, , , , , ,

In my previous post, I found that the Theatre of War 3: Korea package didn’t quite measure up. Yet it seemed close enough that perhaps there is a possibility of redemption. Many of my criticisms were a function of scenario design, and so I wanted to look into putting together a historical scenario and if that improves my opinion.

To help out, I go to a game that, despite its age, is still one of the best treatments of this level of tactical warfare available. I am speaking of Steel Panthers. In one of its current incarnations, Steel Panthers: Main Battle Tank, the second scenario is a potential test for for my ideas.

The battlefield focuses on a ridge between two low-lying areas of rice farming. The North Korean invaders had crossed the Naktong River, where the armies of the South were holding a defensive perimeter centered at the city of Pusan. The ridge at Obong-Ni was a natural defensive point and became the focus of intense fighting over the course of several battles in August and September of 1950. From a wargaming standpoint, these battles were an inflection point, where the U.S. Marines began turning their desperate defensive position into an offensive to retake South Korea.

The particular focus of this scenario on the first hour of the North Korean counter-offensive on August 17th. The communists had overwhelming numbers as well as superiority across different weapon classes (advantages in artillery and armor as well). The advantage wasn’t to last. As the US was able to bring up both superior numbers and weapons, they were able to overwhelm the Korea positions and force them back across the river within a matter of days. In this brief snapshot of the battle, however, the Marines were attempting to turn back an assault with T-34s using bazookas and Recoilless Rifles. However, the U.S./UN had air superiority. The Marines’ advantage is that, while they are likely unable to stop the enemy tanks on their own, they can call in airstrike after airstrike.

Pixelated Soldiers of a Forgotten War

Amazingly, before this exercise, I’ve never played Steel Panthers. The game itself is old enough to drink, having been released by SSI in 1995 (see timeline here).

I think I initially didn’t get it because of the price. Later, it began to look dated and criticisms of the computer opponent and “gaminess” deterred me. Even when the price was “nothing,” there was some on-line arguing about the functionality of the game and whether it actually ran properly. Long story short, I never got the game until just now when I wanted a comparison scenario.

I’m playing with the free version, which has a limited resolution. The choice is either to have a very small window that graphically looks decent, but is kind of hard for me to see text and other details – or to run it in full screen mode where, to quote someone far more articulate than I am, it “looks like every element on the screen was printed with a boiled potato.” I haven’t done enough to really evaluate the AI, but I’ll assume from the on-line criticism it remains a weakness. Nevertheless, in many ways Steel Panthers has yet to be surpassed for what it does. It provides a highly historical (see previous) treatment of small-unit action from World War II through to the present. It does so with a wealth of available scenarios, a battle generator, and editing tools that are limited only by the skills of the program’s fans. I won’t dwell too much on either the game system or this scenario, but I’m sure I’ll be back to it later.

I will mention a couple things that really stood out as I played this. The system seems to have struck a pretty good balance between simplicity and realism. The number of “attacks” per turn feels pretty decent, while “defensive fire” is completely automated (subject to user-defined parameters). It’s slow compared to the more-modern “real time” systems, but better than most out there. Second – smoke. The computer uses smoke to mask its movements both retreating and attacking. This is one of the few games I’ve seen it used properly. I also like the mechanic where sustained firing on a defending hex creates smoke, which then obscures future shots into and through the hex. I’m sure it has been done elsewhere, but it hit me as something I hadn’t seen before.

As to the details of this battle, I found an archived turn-by-turn description for your viewing pleasure.


Observation point, village center, and the northern point of the defensive ridge. The graphics look much better in windowed mode than distorted to fit a full screen at modern resolution

Tank Country

Initially, the UN troops found themselves facing North Korean armor with infantry, generally a distinct disadvantage. One of the reasons is that the American high command didn’t consider Korea to be “tank country,” and so neglected to deploy armor formations. The necessity of countering T 34/85 attacks made them rethink that position.

The first step in creating the battle was to recreate the battlefield. One of the reasons I chose this particular battle is there are multiple sources available to recreate the battle down to the level of detail necessary for the Theatre of War engine.

My first thought was to attempt to edit terrain in one of the existing maps. The game comes with a map editor, which appears to support tools for everything from creating a new map from scratch to editing an existing one. Unfortunately, there is no documentation so I simply tried everything. As far as I can tell, the basic terrain cannot be modified. Terrain height and type appear to be fixed on the provided maps, essentially limiting all scenarios to the nine provided maps. This also seems to apply to the trees, for which an extensive set of menu options exist, but all appear to do nothing.

What is editable are the roads, and what they call “statics,” basically buildings and trenches. Add to this that the each map is actually larger than the playing area, and there is some variation to be had. The terrain can be repositioned within the battlefield window and then the villages, roads, etc. moved to create a wide variety of setups.

So my next step was to find some terrain that seem to approximate the battlefield in question, and then reposition roads, trenches and buildings to get something semi-historic. The process was tedious, but not impossible.


The battlefield taken from a military map from the period. The ridge is the series of crests following the dot-dash line passing diagonally through the center of this picture.

A sketch of the battlefield showing troop positions and movements. Note that this map is rotated about 40 degrees from North-South so that the ridge is straight up-and-down on the map. Also, the arrows depict troop movements from after the scenario we are modeling. (https://www.koreanwar.org)


Google map version of roughly the same terrain as in the first map is enlightening about the actual terrain. Ridges alternate with low-lying rice-farm valleys.


Steel Panthers scenario, initial positions, zoomed out to match the other views.

With all this material to work with, I thought I had done pretty well at reproducing it on my TOW3 map.


From the air, I felt I had the gist of the battle taken care of, if I ignore the mountain where the second ridge should be.

Once I zoomed into the ground level view, however, it was clear to me that while the map may look fine, the terrain I was working with was far from what I was trying to model.


Down in the trenches, that isn’t really much of a ridge, now is it?

What looked like a good-enough approximation of the historic ridge line was really just a patch of rough terrain that really did nothing to block sight lines. The entire battlefield has pretty good Line Of Sight from one end to the other, meaning that as soon as I start running, everyone starts shooting at everyone else.

The TOW3 scenarios I’ve played start out with opposing forces separated by those large mountains you see in the background. If everyone starts out on the same side of those mountains, there does not seem to be much to limit contact.

Still, this is a learning exercise so, aside from the lost effort of placing all those trenches, what else did I learn. Is there still hope?

The next step, having created my terrain, was the placement of the forces. I started with the American side, and was able to mostly recreate the forces defending the ridge, as provided by the Steel Panthers order of battle. I left out the reinforcements scheduled to arrive through the scenario and also limited the off-board artillery and air support to what fit into the TOW3 editors parameters. But having got things pretty close to how I wanted them, I moved on to set up the North Korean attack.

And found that any more than about 3 tanks and a couple of mortars maxed out the unit allowance.

I’ve seen video of large numbers of units on screen at any one time. So the unit limit is likely in the editor and not the game. The game comes with, in addition to the map editor, an “Editor” and a “Simple Editor” for creating scenarios. I’ve used the “Simple Editor” in this. The battle itself is saved in an XML file, so there seem to be ways around the limit if needed.

But again, still a learning exercise. So I decided to scale down the scenario to the first 20 minutes, and exclude everything that wasn’t either initially on the front lines (for the U.S.) or in the first wave (for the North Koreans).

Even with my best efforts, I still couldn’t populate the North Korean side with enough for their human-wave style attacks. The key challenge of this scenario, however, are the North Korean T-34s, so I focused on getting the tank and mortar count right.

Upon running the scenario, the terrain problems were immediately apparent. I had placed the North Korea infantry at the far edge of the map, intending them to provide indirect fire (as in the SP scenario). I did not build any defensive system for them, which meant they were immediately targeted and fairly quickly destroyed. Absent any ability to alter terrain, it may be possible to shield indirect fire units with a combination of trenches and buildings, which can be moved in the editor.

In the Steel Panther’s scenario, the infantry advanced to a point where smoke could be targeted, which then covered the final advance. In the TOW3 version, the advancing infantry was both fewer in number and without terrain cover, and was eliminated before closing. The armor, on the other hand, had similar feel in both games – pretty much immune to infantry. One different was the U.S. recoiless rifle. Not effective against tanks in SP but deadly in TOW3.

Overall, the inability to add off-board support except through “purchases” limits the ability to get the historical situation correct. U.S. superiority via artillery and in the air is a key component of any Cold War or Modern conflict. On board artillery is an option, although that exacerbates the until limitation issue. In my setup, I added two 155mm howitzers to substitute for the missing off-board options. It should have been two 155mm batteries.

Furthermore, the on-board artillery seems to have targeting problems. As nearly as I can figure, the problem is that artillery cannot fire indirect under several circumstances. If the guns are loaded with direct-fire ammunition (i.e. AP shells), they cannot target indirect fire until the loaded ammunition is discharged at a direct fire target. Secondly, if the guns can shoot at targets directly, it seems that they will not fire indirectly. Or maybe not. Whenever I think I have it figured out, I seem to quickly find holes in my explanations. I can say it is very difficult to target the enemy with indirect howitzer fire, although indirect mortar fire and off-board artillery fire will work in the same circumstances.

The manual is of limited use. There are a few bits and pieces of information, but not the kind of detail necessary to resolve the problems I have. The manual seems primary written to explain the intricacies of armor penetration modeling. I assume the details are because the models are, in fact, implemented in the game, although even that wasn’t entirely clear. While it is nice to know that your game has high-fidelity models, the actual effect on gameplay (the experience of playing) would seem to be minimal.

Other Battles, Other Shortcomings

Following my Obong-Ni Ridge scenario, I looked at some of the other Steel Panther scenarios created for this same time frame. That is, the beginning of the U.S. offensive in the fall of 1950. I have not yet tried to implement matching TOW3 scenarios.

A few parting thoughts.

The units and weapons available in TOW3 are “representative” rather than exhaustive. Meaning, creating a historically accurate battle will almost certainly involve some substitutions. In all of the SP scenarios that I’m looking at, the U.S. Marines are using the M26 Pershing. This vehicle was already out-of-date at the start of the Korean War, but as the Marines were caught with a shortage of armor (because Korea wasn’t really tank country), they were forced to use what was available. In TOW3, one can use the M4 (Sherman) or the M46 (Patton). Only.

One of the available SP scenarios is very similar to what the automated Mission Builder will create in TOW3. The U.S. is move a small, mixed armor and infantry force along a road with poor sight lines and a village. Along that road, the North Koreans have units set up for an ambush. The U.S. has no air or artillery to call in. If there were a bit more customization capability to the TOW3 Mission generator, a satisfying approximation could be created. However, if I want strictly limit the types of units, I’d probably need to create the mission in the editor. I have yet to do so.

Another one of the SP missions involves repelling an assault on the recently recaptured Kimpo airport. The Steel Panthers documentation is lacking in this case, but it would seem to be a night action. Visibility is restricted to 150 yards, meaning the large flat ground of the airport does not allow engagement at a distance. This is another shortcoming of the TOW3 engine in which the variations in weather conditions seem, to me, to impact only the lighting effects rather than actual game play.

The last scenario I tried in Steel Panthers was simulating an island landing, securing the harbor in preparation for the main Inch’on landings. It served as an illustration of how much is missing from TOW3. The scenario uses landing craft to place the U.S. Marines on the board and massive naval artillery bombardment to prepare the ground. The available air assets exceed my ability to properly use them, and there are also spotter aircraft both off board and on board (a helicopter). On the North Korean side, there are a variety of defensive positions, bunkers, caves and other cover as well as terrain that prevents the Marines from engaging at a safe distance. The size of the battlefield and the numbers engaged also clearly exceed what I could do with TOW3.

While I wouldn’t expect TOW3 to do all these things, its inability to do most of them points to why I have to, despite giving TOW3 a second chance, once again withhold a mark of approval.