COMBATSIM.COM: The Ultimate Combat Simulation and Strategy Gamers' Resource.
 

Campaign AI: Hub of the Wheel

  by Leonard "Viking1" Hjalmarson

 

  EF2 My first experience with a dynamic campaign engine was with EF2000. Although that wasn't so long ago, there has been a LOT of water flowing beneath that proverbial bridge; it feels like forever! Anyway, here is what I wrote at the time: (March, 1997)

If you have not yet flown in a dynamic simulation, then you simply have not experienced a realistic tactical environment. Although there are military flight simulations that approach the same complexity and involvement without true dynamism, like Janes Longbow with the Flashpoint upgrade, accept no substitute!

When you are striving for mastery of the air, land, or sea in a virtual battlefield, you get a sense of involvement and achieve a "suspension of disbelief" that cannot be had in any other way. If you have been an EF2000 pilot or have flown in Warbirds or Air Warrior, then you know what I mean.

In EF2000, there is ALWAYS something happening around you. Allied flights are heading out, you sometimes run across an F117 flying low, and when you get near the Russian airbases, fighters scramble to intercept you. In EF2000 using Browse Plane or in TactCom using Smartview, you really get a sense of the dynamism of the environment. It aids both suspension of disbelief as well as situational awareness, and impacts real-time tactical decision making.


TACTCOM
A screen shot from the EF2000 v.2 under 3dfx

The major weakness in version 1 of WARGEN was that it did not integrate ground action. WARGEN made use of SAMs and AAA to be sure, but limitations could be seen in EWR modelling, damage and reconstruction modelling, and there really is no ground war to mention. Look for improvements in all these areas in version II of WARGEN in Total Air War.

After Tactcom, the next military flight sim to be released with a dynamic engine was Longbow 2. Longbow 2 has a following far beyond the original because it is much more immersive. How does it accomplish this? Read on.

Around the same time as Longbow 2 was being released, Interactive Magic released their iF22. The campaign engine was very similar to the structure of DiD's award winning EF2000, and though the simulation had similar depth the beauty of the dynamic engine was obscured by mediocre graphics and a variety of bugs and problems.

So what are the essentials of the dynamic engine as seen in EF2000 and iF22? In short, when you fire up iF22 the campaign engine generates a set of missions based on the current situation. These missions are determined by the AI to be a realistic tactical response to a fluid situation. The player selects from a list of missions presented and at any given time may fly a CAP, Strike, Escort etc.

Then at the end of every mission the player generally elects to either exit the program or to fly a new selection that is generated by the AI based on the progess of the war. Its logical, and there is a flow and connection to what has gone before. But there's more to it than this.

Some of the essential elements that lead to the sense of immersion that I value so highly in this type of engine are these:

  • a logical flow and connection between missions, the success or failure of one mission impacts the selection process
  • objects stay dead between missions, realistic repair times
  • resources are a factor: loss of resources (aircraft, weapons) on both sides impacts ability to continue or succeed in the campaign
  • dynamics within the mission itself; no two missions are ever exactly alike. An element of unpredictability.

Other elements that add to the depth of the campaign:

  • levels of strategy: alliances may be made or broken during the course of a campaign that will in turn influence difficulty and complexity
  • tactical control: the player can actually task flights and perhaps control the flights once in progress
  • management of intelligence: the player is presented with updates on the tactical and strategic situation and this informs his decision making process
  • resource management is tactical: the player determines the allocation of resources and priority of resources. For example, he can request new supplies but then must task his limited aircraft to escort truck convoys or incoming supply aircraft.
  • A high degree of interaction between assets; a sophisticated system of communications and control (as of wingmen)
  • Resource management extended beyond military assets to economics and industry: hitting a factory may affect their ability to produce weapons or fuel
  • A real time campaign vs. a mission generation system.
  • Detailed briefings and debriefings including pilot records.

FLIGHT PLAN

Longbow 2 meets all of the first criteria and at least two from the second list. No currently published military flight simulation on the market meets all the criteria on the second list.

Not long after LB2s release we began hearing more about Janes F15. Some of my earliest questions on its structure had to do with dynamics. Early on, I asked Andy Hollis to describe the campaign, since it had't been clear to me as to how it was classifed: dynamic, semi, or...? In an interview a few months prior to the release Andy commented on the campaign engine:

Andy: F15's campaign system is very dynamic. It also is not like any currently existing system, so trying to use labels like "semi-dynamic" or "fully dynamic" as people have come to use them would be inaccurate. The goal of a dynamic campaign is to provide a compelling series of missions that combine together to provide a sense of : 1) overall purpose, 2) progress and cause/effect due to the player's actions, 3) being part of a much larger world, where the actions of others have causal effect, not just your own, and 4) continuity through resource management (planes, ordnance, aircrews) and world integrity (dead things stay dead and regenerate over time as appropriate).

The final key element, though, is variety, which provides for replayability and a strong sense of the unexpected. This can be accomplished in more than one way and each way has its advantages and disadvantages.

The "fully dynamic" approach that is in vogue right now obviously does a good job of each of these things when it is executed well (Yes, Martha, there are bad "fully dynamic" campaigns out there, too. In my opinion, Jane's Longbow 2 was a really good example of executing well on this theme. But then, I am biased. ;-)

The typical flip-side of fully dynamic, though, is that the kinds of mission types tend to be more generic, and the subtleties that a human mission designer can come up with are not implemented by the algorithmic mission/campaign generator. In LB2, this was alleviated somewhat by the "special case" hand-crafted missions that came up once in a while. This added a heightened sense of "what is possible" within the game world.

So, what's another good direction? F15's campaign system is yet another hybrid. For variety within the mission, things happen differently every time that mission "shell" gets played out. Bad guys come from different places, some things may happen this time, some may not. And the combinations of all these variables make playing the same mission shell completely different every time. Next, there are a variety of mission shells that can be selected from by the campaign director as the campaign progresses. These will depend on a variety of elements ... more variety, but with a definite sense of cause-and-effect.

F15 MAP

Finally, missions are not in a tree. There are a collection of whole campaign phase possibilities (groups of mission shell possibilities) that may or may not be a part of the actual campaign you will fly. There's a whole defining set of logic that dictates what phases can come before and after each other, which ones are prerequisites for others, and finally, which result conditions determine a final outcome.

Just to give you a sense of the scope of this, a single run-through of a full campaign will probably encompass only 25-30% of the total phase possibilities that are built into the campaign system."

To me it sounds like Janes F15 sports a strong but semi-dynamic engine, with an unusual and sophisticated strategic management system with a good dose of randomness thrown in. And this is indeed the feel I get when playing it.

F15 is ground breaking in some ways, most especially in the comms and control dimensions. It's that factor that was the largest attraction for me in actually playing the sim, although other factors, like excellent effects and high object detail, help in immersion.

I have long believed that dynamic campaigns were the wave of the future. When I first fired up F15 I thought: "I was wrong. The dynamic campaign is not THE most important factor." But Janes F15 didn't keep my interest. The world was not enduring enough.

In Janes F15 I can stop anytime. If I start a new campaign I know what to expect. And although it's an excellent simulation and a great deal of fun to fire up in a spare hour, I don't find myself drawn to return. Even though individual missions are dynamic and connect in logical ways, I'm not immersed in a believable world.

LB2 Nite Shot
Longbow 2 Dynamic Lighting. Click for 60 K image.

So.. what is it I am really missing? What really composes that enduring world? Is this just my own psychedelic rumination, or is there something definable here?

Click to continue

 

  TAW TASKING

Some years back I was studying literature in graduate school and I came across the writings of JRR Tolkien. In one particular essay, "On Fairie Stories" (In "Letters Presented to Charles Williams") he discusses the philosophy behind the creation of his world of Middle Earth.

Tolkien argues that the creation of sub-worlds is part of the human experience, the right and need for creation that we all possess as humans made in the image of a creative God. He says a great deal about the creation of persistent, consistent and believable worlds and his thoughts brought new light to the creation of fantasy worlds.

But the key point is that we have a desire to create these sub-worlds and to live in them. The question for our purposes then becomes: how to make the most attractive and believable world? What are the key elements that push toward the "willing suspension of disbelief" in military flight simulations? (Incidentally, that phrase itself comes from another writer: Coleridge.)

I've already given my implied answer that the heart of such a simulated world is the dynamic campaign engine. But now we have to go further, because there are a variety of approaches to dynamic engines. Which of these approaches is most effective in enabling "the willing suspension of disbelief?"

Let's go back to the best examples we have: Longbow 2, if22, and EF2000. What limits the ability of these sims to really draw me in?

Answer: many factors contribute to the limits of these sims to draw me in. But is it possible to indentify a determining element? The entire direction of this article implies that it is!

Let's use the analogy of a wagon wheel. At the center of the wheel is the hub; the hub supports every spoke. The hub is THE most critical part of the wheel. Lose a spoke and the wheel is weakened; lose the hub and the whole wheel collapses. The hub is the center and on it the entire structure depends.

F22 AWACS

What is really at the center of "suspension of disbelief?" For a time I thought it was a brew composed of graphics, control, comms and dynamics. Nope! Its not a brew at all. All the components "are NOT created equal." The hub is dynamics. Dynamics is the center, and everything else depends on that center.

But it gets even more complicated than that: there are soft and mushy hubs like those in EF2 and iF22, and then there are HE MAN HUBS... real time dynamics, integration of the ground war and tactical control, and the ultimate gold plating, real time multiplayer DACT.

You see, in EF2000 and in iF22 and in Longbow 2, the war ends between missions. Then a new mission set takes shape. I can choose, and there is no urgency. Even the need to generate a mission set destroys the suspension of disbelief. I am playing on a computer. The veil between the world of fantasy and reality is torn in two, and it's only with a mental effort that I re-enter that alternate world.

F22 Squadron

But all this is about to change! This doesn't mean that the first incarnations will be perfect. Maybe some of the spokes will be mis-aligned, or a bit thinner than they could be. But DiDs decision to separate TAW from F22 ADF is a ray of light in the darkness, reason to hope, because it gives them more time to develop their vision of a real time campaign where the player has tactical and strategic control over an integrated air and ground war in real time.

And TAW isn't the only game in town this summer: Microprose has their own answer to the total air war in Falcon 4.0 which will be the first entry to real time dynamics in a military flight simulation. And not long afterward Microprose will add multiplayer dissimilar air-to-air combat with a MiG 29 add on!

F4 GUI Click for larger image..

And now you know where I have been going with all this. Maybe the idea itself isn't so revolutionary: its Command and Conquer with a simulation overlay. But it is the wave of the future, and its going to be HOT!

Those of us who have enjoyed the AWACS interface in F22: ADF have already had a taste of what is coming. It doesn't take that much imagination to go the extra distance. Imagine this interface integrated seamlessly with ADF in a real time campaign. The clock is always ticking. No longer are you limited to what you see in the AWACS screen; now you can call up new flights at will, so long as the base has the resources.

At any time you may choose from a number of scheduled missions that are based on the current requirements to successfully prosecute the war. Or, you can jump in to fly any current mission that is being flown by any allied F22. You still have the tactical control that ADF AWACS interface gave you, though TAW may extend this control to allow more flexibility.

Furthermore, you can redirect traffic, call up Intercepts, Scrambles, CAPs, Escorts, Vis Idents, Strike missions, schedule refueling, redirect your AWACS or JSTARS. You can also decide the loadout and edit waypoints when you fly (unless you jump into a mission in progress).

But this is only the basics. What is occurring behind the scenes is much more complex. There are levels of strategy: alliances may be made or broken during the course of a campaign that will in turn influence difficulty and complexity. You may start out in a position of advantage, but the wrong move may turn your allies into enemies. (Note: for discussion of the theoretical basis of these new simulations see our reprint of Air Theory for the 21st Century).

Total Air War

And you are required to manage the communications intelligence as it becomes available. You are presented with regular updates on the tactical and strategic situation and this informs your decision making process. The enemy is after army targets: defend them. You know that your priority must be air control, so you manage your resources and plan your strategy accordingly.

Resource management is tactical and critical. If you are low on AIM9xs you will have to wait, and when they come in you must ensure that they arrive safely. But this means committing key assets to an escort mission. If your base comes under serious threat while they are gone... Or if the enemy takes our your roads and bridges..

There is a high degree of interaction between assets. The information war is a key component, and you must win it. Be careful to defend your AWACS and JSTARS and take care of your FACs. Your EWR and GCI network is critical. The same goes for the enemies systems: disable his comm links and you are on the way to air dominance.

And all this occurs in real time. There is no stopping to generate a set of missions based on the last eight hours of simulated time. The clock is always ticking and you have a sense of participating in a real though alternate reality. Is this psychosis, or is it .... Total Air War... or Falcon 4.0?

Von Clausewitz, the renowned theorist of war, stated that "a certain grasp of military affairs is vital for those in charge of general policy." Its interesting that the evolution of the dynamic campaign system has come about partly in response to the demands of the modern battlefield. So art imitates life! While Total Air War and Falcon 4.0 purposefully reflect the theories of John Warden III, the dynamic campaign itself is a necessity for bringing those theories to life. John Warden:

"To summarize: understand the political and technological environment; identify political objectives; determine how you want to induce the enemy to do your will (imposed cost, paralysis, or destruction); use the five-ring systems analysis to get sufficient information on the enemy to make possible identification of appropriate centers of gravity; and attack the right targets in parallel as quickly as possible. To make all this a little more understandable, it is useful to finish by mentioning the Gulf War's key strategic and operational lessons, which look as though they will be useful for the next quarter-century or more.

We can identify 10 concepts that summarize the revolution of the Gulf War and that must be taken into account as we develop new force levels and strategy:

  • The importance of strategic attack and the fragility of states at the strategic level of war
  • Fatal consequences of losing strategic air superiority
  • The overwhelming effects of parallel warfare
  • The value of precision weapons
  • The fragility of surface forces at the operational level of war
  • Fatal consequence of losing operational air superiority
  • The redefinition of mass and surprise by stealth and precision
  • The viability of "air occupation"
  • The dominance of airpower
  • The importance of information at the strategic and operational levels

Does this mean all campaigns must be fully dynamic or "real time" to be fun? Not at all. It means that I wish that they could be, it means that the dynamic campaign system has the ability to most closely replicate a modern military conflict, and it means that for me this type of campaign has the most potential for the willing suspension of disbelief. Your mileage may vary! But either way, it's going to be a great year! Watch for several responses to this piece in the next few days.

Roger Godfrey on Dynamic Campaigns

Join a discussion forum on this article by clicking HERE.

 

© 1997 - 2000 COMBATSIM.COM, Inc. All Rights Reserved.

Last Updated November 15th, 1999

© 2014 COMBATSIM.COM - All Rights Reserved