The Fog of War

A key element of General Staff gameplay is the notion of ‘fog of war’. In essence we keep three maps: what blue thinks the situation is, what red thinks the situation is and where all units actually are on the map.

A courier reports that an enemy unit was observed four minutes ago. Note: the enemy unit has almost certainly moved on since then. (Click to enlarge.)

A courier reports that an enemy unit was observed four minutes ago. Note: the enemy unit has almost certainly moved on since then. (Click to enlarge.)

Location of your headquarters (HQ) unit is vitally important in General Staff. If an enemy (or friendly) unit is not directly observable to your HQ unit, the information of an enemy unit sighting is passed to HQ via courier. The amount of time for the courier to travel from the observing unit to the HQ is precisely calculated (see screen capture, above). Friendly units report their position every hour and dispatch a courier to HQ with this information as well. The time for a courier to travel from the reporting unit to travel to HQ is also calculated.

In essence, then, locating your HQ unit on a hill, ridge or other prominent feature is important if you want a clear view of the battlefield. Otherwise, you will be dependent on your couriers to penetrate the fog of war.

The phrase, ‘Fog of War’ (or Nebel des Krieges in German) was first written by Carl von Clauswitz in his famous treatise, On War (in 1832, first English translation in 1873):

War is the realm of uncertainty; three quarters of the factors on which action in war is based are wrapped in a fog of greater or lesser uncertainty. A sensitive and discriminating judgment is called for; a skilled intelligence to scent out the truth.

— Carl von Clausewitz

Update: some readers at Kriegsspiel Forum have quite correctly pointed out that at great distances units would certainly not be able to ascertain the name of the unit that they are observing. Consequently, we have changed the information provided to this:

Dispatch from observing unit modified to include distance and just unit type. (Click to enlarge.)

Dispatch from observing unit modified to include distance and just unit type. (Click to enlarge.)

Now the question arises: how far could you see with a 19th century telescope? Do we need to include a maximum range for line of sight? Interesting question. Would like to hear your comments.

2D or Not 2D That is the Question (Redux)

We have been going back and forth on the question of: should General Staff have a 2D or 3D battlefield? Arguments can be made on both sides. The key points for 3D are:

  • we already are storing all the necessary data (we know the height of every square meter on the battlefield)
  • we have topographical maps that we can use for the 3D texture
  • being in 3D may increase General Staff‘s popularity
01TopographicalMap

A General Staff topographical map in 2D. (Click to enlarge.)

So, after going  back and forth on this issue for many months, we finally did some tests (thanks Ed Isenberg!). And here is what the same topographical map looks like utilizing our stored elevation data:

The same topographical map rendered in 3D. (Click to enlarge.)

The same topographical map rendered in 3D. (Click to enlarge.)

Well, it’s definitely 3D. But, it appears to me, that we’ve lost more than we have gained. The user hasn’t acquired any new information; hills, rivers, ridges and valleys were all quite apparent in the 2D map. We’ve also lost that wonderful historical Kriegsspiel feel of Victoriana typography and cartography. Our 3D display looks like a wrinkled pigskin and not a map that was used by the General Staffs of the great armies of the world. Also, not being in 3D will facilitate porting General Staff to smartphones.

We certainly would like to hear your opinion. But, as things stand now, General Staff will be moving forward in 2D.

Slope Weight Added to Least Weighted Path Calculations

An example of slope avoidance in General Staff. Note how the cavalry unit skirts the edge of the hill on the way to its objective. (Click to enlarge)

An example of slope avoidance in General Staff. Note how the cavalry unit skirts the edge of the hill on the way to its objective. (Click to enlarge)

We have added a new feature, the cost of traversing up a slope, to our proprietary least weighted path algorithm. This will create even more realistic (and, frankly, optimal) unit order paths.  A key element to General Staff is its ability to assist the user in calculating optimal paths for units so the user only has to click where he wants the unit to go and the AI figures out the rest. In fact, the user doesn’t even have to click on the map, but can select the unit’s destination from a list of objectives.

The original least weighted path algorithm with slope weighting was created by Dr. Sidran when he was in graduate school. Dr. Sidran said, “I should probably write a paper describing the new algorithm, called EZRoadStar. However, as I am no longer an active member of academia there is no pressure to, ‘publish or perish’.” Instead, he will concentrate on finishing General Staff.

General Staff is expected to be released for Windows, XBox and iOS later this year.

 

General Staff Gameplay

Screen capture of General Staff showing the 3D Line of Sight (LOS) areas that are visible to the blue HQ unit.

Screen capture of General Staff showing the 3D Line of Sight (LOS) areas that are visible to the Red HQ unit (grayish areas are not visible). (Click to enlarge.)

We are very pleased to show screen captures of General Staff that demonstrate some of its unique gamplay features. General Staff is different (at least as far as we know, and if there are other computer wargames that have these gameplay techniques, we haven’t seen them) in that you, the user, are assuming the role of the commanding general with his general staff. You do not have an omniscient view of the battlefield. The only unit locations (both friendly and OPFOR, or OPposition FORces) that you know for certain are the ones that you can directly observe. The above screen capture displays the areas that the Red HQ unit (U. S. Grant and staff) can see from their location on Riverview Hill. The 3D Line of Sight (3D LOS) algorithm in General Staff uses elevation and terrain maps to calculate the LOS for each unit.

An example of how units that are not directly visible to HQ are displayed. The longer that a unit remains unobserved, the fainter it becomes. (Click to enlarge.)

An example of how units that are not directly visible to HQ are displayed. The longer that a unit remains unobserved, the fainter it becomes. (Click to enlarge.)

When units are hidden from direct LOS they begin to fade (as seen in the screen capture to the right). The longer it has been since the unit has been observed, the fainter the unit is drawn. However, every hour, every friendly unit dispatches a courier to headquarters with its current unit location. When the courier arrives (his path is calculated and described below) the unit’s position is updated and the unit is displayed in full color. In the above screen capture, arriving messenger reports are displayed in the bottom of the screen. Enemy units, however, that have not been observed for an hour disappear completely from the map.

CourierTime

Screen capture showing easy to use interface for giving orders. Also displays unit information, distance from HQ, morale and fatigue. (Click to enlarge.)

Orders are not given directly to units, as in other wargames, but are sent, via courier, from HQ to the unit.  The screen capture (right) shows the interface for constructing a unit’s orders. Orders are entered via unit-specific pop-up menus that specify the formation, direction, facing and speed for the unit. Also displayed is how far the messenger will travel and how long it will take for to deliver the orders. Couriers travel at 17.5 kilometers per hour and will attempt to use roads whenever possible.

This screen capture displays the path from Blue HQ to the unit that the courier will travel, the future orders for the unit and confirmation of the orders. (Click to enlarge.)

This screen capture displays the path from Blue HQ to the unit that the courier will travel (in green), the future orders for the unit (thicker blue) and confirmation of the orders. Note how the courier automatically takes advantage of roads. (Click to enlarge.)

The above screen capture shows the path the messenger will take (in green), the orders (in transparent blue) and a confirmation pop-up that shows the time the messenger will arrive at the unit with the new orders.

We have since updated the Messenger box to look like this (in keeping with the 19th century late Victorian theme of General Staff):

A screen capture of the new Messenger information box.

A screen capture of the new Messenger information box.

We believe that this structure not only is an authentic depiction of warfare in the 19th century and how the generals experienced ‘fog of war’, but will greatly enhance the gameplay of General Staff.

 

First Scenario Created!

ScenarioScreenCap2

Screen capture of the Select Scenario screen from General Staff. Click on image to enlarge.

The above is a screen capture of the Select Scenario screen from General Staff. We anticipate shipping with about 20-30 scenarios and this is the first. This first scenario is an – obviously – fantasy match-up between Napoleon commanding portions of the Young Guard and U. S. Grant commanding the famous Iron brigade supplemented with an extra artillery battery and two regiments of cavalry.

We anticipate producing three or more scenarios for every map that we create. Most of the maps will be of historical battles from the 19th century and before. Are there particular battles that you would like to see in General Staff? Please drop us a line.

This first scenario was created to test the AI (crossing rivers and defending bridges), taking advantage of hills and the road net.

New, Faster Pathfinding AI

A screen shot showing traditional A* (pronounced A Star) pathfinding. The green areas are 'nodes' that the algorithm explored on its way to finding the optimal path (in Brown).

A screen shot showing traditional A* (pronounced A Star) pathfinding. The green areas are ‘nodes’ that the algorithm explored on its way to finding the optimal path (in Brown). Click on picture to enlarge to full size.

Artificial Intelligence (AI) plays an important role in wargame development; it’s what separates a good game from a great game. One of the most important algorithms employed in wargame AI is the A* (pronounced ‘A star’) pathfinding algorithm that was created in 1968 by Peter Hart, Nils Nilsson and Bertram Raphael. The paper describing it, A Formal Basis for Heuristic Determination of Minimum Cost Paths can be downloaded here. I did my doctoral Qualifying Exam on optimized pathfinding. My paper, “An Analysis of Dimdal’s (ex-Jonsson’s) ‘An Optimal Pathfinder for Vehicles in Real-World Terrain Maps'” can be downloaded here.

How long will it take for your orders to arrive?

How long will it take for your orders to arrive at this unit? How long will it take for the unit to send a courier back to headquarters with its current location?

Pathfinding is important in wargames because it’s how units, under computer control, move around on the map. Also, and we’re announcing this for the first time here, when you give orders in General Staff a courier has to ride from your headquarters unit to the unit that is to receive your orders. Also, units on the battlefield that are not directly visible to the Headquarters unit (this is done with a 3D Bressenham line algorithm; more about this later) slowly begin to fade from view on the map. However, every hour a courier is dispatched from every unit to headquarters with an update on their position. As we can see from the information box, above, the courier will take 41 minutes to deliver the new position information to headquarters.

The top screen capture shows an implementation of the classic A* algorithm for calculating the optimal path from Blue’s headquarters unit to a far-flung cavalry unit. Note, this is an especially difficult path to calculate because the unit is across a river and there are only three bridges across. The A* algorithm performs perfectly but it is just too slow to be used with a real-time tactical wargame like General Staff. After some thought I wrote a major optimization of A* which we present here for the first time.

An example of the new EZRoadStar pathfinding algorithm created for General Staff. Compare it to the top screen capture which uses the classic A* algorithm. Click to enlarge.

An example of the new EZRoadStar pathfinding algorithm created for General Staff. Compare it to the top screen capture which uses the classic A* algorithm. Click to enlarge.

Above is a screen shot of the results of the new EZRoadStar algorithm. It is almost identical to the original A* algorithm but runs thousands of times faster (my fellow computer scientists would probably prefer if I did some tests, wrote a paper and published the exact figures and I promise I’ll get around to that, some day).

In the screen shot, above, you can see the path of the courier (in green) from the Blue HQ unit to wayward cavalry unit. The new pathfinding algorithm, EZRoadStar, first looks for roads and then calculates how to get on and off the roads. This is much faster than the A* algorithm.

 

Latest Look at General Staff!

Screen capture of General Staff on April 28, 2016

Screen capture of General Staff on April 28, 2016. Click to enlarge.

We are very pleased to show the first actual screen shot from the development of General Staff. This will give you a good idea of what it looks like and how it works. Currently, General Staff is in 2D. We’ve always planned to have it in 3D but, frankly, this looks so great as it is we’re thinking that maybe we’ll just have a 3D view option. Either way, as you can see the elevation display in the lower left hand corner, we have a complete 3D map of the battlefield ‘underneath’ the gorgeous topographical map made by Ed Isenberg. We will use this data for calculating line of sight and movement.

This is what the invisible 3D height map for the same battlefield looks like:

The 3D height map for the above topographical map. Click to enlarge.

The 3D height map for the above topographical map. Click to enlarge.

Other interesting gameplay features that are visible in the screenshot: there are certain areas of each battlefield that are worth ‘victory points’. This will help establish the goals for the battle. The most important ‘victory points’, however, are the Red and Blue retreat routes.

We anticipate shipping with about 20 different battles from ancient history to the 19th century.

GrogHeads Interviews Dr. Ezra Sidran

Grenadier of the Old Guard in 1813; an original Grognard. Public Domain from Wikipedia

Grenadier of the Old Guard in 1813; an original Grognard. Public Domain from Wikipedia

GrogHeads, a blog for very serious wargamers, recently interviewed our Dr. Ezra Sidran. The interview can be found here. The term ‘grognard‘ (French for ‘grumbler’) came from Napoleon’s Old Guard; who were the only unit that had the privilege of grumbling in the Emperor’s presence. The term grognard is now used to describe very ‘hard core’ wargamers.

We were very happy for this opportunity to talk about wargames with interviewer Jim Owczarski.

Please jump over to GrogHeads and take a look at this interview and look around.

Computational military reasoning.

Computational military reasoning is a phrase that I coined to describe the process of a machine performing human-level analysis of tactical and strategic problems. I have spent the last 30 years of my life working on this problem. It was the theme of my doctoral research in computer science. The abstract for my doctoral dissertation reads:

We present here TIGER, a Tactical Inference Generator computer program that was designed as a test-bed program for our research, and the results of a series of surveys of Subject Matter Experts (SMEs) testing the following hypotheses:

Hypothesis 1:  There is agreement among military experts that tactical situations exhibit certain features (or attributes) and that these features can be used by SMEs to group tactical situations by similarity.

Hypothesis 2:  The best match (by TIGER of a new scenario to a scenario from its historical database) predicts what the experts would choose.

We have conducted three surveys of SMEs and have concluded that there is, indeed, a statistically significant confirmation of Hypothesis 1, that there is agreement among military SMEs that tactical situations exhibit certain features (or attributes) and, that these features can be used to group, or identify, similar tactical situations. The statistical confidence level for this confirmation of Hypothesis 1 is greater than twice the prior probability.

In order to test Hypothesis 2 we constructed, after SME survey analysis, a series of algorithms, which we present here, for the analysis of SME identified tactical features (or attributes) including: interior lines, restricted avenues of approach, restricted avenues of attack, slope of attack, weighted force relationships and anchored or unanchored flanks. Furthermore, the construction, and implementation, of these algorithms, required the design and implementation of certain ‘building block’ algorithms including: range of influence, optimal FindPath, ComputeGroupsByThreshold and ComputeGroupsByNumber.

We further present an overview of TIGER, itself, and the built-in utilities necessary for creating three-dimensional tactical situations, complete with terrain, elevation and unit types as well as our implementation of Gennari, Fisher and Langley’s CLASSIT classification system.

Lastly, we present TIGER’s classification of twenty historical tactical situations and five hypothetical tactical situations and the SME survey results of TIGER’s classification that resulted in TIGER correctly predicting what the SMEs would choose in four out of five tests (using a one sided Wald test resulted in p = 0.0001 which is statistically significant).

TIGER logo from my doctoral research.

TIGER logo from my doctoral research.

The entire dissertation can be downloaded here

I have also written a number of papers about implementing tactical maneuvers: “Implementing the Five Canonical Offensive Maneuvers in a CGF Environment,” which can be downloaded here. And “Algorithms for Generating Attribute Values for the Classification of Tactical Situations,” which can be downloaded here.

What will make General Staff stand out from other wargames is that it will be the first commercial computer wargame to implement this research. I have high hopes that General Staff will have the most advanced tactical AI ever produced in a computer wargame.

2D or not 2D (that is the question).

Let’s just start off by saying that General Staff will be in 3D. It’s the only way to display the blocks that represent units. But the question is: should the map be flat with just 3D unit blocks (simulating the original Kriegsspiel ) or should we employ a technique that I used for a project for the  U. S. Army in which a 2D topographical map was used as the skin for 3D elevation that was extrapolated at runtime from USGS (United States Geological Survey) data?

A screenshot of a project that I did for the Army. Click to enlarge.

A screenshot of a project that I did for the U. S. Army. Click to enlarge.

There are certainly pros and cons for both ideas. Frankly, I like the idea of using a flat, 2D, map with only the unit blocks in 3D. However, the one thing I don’t like about ‘traditional’ Kriegsspiel is that the unit blocks are rigid and always perfect rectangles that do not conform to map contours or allow units to change formations.

On the other hand, I’m concerned that if we go full 3D (like in the above screen capture), it’s going to be to similar to current 3D wargames (I won’t mention names, here).

I’ve tried to keep the overarching theme of ‘simplicity’ for General Staff in clear view. General Staff is supposed to be a fun, simple game where the graphics don’t get in the way of a pure tactical, enjoyable real-time game.

Either way, we will be employing my optimized 3D Line of Sight (LOS) algorithms. That is to say, units behind ridges will not be visible to opponents.

What do you think? Send us a note or leave a reply.