[quote=BeachAV8R
I often hear people refer to MiG Alley's campaign as a pretty good one..but I don't know anything about it..I never played it so perhaps someone can enlighten me on things like object permanence and supply and the ground war - I'd be interested to know how that sim was. [/quote]MiG Alley's campaign is probably best described as a turn-based strategy game. The clock doesn't run by the second as in F4 or EECH, but rather each day is divided into a number of phases (morning, afternoon etc). For each phase the player gets to set the tasking for every flight in the theatre. This can be as simple as clicking on a target and let the computer sort everything out for you (though this doesn't really work very well) or setting up the entire mission profile yourself. Once this is done, you can choose to fly any aircraft in your OOB. Flying as a wingman works particularly well compared to many other sims. After the mission, the time advances to the next phase. Everything that happened during the 3D mission (including non-player flights) is taken into account, including airframe damage and pilot fatigue/morale. Damage is persistent, facilities may be repaired over time depending on the local supply level. Supplies can be bottlenecked so that they may be abundant in one sector but depleted in another. Reconnaissance flights are important to assess damage and identify hidden supply bases (in time, you will find that the enemy starts using mules rather than trucks and hides supplies in the woods rather than in warehouses).

If it wasn't for the frustrating CTD's, I would rate the Mig Alley campaign as the best I've ever seen in a flight sim.

Mark - That is combat flight sim porn at it's finest! <g> Ah, those type of screens take me back to the day of the Tornado mission planner.

Thanks for the description of the MiG Alley campaign Johan - that sounds interesting and definitely approachable from a programming perspective.

BeachAV8R

I think ED is moving in this direction, they already have a semi decent fast mission generator; I mean it goes above and beyond Il2s QMB, and I would say, that it surpasses strike fighters campaign mission generator...In mission generation only. It falls apart with briefings and stat tracking.

This is all super odd though, because DCS as it stands now has all the stat tracking in it! there's pilot a log book. There's even a log at the end of every mission. If you can save a replay you can save a screen shot of everything that happened last mission.

Thinking further, if DCS can save and replay tracks, all that is missing is something in the background to save the tracks from the previous mission, open that track, and...basically run through it in .5 seconds and figure out the final positions and states of all the units on the map. Then it can run through logic and utilize those fancy airbase supply features. finally it can generate a new plausible mission for both sides.


That's all IMO of course. It does seem to me they are putting together the foundation of something.

I think it's silly to try to make "notes" from one session and use it to help generate the next. By the time you store enough info to make a good enough basis for the next "generated mission" you might as well just save-state the entire thing.

I mean when you play Half Life 2 and press "save game" and "load game" it isn't kinda like before. It's EXACTLY like before. Why can't flight sims have that? Then you don't have to worry about "mission generation" because there's only one long mission.

If you want an idea what the Mig Alley campaign was like fly BOB2 right now. Same company Rowan admittedly significantly tweaked by BDG.
Like some of the previous posters I am sim regressing. I like DCS in all its forms but love BOB2 and F4 BMS4.32; I like Steel Beasts / Steel Fury but love Panzer Elite. What's lost in graphics is more than made up by immersion, replayablity and plain educative entertainment.
At this rate I'll be playing micropose Fleet defender and M1A2 Tank Platoon by the end of the year!

TIM

Well, if you are flying as a single pilot campaign, that would mirror reality in most situations. You go, fly your mission, then you wait X number of hours or a day or so and fly your next mission. You don't really need all the real time movement in between. The only situation where that would really matter would be if you were doing multiple sorties one after another (like the A-10 did in the Gulf War), but that could be solved by just making that mission one long mission. Would it be better to make it one continuous campaign..sure..but the difference between continuous and one where you take 8 hour breaks where it generates data from the previous mission isn't much different when presented to the pilot flying subsequent missions. If that makes any sense.

Of course, if you just want to eavesdrop on the whole campaign in real time, that is another matter.

An example would be when you go to sleep at night. Lots of stuff happened around the world at night while you were sleeping that eight hours away, but there is no difference in your perception when you wake up at 7AM if that stuff was generated automatically or if it really happened. Lots of trees fell in the forest is all you read in the paper. Or DID they?

LOL..

BeachAV8R

Good points, Beach. This is especially true if your preference is to fly only certain times of day, and sometimes you have no choice.

For example, I remember flying the F-117 in FreeFalcon, running the DC at 64X speed (not recommended) while waiting for the sun to set. Seemed like it took forever for the first stealth fighter missions to appear. In this case I would have preferred EF2000's 8-hour cycle (which seems very similar to MiG Alley).

So maybe a 24-hour ticking clock isn't so important, as long as the world is persistent and missions are generated on-the-fly for the current situation.

I applaud EF2000 for giving an optional fast-forward 8-hour display of the DC in action, as shown in my crappy video. Although it takes longer than the automatic skip to the next 8-hour period after a mission is over (calculations are instantaneous), you get a better idea of what's happening in fluid motion.

Just the fact that in EF2000 I can observe (in fast-forward 2D) an entire campaign play out to conclusion without any participation whatsoever, is to me the essence of a DC.

Hi all,

About dynamic campaign in DCS.

We in the Israeli community and the Greek community have been working hard to create such campaigns.
Obviously, DCS does not have a dynamic campaign just yet, but that does not stop us from using scripts and the DCS features to try and make it as dynamic as possible.

Right now we have managed to hold 3 hour sessions for 64 players with no problems.

Key features of the Campaign:
-Register as a Squad and get assigned to your side & dedicated Airbase.

-Select available Aircraft numbers & types for your Airbase Warehouses.

-Each side starts with an amount of virtual currency in order to ''buy'' aircraft depending their cost.Aircraft losses decrease their initial available number.

-Squadron Leaders will have to plan weekly missions in order to serve their Sides master plan.

-Planning,adding Waypoints & other info into the mission file through the Multiplayer Mission Planner (Browser - Based) during the week by the Squadron Leaders.

-Capture Enemy Airbases by destroying all their facilities /ground forces and by transporting friendly units with UH-1H's /Mi-8's near their location.

-Relocate your squadrons aircrafts in case your Airbase is captured.

-Plan Special Ops Transport Missions with Choppers to temporarily capture or destroy enemy assets.

-Destroy enemy Radar sites to reduce the available enemy intercept fighters and enemy GCI picture.

-Intercept AI fighters will only take off when the Airbase Radars spot enemy aircraft threat.

-Provide CAS and escort friendly Ground units to capture and hold cities
Attack enemy naval units that will attempt to provide Air Defense cover or Ground Attack.

-Ground Commanders can provide GCI services & control Air ,Naval & Ground AI units in order to support the war.

-Resource Managment . Ammo Depots resupply the Airbases. Squadrons will have to protect them in order to keep the level of available ammunition high.

-In order to provide more intensive air to air fights it is decided to use only the following air to air missiles during the campaign. Available Air to Air Missiles : R-60M / R-73 / AIM-9M / AIM-9P / AIM-7 / R-27R

-The state of every mission is saved at the end, all ground units status and new positions are "loaded" in the next session.

-Pilot & Team statistics avaiable after each mission (Example: http://89.163.173.82/DCS/Default.html )

ED post for the campaign:
http://forums.eagle.ru/showthread.php?t=131726

SOP - http://www.anon6.com/forum/m/13804678/vi...arts-oct23-2014

Twitch broadcasts -
http://www.twitch.tv/morrtz/b/581072024
http://www.twitch.tv/morrtz/b/582264750

The campaign is saved from mission to mission, ground forces are saved, resources are managed by the DCS engine and by us via 3rd party scripts to handle the statistics - http://89.163.173.82/DCS/Default.html

The only limit we have currently is the number of units we can have in the mission which restricts us from doing a very big moving campaign.
It will probably be resolved when ED publishes the dedicated server.

If anyone wants some help in setting these kind of campaigns up, you can send a PM to me and I'll send you the info about it.

I think that is an important point. Whether the campaign skips forward in increments (2, 4, 8, or 1 day) or runs seamlessly, it would be nice if a spectator could witness the whole thing. The degree to which a player influences the campaign should be tangible though in a "game" setting and you would expect lack of participation to result in losses more often than victories. That isn't true, of course, in real life (one pilot would rarely change the tide of an entire campaign), but personally I want DCS missions to be weighted more toward gameplay than realism.

BeachAV8R

It's nice to have a toggle like "Player Performance Influence Campaign?"...



And I like the way the option is implemented in EF2000, a player score combined with airbase strength (score = who has opportunity to advance, map = how strong are defenses).

You can start off well enough...




...but then it can go downhill as the Russians try to push you into the North Sea...





This is when you pray the Swedes will join the fight with NATO (although in this case it's probably too late)!


An interesting aspect to DID campaigns is national alliances...

I've fiddled a bit with this. It would be an awful lot of work, whether it was a true dynamic campaign, or a serial mission generator with history, or some hybrid.

But, as illustrated by the impressive Summer Rain project, it is possible. The DCS engine is open enough to allow a third party to create an SP or coop campaign generator. It's not ideal, as there are some engine limitations that make this harder than it should be, but the Mist lua library has helped.

If you were to do a true dynamic campaign, you'd need need to do a lot of work communicating with the DCS server, making operational and strategic commands in your code. You'd leave the tactical decision to the DCS AI, assuming they can improve their AI to the point where it could fight a day long engagement unsupervised. Failing that, you'd have to implement a TacAI as well.

If you went with the serial mission generator, you could abstract a lot of the conflict resolution and potentially produce more believable results. The onus would be on you to turn those abstract results into a mission file. That is quite difficult, as you have to evaluate terrain and other features to decide where to place units and statics.

Besides the daunting task of programming logistics, AI and interface, I think you'd also want more static and ground unit types than are currently available. So, you'd be taking on some asset creation as well.

Sounds like a fun project.

That's the thing. It is much different. Every campaign that tries to generate a string of missions based on the last has really really sucked and sucked hard. They never get close to continuity. Ever mission feels too coincidental. Oh look a Bf-109 patrol crossed our patrol path at the 58% point... again. It's theoretically possible to make good continuity between separate generated missions but it hasn't been done and I have no confidence that it will be.

The Silent Hunter (3, 4, 5) patrol missions are a good example of how you can have a long-duration campaign that isn't a "string of missions" with good continuity by use of "real" save games and time acceleration.

I still don't see how it's different. LB2 generated missions based on the last with no continuous err..continuity. And one could certainly say the repetitive wall of MiGs in Falcon 4 was no different than your Bf-109 example. You ALWAYS encountered a wave of flights on every Falcon 4 mission. Whether that happened continuously each hour for eight hours or if you were only there to witness it once every eight hours doesn't really much matter in my opinion. Unless you can get the AI to start writing AARs..then we are on to something..

BeachAV8R

Surprisingly, creating a scripted mission is actually easier, it is very meticulous work but it is easier as you know exactly what will happen and how to plan for each event.
When designing a dynamic scenario, there's many accountable things you have to take into consideration to validate a balanced and randomized scenario.

There are several ways to resolving the dynamic campaign problem of balancing and randomizing, none of them easy, especially when trying to satisfy many aspects and visions of a dynamic campaign. For example, do you give someone one life or several lives? how will it impact the campaign continuity? will the campaign run on a weekly basis or on a 24/7 mode?

These kind of questions will bring many obstacles to overcome, as you can see in Summer Rain, we have broken through many barriers. The biggest I would say is the network stability and continuity from mission to mission, we have gone through a long road to have these features available today.

This campaign is the result of more than a year of learning, building, testing, reasoning and eventually drawing the best conclusion.

Thanks for your post. It's always great to see someone that shares my beliefs regarding this issue.






Well like BeachAV8R said, you have LB2 campaign! It's a campaign that "generates a string of missions based on the last" and at the same time it's awesome (one of the best dynamic campaigns ever made, at least IMO) which really gives the player a sense of continuity (for example you could see that an area that was occupied by a ground formation in a mission, will still be occupied by the same unit in the next generated mission).
Another great examples of campaign based on this principle (and that didn't "sucked hard") were the ones that I already mentioned such as WOFF and even some IL2 dynamic campaign generators such as DCG aren't bad at all. There are many other examples as well.

Regarding the "one mission big scenario" that you seem defend for DCS, there's IMO a big problem of this approach within DCS:
- DCS is very demanding on even the best PC configurations when there is a big number of units in the same map/mission. So having a "one mission big scenario" as a dynamic campaign for DCS means that you must have all or at least most of the available units (for both sides) in the same map/mission and in order for such campaign to be good, immersive and believable this means a very big number of units, which by its turn could bring to its knees even the most advanced PC.

We'll see how much of a factor this is with EDGE, since the DX11 engine should reduce the hit for object volume.

AI unit count is a CPU issue far more than a GPU issue, so I doubt DX11 will help. DCS would have to put some work into improving AI performance, possibly by bubbling and partitioning, which would allow for simplified AI in non-critical regions, and using extra cores.

It will definitely help in rendering objects, though you're right about the behavior of said objects being CPU based. I'm not a huge fan of the bubble as it exists in Falcon 4.0 (though this was necessary for CPUs of the time), but I definitely believe that different AI behavior routines could be bubbled (in lieu of an all-or-nothing approach) in order to have our cake and eat it too.

Which campaign have you flown? On the normal KTO campaign, in 8-24 hours, almost all major enemy airbases get slaughtered and you basically wont see any enemy air activities after that.

87 missions in the Rolling Fire campaign: HERE!

I found that the enemy air initially went down..but bounced back..but enemy ground forces after 9 days were really getting pounded...



I would guess that depending on initial settings, PAK settings, types of missions preferences, my campaign may not be similar to someone else's (I always picked sliders that focused on SEAD, Strike, CAS, and generally ground pounding).. Also, this wasn't the most recent BMS offerings - so I'm sure changes have been made to the campaign engine over the years (decades!..lol..)



Ah..BUBBLES! That takes me back too..<g>!

BeachAV8R