The following is part I of a series on the design concepts and thery behind a game on trench warfare in the Great War. The game itself will follow as the final installment.

Veteran gamers will no doubt find similarilities between TRE'NCHFOOT and other-games, and nitpicking technical experts will probably find factual errors (they always do). But the purpose of this series is NOT to present a spectacular new game full of original ideas (it DOES contain some new concepts, such as the gas attack rules); rather to detail the development of a game as a model of 'reality' from conception to playtest edition. The emphasis is on WHERE the game from, not on the game itself - a game which recreates battles in that most horrible of wars where a soldier's two constant ccmpanions were flying metal, and TRENCHFO0T.

THE DESIGN CONCEPTS IN TRENCHFOOT

The play of any game, the outcome of ary simulation, is dependent to an overwhelming degree on the design concepts and assumptions which went into it. This is relatively obvious when one looks back upon the history of the uses and abuses of games and simulations. But on the other hand, one cannot create a model without making assumptions. Thus any simulation is bound to have many assujmptions--both implicit and explicit.

Simulations are generally created to mcdel complex situations Thus either the simulation itself must.be prohlbitively complex (for our purposes) or many simplifying assumptions mist be made.

But these assumptions mat be made with two factors in mind: historiccal accuracy and simplicity (otherwise known as "playability.") Simulations on this level at least nust be as simple as possible to allow two average individuals to understand it and be able to pIay it in a finite amount of time, maintaining their interest to the end. But the simulation loses its value as an educational tool (and as a model of accuracy not accurately reflect the model of reality). So, compromise must be reached. This is where Design Theory comes in.

Many assumptions are made in designing Trenchfoct. All are evident to some degree in the play of the game. Many of the assumptions are interrelated and each was made to solve some specific "kink" in the model.

Overt Assumptions: Restriction of the Model

1. All nationalities will be treated identically.

Except for the optional rule, there is no differentiation between nationalities in Trenchfoot. This in clearly an unrealistic assumption for each nation trained and equipped its soldiers differently, and each nation had a different motivation for fighting--along with its implicit amount of motivation (othervise known as morale). British troops were known to be able to absorb and inflict more punishment than their French Allies. German artillery van known to be more efficient, as was Gorman organization.

But modellling this fact actually presents a very difficult task far out of proportion to the realign it adds. Few casualties were caused by Rifle fire. Most were caused by artillery and many by machine rune, All things taken together, there was very little effective difference between a French or British or German machinegun or by 1916, a French or British or German machinegunner. Presented with a target and with orders to fire, all reacted similarly -- artillery can also be similarly equallized. While German ammunition might have been better or Krupp guns more efficient, Allied armies might have more ammunition or puna to make up for it. The most realistic simplification of all In the Infantry. While the nationalities might vary in combat effectiveness. their use of the Infantryman essentially the same: as cannon fodder. 1n 1916 infantry training had been reduced to a minimum -- it was necessary only to carry a rifle, bayonet, and to charge the enemy trenches. The infantrymen was expendable.

Differences in combat effectiveness could have been shown by varying the combat factors. But since these variations were usually not major, a now scale for combat factors would have to be introduced. For example, a scale from 1 to 10 could be used, with the German Assault Factor being 8, the Britich 7 and the French 6. But In the long run, this variation has little effect on the play of the game, (for the variation is too alight to affect the odds ratios), and the larger factors are unwieldy to handle.

Thus, except for gross variations in combat effectiveness, which would show up in actual play of the game. It is unnecessary, and adds little to the model, to consider these other variations.

2. All units of the same type are the same

In actuality, each unit of troops is heavily dependent on the men it is composed of. Some companies my have more group spirit than others. Some will be larger and others smaller. Some may be gifted with a larger than normal allotment of welltrained man. But again, in a war when man are used as cannon fodder, and in a simulation where there are many units on each side, such differentiation adds nothing at all to the modal. except needless complication.

3. Ammunition

No provision was made in Trenchfoot for ammunition. It is assumed that resupply takes place automatically if it is needed. Once again, this is a reasonable assumption, for the typical Trenchfoot scenario covers only a few hours of real time, and units are not "firing" every turn.

4. Terrain

It was assumed that there are really only seven types of terrain, and that they are (mostly) mutually exclusive. that is that there in only one type per hexagon. It was necessary to limit terrain to one type per hexagon to maintain simplicity. This actually involves minor distortion of geography to conform to the hexagonal grid. Limiting the number of types of terrain is only somewhat unrealistic, for other types of terrain can really be approximated by one (or more) of the types given. (Certain types of terrain such as marshes were excluded from this model to keep it simple.)

5. Exhaustion

Because the simulation covers only a few hours of time, exhaustion is really not a crucial factor which occurs in this time interval. Actually, however, exhaustion was (believe it or not) built into the model.

Exhaustion affects combat efficiencyand morale. It was assumed that exhaustion had already set in (as indeed by this phase of the offensive, after months in the line and several days of artillery bombardment would be the case), and thus units are reduced in efficiency by this effect. Further, the relative ease of 'eliminating' units is also due to the exhaustion of the men which reduces their resistance to disintegration due to panic and loss of morale. An optional rule permits relatively 'fresh' troops to be used.

6. The Grid System.

For mapboard. simulations, there are two forms of movement possible: free movement and grid movement. Free movement involves the use of a map with no grid movement is done with a ruler, (which is also used to determine firing ranges and lines-of-sight). The Grid System involves the use of a grid to mature movement range, as in Chess.

Free movement is usually preferred for the High Level Government and Military Simulations. for there in no distortion in movemant, and any nap of suitable scale can be used. This system, however, presents many problems to the amateur. and special problems in two player game design. Although the absence of a grid allows units to be moved in any direction without distortion at their exact speeds (to the "scale foot,* oven) it also allows a good many vague situations to air--units may be half in one type of terrain and half in another, for example. An argument over a sixteenth of an inch (in range or movement) often results when two players are deeply involved in the game.

The Grid System, on the other hand, allows a more rigid net of rules, essential when two players compete, but at the expense of some realism. The nap may have to be distorted to conform to the grid, so that the terrain in each polygon Inexactly dotermined. Movemant in polygons is therefore restricted to multiples of the length of a polygon (20 motor polygons results in all movement as multiple of 20 motors, for example). And distortion results when moving from polygon to polygon for the route between two points my not be a straight path.

Once the need for a grid system is established, the next problem to to determine the type, of grid (determined by its polygon). It is easily established that the grid pattern must consist of regular polygons. Consider one polygon. If the polygon to not regular then the distance across it in one direction will be greater than in a different direction. Since distance will be measured using 20 polygons as the distance of measure, it in important that x polygons represent the same distance regardless of how it is measured. Thus, if x polygons represent 1 kilometer, each polygon must be 1/x kilometer across since the same number of polygons must represent one kilometer in every direction, each polygon must be 1/x kilometer across in every direction. Thus, it must be a regular polygon.

The simplest case is a two sided polygon. Since this just a line segment, it obviously cannot be used (it his no area).

The next case is the equilateral triangle (sea Figure A). It can only be shown that this results in gross distortion. Let the height of each triangle be h. Then the length of any side is h [square root] 3 (or approximately 1.7 h). Allowing movement through the vertices, the distance from triangle A to triangle B is 2h through 1 and 4 to B, or 'three triangles". The distance from A to C is also 2h (proof left to reader) but to got from A to C a unit must move through triangles 1, 2, and 3. If diagonal movement (i.e. through the vertices) is not allowed, distortion is even worse, for the distance from A to B becomes 5 triangles, while the distance is still 2h. Further, the distance from the center of triangle 1 to the center of adjacent triangle 2 is h, but from the center of triangle 1 to the center of triangle 4 is 1.7 h, yet both are represented by one polygon (triangle) of movement.

The square is the next case. As any checkers or chase fan 'knows, there is also distortion in this system. Moving diagonally, one can get from square A to square B through square 4 alone, while if diagonal movement in prohibited one want move through squares 1, 2, and 3. The distance from the center of square A to the distance of square 4 is 1.4s, while the distance from the center of square A to the center of 1 is s and both are represented by one polygon (square) of movement.

What is essentially the problem is the difference in distance from side to side (of the polygon) versus from vertex to vertex; and the variation in "diameter". For a triangle it varies from h to 1.7h For a square the variation is from a to 1.4s.

Thus the best polygon for a grid would be the one whose diameters are all equal (or, in other words words, the distance across the polygon in any direction is the same). This suggests circles, however, circles will not form a grid unless they are "squished" somewhat so that, no holes exists. Now we can eliminate those portions of the circles which overlap (since they are confusing anyway) and approximate the air by polygons. The result is a hexagonal pattern. An analysis similar to that done for the triangle and the square will show that there is almost no distortion at all, since no movement passes through a vertex. The distance from A to B through 1, is almost identical to, the distance from A to C through 1 with a distortion (from center to center) of less than 11%, far 1ess than the 40% using squares or the 70% using triangles.

We thus adopt the hexagonal polygon for our grid, system.

DESIGN THEORY AND IMPLICIT ASSUMPTIONS
CONSTRUCTION OF A MODEL II

1. a) Combat level--Area Size

Trenchfoot is a model of trench warfare on the Western Front 1916-1918 on a tactical level. That is, it is a model of the lowest possible level of combat which will still present the complete offensive.

Offensives in the Great War were usually conducted along a kilometer front, and rarely penetrated enemy territory more than or 6 kilometers. Thus, a study of a typical offensive can be limited to an area 25 by 3 kilometers. Further, the offensive was generally similar all along the front; rarely was the offensive conducted differently from kilometer to kilometer; most often men were lined up along 2.5 kilometers of front and pushed forward. Thus, it in rot necessary to study the offensive along the whole front, but only portion of the front, for what is happening an this portion if the front is occurring elsewhere.

So it in possible to study the offensive in depth by concentrating on only a small portion of the front. But the area we concentrate on must be sufficiently general to reflect the whole offensive. So although we can concentrate our study to only several kilometers of front, we must retain the depth of several kilometers, to be able to investigate all impacts of the offensive from its launching in friendly trenches to the stalemate several kilometers into enemy territory.

Thus. the area which the map must represent should be at least 3 kilometers wide by 5 or 6 Kilometers deep. This permits sufficient concentration on a small area without losing any generality.

b) Combat level-- Unit Size

Since we are dealing with a tactical level simulation, it would be wise to use tactical level units. As we would not study individual men when investigating the whole war, we should not use units too large for our tactical study.

One basic tactical unit in the Great War was the battalion of between 600 to 1000 man. There were three such battalions to regiment, six to a brigade, twelve to a division, and 24 to corps, which was the basic strategical maneuver and combat unit. However, our combat system dictates that we be able to break down the battalion into smaller units. Thus we arrive at the basic unit of company size (approximataly 2OO men). <>c. Combat Level--Time and Distance

Having arrived at a decision for unit size and, the site of the territory to be covered, we must now deal with the harder problem of time and distance. To arrive at a decision for this problem, we must take several factors into account:

(a) the "real time" length of an offensive
(b) how far units can move in a given time interval
(c) movement factors should be kept within reason i,-.e. small)
(d) weapons ranges
(e) unit sizes
(f) board size (in hexes)

Great War offensives were generally fought over a period of weeks, or even months. The crucial points of the offensive, however, were generally measured in days. Further, it might be wise to focus on various stages of the offensive individually. instead of the entire offensive from start to finish. This would leave out most of the dull and unimportant periods but, still leaving the crucial periods. not lose any generality. Thus the tine interval for each point of focus (a 'scenario" could be measured in days. or even hours.

We must now take into account the movement rates of the units. Infantry could generally walk three kilometers per hour, one and a half to two kilometers in combat. Thus, if each hexagon is two kilometers across, then units could move one hexagon per hour. Since we want movement factors reasonable (4 or less from previous experience) each turn should represent two to four hours.

The time distance factor in more easily determined when the problem to be simulated in more clearly defined. For example, since we will want to focus on the initial phase of the offensive, (the assault across no-mans-land) in detail, we want the units to represent several turns-in no-mans-land. Since no-mans-land was generally about 400 to 800 meters across. we will want units covering no more than 400 meters per turn. Thus each turn could be 20 minutes and each hexagon 200 meters across, giving us an infantry movement factor of two hexagons.

Weapons ranges must also be taken into account. Units should not be able to move from "out of range" to assault range ( contact) without being shot at. Due to our combat procedure, it will be necessary to have units take at least two turns to move the distance that weapons can shoot (so that they will be targets for one turn at least.)

Since infantry weapons and machine guns have ranges of 400 and 600 meters, respectively, units should move no more than 400 meters per turn.

Unit size must also be take n into account. Each company (of infantry) requires between 80 and 160 meters of front (in a loose skirmish line) , so each hexagon must be at least 80 meters across. This is essential if we are to use a grid system, so that each unit will be in one and only one hexagon at any given time.

Finally, the size of the board should be kept in mind. The board is to represent approximately three kilometers by five kilometers of territory, We want to keep the board small enough to fit on a table (we would not want each 5/8" hexagon to represent one meter (across), yet we will need enough hexagons to deploy at least a corps (100 units) without overcrowding.

With all of these constraints, a scale of 100 meters (across) per hexagon seem best. Weapons range would then be 4 to 6 hexagons and no-mans-land would be four to eig.ht hexagons across. Since we want infantry to move no more than 400 meters per turn, they are to be given a movement factor of four hexagons (400 meters) per turn. Each turn thus represents approximately 12 minutes of real time. This permits us to examine the offensive in great detail, once every twelve minutes.and by each 100 meters. The board would then be about 30 by 50 hexagons, which measures 23" by 28" -- good size for a board.

To be continued