by Leonard Hjalmarson and Neil Mouneimne

The driver's position in tank sims will be more useful when the terrain has enough features and hazards to make finding a good path through more important than it currently is, but the driver can already benefit from an easier-to-read terrain engine by simply being able to discern the path of least resistance through the broken terrain. iM1A2 has terrain that was extremely difficult to read. Armored Fist 2's terrain was detailed at extremely close range, but it's shape and exact features were very difficult to read from the driver's position. Spearhead looks like it will do better in that regard, and M1TP2 again looks like it might top the competition in having intuitive terrain. You can fudge terrain readability severely in fixed-wing and helicopter sims, but ground warfare simulations are severely compromised when they don't address these issues effectively.

The two armor releases of 1997 make for an interesting comparison in terrain and its influence on realism in tactics. In many ways iMagics sim was a good one, but the weakness of the graphics engine really hampered tactical play. In the same way, Armored Fist IIs exagerrated terrain and missing treelines limited gameplay. AF2 involved the player in other ways, with lots of action and pleasing graphics, where iM1A2 attempted a higher level of simulation realism and tactical play but didn't put it all together with a realistic graphical environment.

Its not fair to overly criticize iMagic for their release of Abrams. Remember, this sim would have been in development as early as 1995 when 3d hardware was still a gleam in the eye of certain thoughtful engineers. But the outcome is the same: the graphics engine was not up to the task of providing a realistic or attractive environment, where the engines of late 1997 (witness Longbow 2 and F22: ADF) are more than capable. And thats the good news this year!

Where Novalogic had the graphics engine and iMagic had the systems and realism, this year will see the first armor simulations capable of bringing it all together in one package. Its difficult to say how much iPanzer '44 may be hampered by a rehash of iMagics graphics engine since there have been some enhancements, especially in the use of trees. Tim commented:

Forested areas will provide cover as well as establish maneuvering and firing lines. The forests will allow tank platoons to advance with some cover, but there may be a Tank Destroyer platoon waiting around the corner and infantry may be laying in wait for the player to come closer. Making use of cover to mask advances and coordinating infantry attacks to ferret out any surprises will be necessary to succeed in battle.

Likewise Panzer Commander, having being designed from the ground up for 3d hardware, may lack the texture detailing that could make it completely state of the art but still makes use of tree lines for cover. What about the modern sims: Spearhead and M1TP2?

Having just completed a preview of M1TP2, I can say without hesitation that we will see a truly state of the art simulation from any perspective. If the AI holds up to the startling level of its other dimensions, M1TP2 may set the standard for years to come. In the shot below right you get a sense of the detail in this simulation. It is simply astonishing...

The reality is that you sometimes feel you are driving in an artists canvas. It has a slightly French Impressionist feel to it... until the lead starts to fly! While this might worry some gamers with regard to frame rate, under 3dfx there is no need to worry whatsoever. Even my AMD 233 system, with its slightly weak FPU, moves this sim along fluidly under 3dfx hardware. In fact, I've been wishing that Microprose included an 800x600 mode!

Eye candy aside, the simulation benefits by tree lines, individual trees, and buildings that look real and take damage in realistic ways. The terrain engine allows great depth-of-field and "fencepost" type treelines. It also does an excellent job of showing true distances at low altitude. However, there is more to it than this. Resolution of enemy vehicles at a distance is much simpler, even with optical aids. Add the gunners 3x and 10x site, and you are really cookin! Between the terrain engine design and the strategy options you can go "hull down" a whole lot easier than in the original M1TP (and most other tank games for that matter). Now what about Spearhead?

In the last Spearhead beta I saw the terrain and object detailing couldn't be compared to M1TP2 and trees in the desert will be rare. However, detail and terrain are substantial enough to allow solid tactical gameplay. In combination with a multiplayer goal of 18, detailed systems modelling and significant tactical control over one's own platoon (as appropriate for a modern simulation) Spearhead should have enough depth to satisfy the hard core crowd.

Damage Modelling: the Achilles Heel of Armor Sims

Damage modelling falls under the category of physics in general. In this department there are probably more differences between the modern simulations and the historical ones. After all, armor itself has become a LOT more sophisticated since the second world war.

Damage issues that an armor designers must consider seem to go on forever:

• What side of the vehicle was hit?
• How much armor is on the hit side?
• What kind of armor(s) is/are on that side?
• How sloped is the armor/What is the bite angle of the impact?
• How effective is that armor on the kind of projectile that hits it?
• Is the armor's effectiveness degraded if penetrated? If not penetrated?
• What systems can be damaged from the hit side if penetrated?
• What kind of secondary effects are possible? (Ammo detonation, fuel fire)
• Even if the armor isn't penetrated, what kind of external equipment might be damaged?
• (Optics equipment, external fuel drums, etc)
• Does the projectile's effectiveness decrease over distance?
• Can soft cover reduce the effectiveness of the projectile or stop it completely?

On the heavy T-80 variant, you might have to expect to contend with as many as three different kinds of armor! In front you have reactive explosive armor (armor that explodes outwards when hit to deflect the force of incoming shells), and then composite armor underneath. Hit it from the back or overpenetrate and then you'll have to deal with steel armor in the rear.

Again the differences in comparison to flight simulations are substantial. In the flight sims we tend to view discrete system hit bubbles as a luxury. Few sims reward carefully aiming at the pilot or engine, in the first place because a missile is only aimed at a large object in general, so you just hope that the damage model picks a favorably lethal system to have been hit! But in tank combat, where the exact location of the shell hit determines a lot about the results, we need to be able to distinguish between the six sides of both the hull and the turret at the very least.

Ideally, we'd want to treat each major piece of the tank and each section of armor seperately, and then understand at exactly what angle does the shell hit the armor before and after the reactive armor goes off (if any). Suddenly the ballistic modelling is also figuring into this equation. All these factors are important in making armored combat realistic, but the processor cycle requirement is very high, so designers need to choose very carefully which aspects they can gloss over without harming gameplay and which are critical to maintaining realism.

One would expect the greatest detailing from the modern sims, but Rick Martinez commented on Panzer Commander that "we are taking everything into account. Damage resolution is the same: type of armor, slope of armor, velocity and explosive power of the projectile, and you can target a particular location for maximum damage." Meanwhile, M1TP2 is also looking great, especially in general physics:

If you drive the tank in a sloppy manner you can break the tread and immobilize yourself, like in a real tank. Penetration and overpenetration is well modeled. You can fire a sabot round straight through a treeline to hit the unit behind it, or shoot two enemies with one shot if they are lined up properly and the shot retains its energy.

The killing power of sabot tank rounds decreases over distance but missiles have the same killing power at all ranges. Reactive armor gradually loses effectiveness as it is used up in the course of a battle. If an M1's turret is hit and the ammo goes up, the blowout panels will blow off to the sides to vent the blast away from the crew. Small arms fire against the M1 will ricochet in a small shower of sparks.

And here are some comments from John Williamson of Zombie on Spearhead:

We model all 6 sides of all entities (top, left and right sides, front, rear, and bottom). The front of a tank is the most heavily armored (and the most slanted) and as such it is the hardest to kill from that angle. The rear is the least well protected and a shot to the rear, if not destroying the vehicle, will likely destroy the engine.

A shot to the side has a good chance of damaging the treads. The type of round used as well as the distance from the target also comes into play. Basically every element of the tank has a chance of being damaged, this includes: the engine, the laser, thermal, night vision, CITV, IVIS, radio, engine, treads, main gun, 50 cal, coax, etc.

For overall systems accuracy, its unlikely that anyone will touch Microprose this year. Much of the research done for Falcon 4 weapons systems can be applied to M1TP2. And M1TP2 is the only armor simulation slated for release this year which has a fully dynamic campaign and allows the player to act as Theatre Commander with control over air support, ground forces, artillery and armor.

M1TP2 will simulate the complete CITV system and IVIS controls. The virtual cockpit in the M1A2 is top notch and you can move seamlessly using your mouse from the Tank Commanders station to the Gunners station and to the unbuttoned position also (the driver position is not modelled although you can take control if you need to). The virtual cockpit reminded me of Silent Hunter and is composed of interactive photos from a real Abrams. Most buttons and knobs are clickable.

Spearhead won't be left behind in the systems department, however. Gunnery itself is simple, but has some nice touches. One button will take the laser rangefinder reading, and the other will fire the gun (on a two-button joystick). What's interesting is that the targeting computer - and hence, autostabilization - can be turned on and off at will. Unlike previous attempts at autostabilization, the gun doesn't stay magically lined up - the computer has to race to keep things straight. It handles the job just fine over smooth terrain with minor undulations, but put it over really broken terrain and rough transitions, and the gun tube will momentarily be misaligned until the system can properly compensate. Seems very simple, but it is an important facet in making the sim behave more like a tank and less like an arcade game.

The commander's CITV is shaping up to be the most realistically modelled yet seen in a simulation. Other tank sims seem to make some strange interpretations of the operation of the commander's station, and as such they don't seem to be very useful in the game at all. Spearhead now makes the commander's station much more potentially useful without it overshadowing the other positions.

Spearhead's CITV As one may expect, the commander's CITV view sits over the turret and can be rotated independently of the turret itself, although it's direction is affected by turret movement. As commander, you can spot targets independent of the gunner and order the gunner to aim at the target - during which time, the view might be temporarily lost as the CITV and turret both try to simultaneously align with the target. Alternatively, as commander, you can override the gunner's controls and lay the gun on target before handing it over to the gunner. This would be even more ideal in a multi-player/single tank configuration, but it is most likely that the multi-player/multi-tank setup will be preserved to satisfy deathmatch play requirements.

Campaign Dynamics and AI

All the armor sims of 1998 will model infantry in one way or another. In our interview with tanker Jeff Babineau, the underlying importance of infantry/armor cooperation was strongly affirmed as an area in need of improvement in PC armor sims. In Panzer Commander and iM1A2 Gold, infantry is composed of sprites. Control of infantry in Panzer Commander hadn't been clearly determined when I last spoke to Rick Martinez.

In iPanzer '44 and M1TP2, infantry will be under the control of the Platoon Commander. Control will be possible from the map view. Tim of Charybdis put it this way:

We will be representing the various types of infantry teams and once they are armed appropriately the infantry doctrines start becoming obvious tactics. A single infantryman can take down a tank with a bazooka or a well placed grenade so using friendly infantry to eliminate enemy infantry will be important.

They all have 3D articulated models, and will appear in combinations, with attached elements etc. The infantry will be controlled by the AI primarily, but you can give them orders on the map like any other unit under your control. The largest goal with the infantry is to make them an integral part of the way the battles are fought in the spirit of the way battles were fought.

There are "soft" targets in Spearhead in the form of occasional troops encountered, and these troops can be friendly, enemy or neutral. They will engage appropriate targets and inflict damage, but the player does not have any control of them.

As for general campaign dynamics, Spearhead and Panzer Commander have a campaign but the missions are linear, with no carry over of damage effects or recalculation of elements based on the two sides interactions. iPanzer '44 has what is termed a semi-dynamic engine where success or failure in one mission determines the choice of the next mission and there is some carry over of actions and decisions from one mission to the next.

Another key difference between iPanzer and Panzer Commander is in vehicle modelling. Panzer Commander allows the player to choose from twenty four drivable vehicles, and models morale elements in gameplay AI. Rick Martinez comments:

We've developed what we call the "thermometer of morale." If you have troops who are green they are going to have a higher chance of morale problems in a difficult situation. But this gets pretty sophisticated. We have a system that actively assesses threats given the present tactical situation. This varies widely given a whole number of factors. For instance, say you've just been flanked and you don't know it, but you are beginning to take fire from the right flank. If you are unbuttoned you will hear the fire from the right, but if you're buttoned this isn't going to affect your assessment as quickly because you think the fire is still coming from the front.

Now what about the attrition and loss issue? If you have superb troops and send them into a situation where they are hopelessly outgunned, you may lose them all because their thermometer of morale has a different scale. The top veteran tankers are more prepared to die in battle than the green ones if they receive a high commmitment order. We have levels of orders that go from probe to recon to attack to decisive attack level. These levels include in them a willingness to take casualties to take the objective.

This also affects other interactions. If green troops are getting slaughtered they may panic. Experienced troops will retreat in a more orderly fashion.

M1 Tank Platoon II is the only armor simulation due for release this year that features a fully dynamic campaign engine. This means that no two missions are ever the same, and that the players actions in one mission carry over to the next one. Fans of Longbow 2 or EF2000 know what this environment is like.

Each campaign is dynamic and progress is based on player performance. If you have a successful offensive mission, your next mission may involve fighting a retreating enemy position, or one where the enemy has fallen back to reinforcements. If you successfully defend a position, then your next mission may be offensive in nature so that you can take advantage of a weakened enemy. This is the key to making the campaign different every time you play.

Although you are the "small cog in the machine," an algorithm based on your performance is applied to the overall battle. If you do very, very well then your side will also do well and the campaign might only last for ten missions, but if your performance is weak you may be slugging it out for twenty five missions. As in Longbow 2, you will never see exactly the same scenario twice, though the start of a campaign may look very similar.

Toward the Virtual Battlefield

Finally, the only armor simulation this year that is part of a larger design goal is M1TP2. M1 Tank Platoon II will allow future games to interface with it even if they are more advanced in some respects. The type of vehicles that will be modeled hasn't been revealed, but our guess leans towards the A-10, with the possibility of a Bradley or Apache as well (Gunship 2010? Click shot on right for a larger image).

Obviously, this is cause for a lot of excitement! This kind of integration of the ground and air war in a multiplayer environment has been something of a holy grail for simulation designers as well as simulation lovers for years now. Its the advent of 3d accelerators combined with the Internet and advances in communications hardware and software that are enabling us to finally see the reality.

And attention to detail in M1TP2 won't stop at the tanks themselves; the same physics and graphics detailing applies for heavily armored aircraft such as the Hind or A-10. And lest you think only the A2G and G2A engagements are modelled, let it be known that airborne units carry AA weapons and will engage EACH OTHER as well as ground units. TOW missiles have sighting flares in the tail and fly in a matter that simply *looks* like they're being steered optically (click shot on right for larger image).

If you want to play M1TP2 as strictly a tactical game you can do so, setting waypoints and actions in advance and then observing the action from either the tactical map or an outside "Flying Carpet" view. You can still still issue commands with keyboard, while leaving other decisions in the hands of a virtual Tank Commander. And if you want to jump into more direct involvement in a tank under your command you are free to do so.

At this level it feels more like Novalogics' Armored Fist II, but with sharper graphics, much better depth of field, and greatly expanded tactical control. You can instruct your platoon to merely assault the enemy, get hull-down and engage the enemy, perform a hit-and-run manuever, bypass the enemy completely, or retreat. The "Flying Carpet" view is represented in most of the M1TP2 shots in this article.

Summary

The armor simulations of 1998 are set to take us to a new level of immersion and realism in their respective settings. All these sims will offer more tactical depth and more fun for armor fans world wide. While raising the bar for armor simulation in general, these sims universally pay more attention to the kind of issues that armor fans deem critical, and they also address many areas of game play that have been somewhat neglected. With new depth of AI and new levels of battlefield integration, armor fans will have more choices and more fun this year than ever before!

Check out chapter one of the Tank Platoon Manual.