Admittedly, much of this was corrected in CMBB. My stay over there was mostly right after CMBO was released.

 

Its ironic you bring up Combat Mission, because that game is where I learned to appreciate a well designed scenario. The community made PLENTY of them and I always enjoyed them more than QBs, both against the AI and other players. It was far more fun to be forced into an accurate fight, even if very difficult. It made you appreciate how heroic a guy like Michael Wittman must have seemed after something like Villers -Bocage. It made you appreciate how tough those gentlemen in the 101st Airborne were at Bastogne. Overall, it gave you a much richer more satisfying experience, IMO, at least. :)

 

We are different in that regard, as I enjoy QB's more then hand made scenarios. Could be I'm used to make missions myself and usually spend more time in editors then the actual game (SBProPE, SH, ArmA, CMSF etc.).

 

Right, so the first version has some flaws. I wonder though. What strategy sim was so much better then CM and with more historical accuracy back in those days? Maybe even today with CMAK and CMSF? Let me know. I should mention I'm not so keen on hexagonal 2D turnbased games so leave them out :D

[bratwurst]

This probably isn't the place to get into lengthy discussions about Combat Mission and what the grogs did or did not want with regards to CMSF but I think I can speak with more authority than ANYONE ELSE here on this matter, since I am one of the designers of the game.

 

While people did express a desire to have the old-style automap maker and QB generator in CMSF, it didn't have anything to do with them not liking the prebuilt scenarios or what was in CMSF needing to "catch up" to something in a earlier generation of game. More than anything they just didn't appreciate that the way those tools worked in the CMx1 engine would not work in the newer CMx2 engine.

 

The automap builder wasn't possible (at least not yet) because of the massively more detailed and varied underlying terrain mesh and the older style QB generator wouldn't work in the new system for a host of reasons but most notably because of all the advances in the A.I. routines.

 

In CMx1, the AI in play in QB's is pretty simple and based on controlling basic objective points (by and large). However, the entire scoring and mission structure in CMSF is more complex (due in part to the nature of the conflict depicted but for other internal reasons as well) and it simply wouldn't translate the same. CMSF uses a different approach in the Quick Battles, notably the fact the maps must be pre-created and the A.I. pre-issued various "action plans" which are embedded into the QB map itself to help determine how they are going to behave, based on the intended objectives and requirements of the QB itself. That is why you have certain QB maps now in CMSF that are meant for say a Meeting Engagement centered around a small town and another which is the same exact town and terrain map, but meant for a Static Defense. It's not the map itself (in this example) which changes, its the underlying strategic component of the AI that differs in the two QB maps. This sort of thing was simply impossible in the earlier engine.

 

So, when people first got into CMSF and found that one tool (the automap) was gone and the other (QB's) was so different, it took some time for them to get acclimated to the new system. In subsequent patches (and we have a new one about to be released hopefully next week) we added even more variety to the QB maps and that has also gone a long way into maturing the QB aspects of the game even more.

 

I hasten to add though that QB's were never and WILL never be the focus of the CM game series. We believe in professionally created, designed and "tweaked to a shiney gleam" type scenarios and campaigns and feel that they will always out-perform something generated on-the-fly.

 

Anyway, if you guys want to get into more details on the QB elements behind CMSF, it would be best to either do it in private, or on the Battlefront.com forums. That is why they exist afterall...

 

My appologies to the Mods if I stepped out of line by clarifying this point.

 

Madmatt

 

Sounds like good news to me. Looking fw to that patch :thumbup: CMSF however don't suffer much from this QB issue because the editor makes up for it. Making dynamic mission using the existing QB maps are only a matter of minutes. Now, back to BS.

[bratwurst]

I think both types of campaigns have their benefits. Hand crafted campaign missions are usually well built and challenging. On the other hand they are repetitive, once i've finished a mission I usually have no interest in flying it again. Also if it's a really tough one, I don't like having to re-fly it over and over again until I succeed....especially if it's beacuse there was one vehicle left I couldn't kill beacuse I ran out of ammo. :)

 

Also other then each vehicle thats specifically placed the world is often completly empty. There usually is no "front line" that you have to cross to get to the target. Overall I find most non-dynamic campaign missions I feel like I'm flying in a barren world until I get to the target. I'd love to get shot at 1/2 way to the target. :)

 

Dynamic campaigns are good because you never know whats going to happen in the mission, and it's always different. The F4 dynamic campaign is by far the best I've ever seen ( I honestly can't imagine why anybody would dislike it). You can be on a CAP mission and nothing happens, or all hell can break loose. I find the not knowing the most exciting part. You have a brief and a rough idea what to expect and that's about it.

 

Don't get me wrong, I love BS and am enjoying the first campaign. This sim is long overdue and I'll be playing it for years to come. I can't wait to see what DCS comes up with next!

 

Well said :thumbup:

[deviletk]

Is it hard to understand that with triggers it can be different everytime? And getting shot halfway to the mission area, well, not hard to do either. And it can be setup so its different intercepting enemies that maybe or maybe not comes halfway to the mission area. Very simple and very dynamic feeling.

[GGTharos]

I think both types of campaigns have their benefits. Hand crafted campaign missions are usually well built and challenging. On the other hand they are repetitive, once i've finished a mission I usually have no interest in flying it again. Also if it's a really tough one, I don't like having to re-fly it over and over again until I succeed....especially if it's beacuse there was one vehicle left I couldn't kill beacuse I ran out of ammo. :)

 

That's an odd thing to say - I suppose if, in a DC, you have to re-bomb the same airfield it somehow feels like a different mission to you? ;)

You're also mixing up 'scripted mission' with 'must kill everything' missions.

 

 

Also other then each vehicle thats specifically placed the world is often completly empty. There usually is no "front line" that you have to cross to get to the target. Overall I find most non-dynamic campaign missions I feel like I'm flying in a barren world until I get to the target. I'd love to get shot at 1/2 way to the target. :)

 

No real pilot would. You fly to the target for 2 hours, stay over it for 15 minutes, and then you fly 2 hours back to base. You get a couple refuelings inbetween. And you are once more mixing up a hand-made mission with a 'barren to the target' mission. Not only is your mix-up not necessarily realistic, but the two don't have to go together, either.

 

Dynamic campaigns are good because you never know whats going to happen in the mission, and it's always different. The F4 dynamic campaign is by far the best I've ever seen ( I honestly can't imagine why anybody would dislike it). You can be on a CAP mission and nothing happens, or all hell can break loose. I find the not knowing the most exciting part. You have a brief and a rough idea what to expect and that's about it.

 

I tend to have a pretty good idea of what to expect ... maybe it's because that DC is a uhm, logistical calculator and I realize this? :)

 

Don't get me wrong, I love BS and am enjoying the first campaign. This sim is long overdue and I'll be playing it for years to come. I can't wait to see what DCS comes up with next!

 

See ;)

[Rhino4]

On Topic: The Ai in DCS:BS has a loooong way to go in order to be convincing. As it stands now, the AI only does what it's told, rather than what is in its best interest for survival. Units that pop smoke and get the hell out of the AO when faced with overwhelming force would be a breath of fresh air...As would units that split forces and attempt flanking maneuvers/feints/retreats to fortified positions/ect and generally make a mess of anything that the mission planner is...planning:huh:...The AI in BS flows on rails defined by the roads or waypoints that the mission planner gives them. As such they are entirely predictable and easily dispatched by someone sufficiently skilled.

 

Off Topic: The only way to fix this (disclaimer: IMHO) is to overhaul both the AI logic and the mission editor. Instead of allowing the mission planner to decide the route, allow the mission planner to set targets, mission objectives, attack points, fallback points, defensive and offensive priorities as areas that the AI may move within and let the mission planner decide which units spawn in that area. These areas would be color coded in the mission editor and their size would be determined by the mission planner. Then, allow the ai you put there to do whatever it wants within those areas to accomplish the goals you set for it by following its own AI logic such as attack/pin/flank/call fire for effect (if you give your forces artillery)/retreat to a fallback point after attrition/ect...This allows the mission planner to stack the deck one way or another while still creating the feel of a "dynamic" land force and allowing the mission planner to still set mission completion goals. Because of this, the way the AI operates within the game is both more "realistic" (have to be carefull slinging that word around:smilewink:) and yet remains controlled.

 

Way Off Topic: Another thing that would help immersion would be allowing the player to interact with ground forces in some meaningfull way over the radio. Having the AI call for artillery if it's present is nice, but it would be nicer if they could call for your help too.

 

Way Waaay Off Topic: Everybody here is discussing the pros and cons of dynamic campaigns. Some said that they don't have the time/energy/motivation to create missions that they would then play. Others said it would ruin the fun if they knew what was coming (paraphrasing heavily). That's why I don't play missions I create except to make sure they work well for others. If I want suprises, I download missions from others...As for whether or not DCS should get a dynamic campaign, I'm all for it as long as they can find a way to make it work within the framework of the existing mission editor and fix...I shouldn't say fix: improve the AI first. Make some kind of campaign builder where your given control over unit makeup, area assignments, mission objectives, and overall victory conditions. Then let the builder insert prebuilt missions at strategic points to help further the plot line or overall force balance. Then start the campaign and let the good times roll. Once the campaign is started, don't let the player control unit assignments or mission objectives. He is, after all, just a lowly pilot...Combine these two arguments into one success that addresses some people's need for random elements while still letting the mission planner reign supreme. It can't be easy, or else everyone else would already have done it, but there has to be some way to combine scripted missions with a dynamic campaign to improve everyone's experience. I have definite ideas about how this would be accomplished, but this post is too long already...Sometimes I just talk too much...