Sunday, December 24, 2017

The Tactical Uncertainty Principle

Advanced Squad Leader did an excellent job with morale. Previous editions of this particular game manfully attempted to address communications and control, albeit in a clumsy manner that was not very gameplay-friendly. But few, if any, games, have successfully addressed the issue of information in war on the tactical level. What passes for “fog of war” is largely limited to something like the ASL concealment mechanism, in which line-of-sight immediately results in full and correct disclosure of the enemy forces. And while this makes sense in light of the obvious limitations of tabletop gaming, it is somewhat strange that computer wargames have not substantially addressed this issue to the best of my knowledge.

So, my thought is this: what if line-of-sight provided unreliable information in the manner Clausewitz describes? What if one benefit of unit experience was the ability to correctly identify enemy units as well as their quantity? There is ample historical precedent after all. One need only read a history of any post-Normandy WWII battle to realize that American troops were prone to misidentifying almost any German tank as a Tiger.

To most American troops, any big tank trying to kill you was a Tiger. Admittedly, the sloping frontal armour of a King resembled that of a Panther Tiger, but a soldier in trouble doesn’t stop to count the road wheels or turn to his recognition handbook. Similarlly, many Allied memoirs talk of being under fire from ‘eighty-eights’, whereas msot German field and anti-tank guns were of 75mm calibre.
Snow and Steel: Battle of the Bulge 1944-45 by Peter Caddick-Adams

The mechanism is relatively simple to implement. For each unit, there is a short catalog of mistaken quantities and types based on probability. One is much more likely to mistake a 75mm field gun for an 88mm than for a 37mm, after all. It is obvious that troops tend to overestimate their enemies rather than underestimate them, both in terms of quality and quantity, so the tendency should be to err upwards. Both the visual and statistical information reported to the player will be inaccurate, then updated as other units make contact or sustained contact causes the initial units to correct their initial error.

It seems to me that this could be as important an advance in wargaming as the ASL morale model was. So, what do you see as the likely costs and benefits of implementing an Unreliable Information model, or if you prefer, Tactical Uncertainty principle?Just because a number of gamers and game designers have expressed interest in it, I’m going to run through my thinking on the subject in public. The more I’ve thought about it, the more I conclude that this could be the most useful new wargame design mechanic since John Hill introduced the morale model in Squad Leader. Also, I’m not going to concern myself with how to implement the principle in a board-and-counter or tabletop game yet, since my primary interest is developing a model that will work for Striker (3rd edition) in the 3DV engine, which is to say, on the computer in a 3D tabletop. While the principle remains the same in either case, the precise model of optimal application will necessarily differ.

The basic Tactical Uncertainty Principle, which is based on the Clausewitzian concept of friction in the form of information, can be summarized thus:

  •  Information reported by LOS is unreliable.
  • The reliability of the information reported depends upon the quality of the unit or leader reporting it.
  • Units reporting inaccurate information tend to inflate both the quantity and the quality of the enemy units reported.

Today, I will focus on the second aspect of the Principle, which states that the reliability of the information reported about the enemy units spotted depends upon the quality of the unit or leader reporting it. For now, I will utilize the five-level unit-quality system of ASL. Here is how I envision the application in terms of pre-modification statistics, with the number representing the percentage chance that the unit will accurately identify the unit(s) sighted.

Elite 85
Veteran 70
Regular 55
Green 40
Conscript 25

A unit with a leader of differing quality will identify an enemy unit on the basis of the leader’s quality rating rather than its own. A unit with the same quality leader will receive an addition 5 percent identification bonus. So, a Veteran unit with an Elite leader will identify enemy units with an 85 percent success rate, but an Elite unit with an Elite leader will do so at a 90 percent success rate.


Will multiple units with a LOS on the same enemy unit make independant roles and report potentially contradictory information (and is there a clean UI for showing this)?

Yes. Yes.

How often will units reroll for possible updates?

Once per enemy movement turn.

Can there be multiple units stacked in a given location and if so, can a unit go entirely unnoticed?

Yes, in the case of a board game and no in the case of a miniatures game. And yes.


  1. This would make ambushes really interesting.

  2. There is merit in assigning a value to the experience level of the reporting unit as it deals with battlefield reports and I believe this is a fairly unique concept. Another factor to consider is the robustness (or usually lack thereof) of tactical communication networks. I'm currently playing some France 1940 scenarios and the game engine reflects both superior German doctrine and communications by allowing the panzers to shoot twice per turn versus only once per turn for the superior French Souma tanks. The experienced company commander may correctly identify his opponents but the chance that higher HQ receives this information is enhanced by the number of radio sets in the company (from each tank having a set, to only the company commander's tank, to none at all for early war Russians).

  3. FYI - The "Combat Mission" series from has used a FOW & unit experience system for spotting/ID'ing enemy units for many years. Also take into account terrain and weather conditions.

    Check it out. Maybe something to be emulated.

  4. I've been deeply disappointed in how real-time strategy games provide real-time omniscient data within line-of-sight. The fun and challenge lies in trying to make accurate determinations from imperfect data. Your scheme ought to be easy to implement particularly in the context of a computer game. You might also want to consider incorporating deception - a small unit deliberately attempting to look bigger, for instance, ought to be able to tilt the odds further toward a misidentification by the enemy. Mistakes, misinformation, and deception are an inevitable part of war that I've never seen properly modeled in a game. A parallel idea that would not work well in a board game but could be implemented in a computer game is lines of communication and realistic time delays in the commander/player receiving updates from front line units on their status or observations. There'd be a whole additional level of trying to maintain your own rapid and accurate lines of communication while trying to attack and degrade the enemy's. Maybe the radios are jammed or the messenger gets lost so the enemy achieves a breakthrough and you don't detect it until they're found rampaging in your rear. Or you've jammed the enemy's radios so they have delays and imperfections in their data while your radios give you real time updates. It would add yet another layer of realism to the fog of war to add this sort of informational friction. Happy to discuss this more if anyone is interested.

  5. I've read that just about every German tank encountered was reported being a Tiger.

  6. SPI's Invasion America had counters that showed the type, but you had to engage the unit to flip the counter over to see how capable it was, conscripts, regular, veterans, elite, etc.

  7. I think you need a "fumble" mechanism that not only wildly overestimates a unit strength, but generates a phantom unit. I'm just thinking of a big operational screw up like Operation Cottage.

  8. I am sure you are familiar with Warlords 3: Darklords Rising. In this TBS game a unit can, within its line of sight, only see two things:

    1. The single enemy unit standing upfront in any stack of enemy units. This is determined by the unit fighting order which is configurable on unit level but not on individual stack level.
    2. The total number of units in the stack.

    To gain more information your unit needs to attack the enemy stack. The maximum number of units you can see when attacking/defending is 8. The maximum stack size is 8 units. However, a city can have up to 4 stacks of 8 units defending out of which only 8 would be shown at the time.

    Without attacking a city a unit can only see whether the 4 city towers are manned or not meaning that 4 units can indicate up to 32 units and 1 unit can indicate up to 8 units.

    This system could be vastly enhanced by adding the ability to manipute fighting order, with drag and drop, on the individual stack level. And obviously by making possible the faking of unit numbers within line of sight not only within the cities but also out in the field.

    Just something to consider.