Life and Death Decisions using Sparse, Unreliable Evidence (Information challenges and
mitigations in frontline military environments)
Martin Hill, Dr John Salt
Abstract: Junior military commanders must cope with evidence provided in various forms across
many bearers from several sources. They need to remember and combine long extracts from
briefings, recent radio messages, background contexts and regulations, verbal shouts from nearby
soldiers and heavily constrained senses (due to smoke, noise, sheltering in cover, etc) to form
awareness of the local situation.
Systems have evolved to share information more clearly and less ambiguously in these environments,
and with little extra cognitive load. However, many were developed for regular warfare with well-
defined front lines, and current small wars and multinational counter-insurgency operations have made
some of these less useful.
This paper describes some of these systems, suggests possibly more useful ways to model them, and
outlines issues with ordinary technical solutions. The intent is to describe such an 'extreme'
information management/exploitation environment in order to draw out concepts that are hidden in
more normal environments, and ideally to spark an interest in readers from other domains who could
Keywords: Military, People-based, IM/IX
Information management doesn‟t get much more interesting than frontline counter insurgency
operations. All the ordinary information management and exploitation concepts apply, including mining
huge data sets, finding and evaluating small significant items, challenging performance requirements
and data fusion issues. On top of that the communications are poor, the information technology is
sparse and vulnerable, and the users are usually poorly educated and often emotionally charged.
We focus on the front-line operators; the soldiers and junior commanders on the ground and their
immediate commanders. These are both the end users and the gatherers of the information, yet
usually those with the least sophisticated technology. Manual information systems are therefore
relevant, along with the technology layers.
We are not so interested in the higher command because, on current operations, those military
headquarters tend to be static and sheltered in an office-like environment, with mains power, desktop
computers, local ethernet and good communications with other headquarters. Ordinary information
management and exploitation techniques and technologies can be applied.
Existing ways of managing information have been introduced somewhat ad-hoc over many decades
and sometimes centuries, and they are not always well understood. Similarly, information technology
products have been recently introduced piecemeal and without coordination. It is not always clear
what these solve in practice, so we should be wary about making changes that may have unforeseen
Even gathering information on current information systems and their effectiveness is not easy.
Generally, it is difficult to embed researchers within operating patrols to directly study the operating
environment. Few good researchers are also good soldiers, and the difference in language and
backgrounds can render even direct communication unproductive.
We rely then on discussions with experts (who act as knowledge middlemen), story-based analysis of
the end-users' experiences, and observing those training exercises that veterans describe as realistic.
The military adapts –in various ways – to changing circumstances, so our understanding of how they
manage their information is often out of date and/or contradictory.
We start here with a brief introduction for the non-military reader (which should not be taken as
doctrine). We then look at the main 'operating pictures‟ that are the core of military front-line
information management, and how they are distributed. We finally consider ways in which technology
has helped, or otherwise, this work.
2. Some Military Basics
A patrol consists of a 'Section' or 'Multiple' 8-12 soldiers, and may be mounted in a group of vehicles,
or 'dismounted' when on foot. Patrols are sent out from Forward Operating Bases (FOBs) which are
spartan fortified compounds manned by a Company of around 100 men.
Figure 1: Clearing a compound, Afghanistan (Jamie Wiseman)
Whatever their mission, an important part of any patrol's task is to gather information. The patrol may
last a few hours or a few days, and the distance covered can be considerable if, for example,
helicopters are used. The patrol is mostly self-contained for its duration; it carries its own food, water,
batteries and other supplies. Typical loads, including ammunition and armour, are around 45kg when
patrolling (Devil 2003), or 30kg with only the immediate fighting essentials. Individuals have small
radio headsets with a range of a hundred meters or so to communicate with each other, and the patrol
will have a more powerful radio to reach the FOB.
The FOB acts as a 'node' for supplies and communications. It provides a connection up the command
chain, with access to support for the patrol such as artillery, aircraft and engineers, and to detailed
information such as maps, images, detailed briefing notes from the 'office' environment of the next
While we will concentrate on the interactions of the patrol members and their immediate commanders,
any operation includes interactions with other specialised arms and services. On current operations,
soldiers must also cooperate with foreign soldiers, and civilians from other organisations. Note also
that the above is only a representative example, sufficient for the purposes of this paper.
Essentially, picture the junior commander on the ground, usually in his early 20s, often poorly formally
educated, in command of around ten men, who not only has to make life and death decisions but is
also one of the many eyes and ears of the organisation behind him. That organisation in turn must sift
through what he sees and hears, combine it with information gathered elsewhere, and pass the
resulting assessment back to him to improve his understanding on his next mission.
2.1 Typical Tasks
For peacekeeping to succeed the local population must support it. This requires, amongst other
Continued effective presence, to reduce reprisals from the insurgents. This means being able
to find insurgents and remove them – arrest, destroy, drive off, persuade – in a manner
acceptable to the local population.
Suitable civilian support, so the military are seen as positive benefits rather than remote and
superfluous or, even worse, an aggravation. Understanding the needs of the communities
requires engagement; eyes on the ground, discussion, cultural awareness, and continued
monitoring of the results.
These require quite a rich understanding of the overall situation that is beyond the stereotyped
'enemy/not enemy' combat engagements. And this understanding has to be gathered and distributed
across poor communication links, and stored and analysed and understood with little in the way of
computing equipment (See also Hall & McChrystal (2009), Gant (2009))
Day to day tasks might include:
Exploratory patrols: going over the ground, often on foot, making contact with civilians, being
available for civilians to make contact (as Fig 2 above), looking over areas.
Vehicle Check Points: road blocks are erected and vehicles checked for contraband and
Strike/Arrest patrols: specific missions to find and arrest particular targets.
There is also the daily grind of personal 'admin' (eating, sleeping, washing) without everyday
conveniences, and the maintenance of the group living area and equipment (keeping equipment and
living space clean, manning guard points, escorting supply convoys).
Figure 2: Assessing a school, Iraq (Trooper M Hill, 2003)
2.2 Example scenario
Again, purely to provide background concept rather than describe doctrine, we give an example
The patrol is driving along in three open topped vehicles. An Improvised Explosive Device (IED)
detonates under the lead vehicle, disabling it and throwing the soldiers onto the side of the road. The
other vehicles pull up into whatever nearby partial cover can be found, and the soldiers dismount.
Bullets are 'cracking' overhead and ricochet from the ground.
All the soldiers are now looking for the ambushers; the wounded will have to wait until they have
suppressed the enemy or they may simply create more casualties. Which means there is considerable
pressure to locate them quickly; but modern cartridges create very little smoke, the sound of firing is
hard to locate in the noise of the passing bullets, and the ambushers are expertly concealed.
Examining potential points of cover using telescopic sights loses peripheral vision. Rules of
engagement may require clear identification of armed men before the soldiers can open fire, and the
enemy know this.
The next actions depend very much on the commander. His own commander at the FOB must be kept
informed of the situation, with suitable detail. The wounded must be brought into cover and emergency
treatment provided. A place for the casualty evacuation helicopters to land must be found and kept
safe. He and his soldiers will try and locate the ambushers, which may take hours and thousands of
rounds of ammunition, and then attack, call in indirect fire (such as aircraft bombers) or withdraw. All
this time he must ensure that he and his men maintain awareness all around.
Figure 3: Dismounted patrol leaving FOB in foreground (Trooper M Hill, 2003)
3. Tactical Picture
Military organisations use several terms with overlapping and confusing meanings for the relevant
tactical information. For the purposes of this paper, we will use the following:
“Tactical Pictures” refer to the tactical information relevant to the tasks and operations in hand. Such
information models cannot be perfect or complete, and attempts are mostly directed towards
establishing the 'best we can do' rather than even 'good enough'.
“Situation Awareness” (SA) is the understanding that the soldier has of the relevant things that are
going on around him, and the intent of his team to change it. This will be informed by what he sees
and hears, by reports and instructions, and by reference materials such as pictures and maps.
Junior soldiers are usually occupied with the immediate ground around them; the 'local SA'. It is the
commanders who are concerned with the wider picture.
This wider picture includes absolute geo-referenced items such as grid references of points of interest,
out-of-date friendly positions, uncertain and out-of-date enemy positions, areas of responsibility and
shared reference points. Also relevant are non-geographic items such as the commander's intent,
“combat markers” (signs that ambush might be imminent, such as no civilians present when they
would be normally), local culture and allegiances, and the condition of his men.
3.1 What's interesting about this picture? Unfamiliar Environments, Background & Context
Generally speaking, when we communicate a situation as quickly and accurately as possible, we need
to extract the things that the other person does not know about it, and tell them only that. If it is not
clear what the other person knows, this takes longer and less reliable.
We need to use words that distinguish between the background and the items of interest. Apocryphal
stories abound of today's soldiers, from mostly urban backgrounds, struggling to describe the location
of a target in woodland: “He's beneath that tree!” “Which one?” “That one! That one there! The green
So it is important to establish a 'normal background' as quickly as possible when deployed to a new
area, and to build a shared vocabulary that refers unambiguously to common features.
3.2 Picture-Centric Blue Force / Red Force
The 'traditional' view of Situation Awareness at a command post is a map, marked with locations of
friends, enemies, areas of responsibility and various common reference points. Many information tools
used by the military – computer based and training and process – are based around these simple
friend/enemy concepts (See Figure 4)
These would benefit from ways to show doubt and uncertainty, but anyway have proved largely
insufficient for the more subtle and varied information required during counter insurgency. Soldiers
and commanders have reverted to ordinary English language to describe the situation to each other.
3.3 Purple Pictures
Counter insurgency and policing operations require a richer picture not least because it is not so clear
who is „friend‟ and „enemy‟, and this changes over time.
Killing suspected insurgents may result in the local picture becoming more „blue‟ (friendly) as the
indigenous population is released from the terror of what are effectively local armed bandits. Or it may
result in it becoming more „red‟ (hostile) as relatives of the killed take up the feud, or property or other
lives are destroyed in the operation.
We suggest the concept of a 'purple picture' to easily grasp the changing nature of loyalties. It
suggests a way of approaching loyalties and how the situation is modified indirectly by activities.
The enemy does not typically consist of a single organisation, and the organisations may vary in their
hostility. The “$10 Taliban” are local youth hired during the agricultural down-time, motivated by a bit
of cash, sometimes the fun of it, often intoxicated, usually enthusiastic but very poorly if at all trained.
Being local, social pressures affect these latter both as encouragement and discouragement (for more
on such social pressures, see Potts and Hayden, 2008). The more dedicated insurgent also comes in
several forms; Kilcullen (2009) distinguishes between those who fight to remove the presence of
foreign armies or remote political power, and those with global causes to pursue.
All these are relevant to how a local commander approaches local populations, and indeed whether or
not to pursue and attempt to destroy the enemy once engaged.
4. Building the Picture
The British Army trains its soldiers together in the formations in which they will deploy. This makes
each soldier familiar with the dialects of the people he will work with, and the informal language and
jargon that is used to describe the battlespace in that community. Some of that language has grown in
that community over many years, and more will be developed during specific training to create terms
with common meanings, reducing ambiguity and improving speed of description. It is important
therefore that training matches operations to ensure the terms are relevant.
We suggest that a formal battlespace description language (including graphical elements), extending
and tying together some of the existing description methods, would improve interoperability between
communities, and would further reduce ambiguity and improve the speed of description.
The main form for disseminating information in the army is the briefing. The team is assembled for a
Figure 4: Simple 'Blue Force / Red Force' Picture
single transfer of information, such as background information or orders for a mission. Orders are
given in a structured form with standard headings to help direct attention, and mock-ups are
encouraged to aid understanding (see Figure 5).
Commanders are issued maps which will then be marked up with relevant geo-referenced information
such as known friendly units and bases, common reference locations, and areas of responsibility.
Other soldiers are another important source of tactical information; what they have seen, what they
have done, what worked and what did not.
On patrol, members communicate by voice, hand signals, and short range radios. The commander
can be updated from the FOB using radio, but the limitations – poor quality, intermittent connections,
dangers of discovery – mean that rich information cannot be passed this way.
On return the patrol commander writes his report while his memory is fresh but also necessarily when
he is fatigued. The report is sent up the command chain to intelligence units to analyse, which is then
used to inform future briefings. This separation of intelligence from direct discussion with the
commander is not ideal.
The FOB commander has faster radios with more reliable connections, but large information sets such
as maps can be put on CD-ROMs and USB sticks and physically transferred.
4.1 Target Indications
Urgent, accurate picture transfer is required when coming under fire; consider our sample patrol above
that was ambushed. The enemy must be located and correctly identified to avoid friendly fire or civilian
casualties. That information must be accurately passed around the soldiers that are scattered about in
various cover with different views on the surrounding area.
The British Army has a set procedure ('drill”) for doing this verbally, using structured phrases, key
Figure 5: Ground briefing (Major AS Phipps 2003)
words, standard angular distances and confirmation procedures. Veterans (ARRSE 2009) and Storr
(2009) recognise these have been effective in the far past, but consider that the current training for it is
so insufficient, and the resulting marksmanship so poor, that it significantly affects operations.
This drill is a human mechanism for information transfer, and it has proved difficult to provide
adequate technology to support it. Acoustic shot detectors can pinpoint the direction and range of a
firing point, but relaying this is not straightforward; if the sensor is carried by a soldier lying in an
awkward pose, or moving around, then the direction from the sensor rarely matches the soldier's
perception. Markers on weapon sights show promise but the technology is still too cumbersome.
4.2 Distributed Pictures
We suggest (Hill 2009, forthcoming) that rather than try to establish a single information model,
populated with a 'single picture of the truth” that requires all information to be sent everywhere, we
design information systems on the basis that the information is distributed around the organisation as
required. This reduces the load on the communication links and the commander's time and attention.
Tom Love (1993) makes the general point that monolithic centralised repositories are a bad idea.
Rather than aiming to provide complete information, we should be asking how little we can get away
with; Simon says “information consumes the attention of its recipients” (Simon 1971).
Enemy are very hard to find, with few if any cues as to their existence or location. As Biddle says
(2004) success in battle relies on nullifying the effect of modern weapons, and much of this is done by
hiding. It is sometimes hard for civilians to realise quite how difficult it can be to locate a camouflaged
expert who does not want to be seen, and the phrase “The empty battlefield” has long described this.
Such information as there is must be separated from the mass of environmental clutter and false
4.4 Trust & Reliability
The British Army's training and ethos promotes units that place considerable trust in the soldiers. (This
trust may refer to a confidence in their reliability, rather than in the information they provide. There is
often someone in a patrol who can be safely ignored on all things). When dealing with external
organisations this trust tails off (McGuinness 2006) even when there is no cause to, and can result in
good information being rejected.
When soldiers work with forces from other countries and organisations, the language, jargon and
cultural differences can exacerbate communication difficulties. These barriers can be partly overcome
using liaison officers; these are people from your organisation that embed with theirs (and vice-versa).
Face to face discussion can help to ensure that the meaning of phrases like “It's a bit sticky” are not
lost, and the subtle differences between the phrases “This stuff is the dog's bollocks” and “This stuff is
bollocks” used by British soldiers is put across.
Communications links can also reduce trust in the information received. Modern radios are generally
secure so impersonations are rare. However voice quality can be poor, leading to misunderstandings.
Mitigation includes 'voice procedure' training that includes key words to reduce vocabulary, words to
aid clarity (such as the phonetic alphabet), procedures to confirm understanding, and certain phrases
used in particular situations.
Language, mannerisms and unconscious social signals can be crucial to properly understanding the
content of the information being relayed. Face to face meetings can help people to evaluate the
source of the information and the content, in ways that are not so easy with a radio or e-mail
4.5 Correcting the Picture
When new information arrives, it may confirm existing information, contradict it, create a new picture or
fill in missing gaps in existing ones. Where the information is contradicted, or assumptions had been
made covering the missing information, the picture will need to be corrected. People are not naturally
very good at this.
Similarly, the picture can alter in different ways when the same information is received several times.
For example, consider the case where three reports are received of a 'technical', which is an armed
pickup popular with some insurgents (see Figure 6) but shares some passing similarities to ordinary
pickups with scaffolding or other materials in the back. We can produce several different pictures
depending on the provenance of those three reports, amongst them:
Three different observations of three different „technicals‟ considerable threat
Three different observations of the same technical confirmation that one exists
The same observation via three different routes (for example, repeated rumour) one
possible unconfirmed threat
The same observation in three different reports (radio, verbal, written) one possible
The false confirmations in this case – where one observation becomes several apparently different
observations of the same thing – is sometimes referred to as 'data incest'. To avoid this, we need to
include with the tactical information some more information about where it has came from.
4.6 The enemy is not your friend
When it comes to evaluating all these activities and difficulties in populating the information models,
Figure 6: 'Technical' pickup, Liberia (Wikipedia, unknown source)
we must remember that the enemy is working hard to make these problems worse. He will hide to
increase the sparseness of your information about him, and do so amongst other people in order to
increase clutter. He will attempt to send you false information to distort your information model and
perhaps to reduce your trust in other allies, and theirs in you. Any information management systems
introduced have to be robust to these attempts.
5. Obstacles to Evaluation
Quite apart from the ordinary obstacles that humans have in evaluating evidence (such as
confirmation bias, deferral to authorities rather than information, assuming rather than checking,
refusing to modify existing pictures, etc), military operations can provide some unusual ones.
'Cognitive Load' is used to refer to the ability of the commander to absorb new relevant information
and make reasonable decisions based on it. Unfamiliar environments tend to swamp the relevant
information with clutter and ambiguity. When under fire there are stresses from the immediate
emotional response, and more from the pressure to remove the threat quickly. These may both be
exacerbated if some of his men are wounded. Exhaustion, immediate physical stress of running with
heavy loads, heat stress, dehydration, lack of sleep and so on all reduce the ability of a commander to
assess and understand new information. (Many of these debilitating conditions feed-back on each
other. Fear may reduce the ability to sleep. Exhausted soldiers may skip eating and so become more
quickly exhausted when moving).
So we can expect busy commanders to handle only a few new things at once. Providing them with all
the information all the time is not only not useful, it can be harmful. Especially so as the inability to
cope with a rich information system will surface more during stressful moments rather than training.
The management mantra is “The right information at the right time to the right people”, but it is not
clear yet what this means in the military environment; current efforts are mostly trial and error, and
there are no good mechanisms (distributed or central) to pass around lessons learned or 'best
Decision support mechanisms exist. Realistic training provides the commander with the practice to
learn what he needs to focus on and what can be deferred or delegated. Repeating procedures until
they become automatic (drills) reduces the cognitive load of both manual tasks and collaboration
tasks. Rules of Engagement provide guidelines for circumstances that allow opening fire, and can vary
depending on the conflict.
6. Automation and Information Technology
Some new technologies have been introduced very successfully and have made dramatic differences
to the capabilities of the military. GPS and overhead images provide accurate working ground
references to coordinate operations. Modern digital radio systems provide email and instant
messaging and file transfers. Modern mobile phones provide every man with a computing platform
even without a cellnet link.
Some have not been so popular. Commercial wireless systems could reduce the cables that adorn the
modern soldier, but as radio broadcasts they also give away his position, and can be used to detonate
The popular saying has it that we “fight the last war”. We should also beware of “procuring for the last
war” and creating information systems that work only for the current missions. However peacekeeping
and counter-insurgency roles are not unusual, and indeed have not been for centuries. Tuker (1948)
claims that historically linear (conventional) warfare is an exception, so it may be argued that
developing systems for current operations will have enduring value.
6.1 Discussion & Social
The rather flippant phrase “All plans fail on contact with the enemy” contrasts with the general doctrine
that “Thorough planning is vital to the success of military operations”. It may be that thorough planning
is not just about assembling a set of tasks, but about meeting and getting to know your colleagues and
their capabilities. Then, as the original plan begins to fail, it is straightforward for the team to
collaborate – remotely and with poor communications – to adjust it to the circumstances.
New information technology such as collaboration tools allows people to, for example, work on a
common electronic map from remote locations. Social networking software (TIGR) has been 'mashed
up' to provide soldiers with familiar peer to peer frameworks for sharing situation awareness. Chat and
email provide ways to maintain long discussions.
So the caveat is not that IT tools are harmful, but to take care not to replace existing face-to-face
planning where that personal and social familiarisation is important to the resilience of the mission
Similarly, providing automatic gateways between information infrastructures can, in principle, allow
different nations to connect up their systems and automatically share data. This provides much more
complete pictures, but the caveat here is to ensure that the liaison officer, who would have carried out
the language and cultural translations, is not lost without some other mechanism to compensate.
6.2 IT to Head
Information may be to-hand and indeed examined, but it does not necessarily become part of the
soldier's situation awareness. A common, long running anecdotal complaint is that radio messages are
commonly ignored – even unnoticed – when the commander is busy, despite being broadcast straight
into the ear.
More recent examples are the automatic map displays that can be glanced at and assessed with no
long term memory being laid down. Car drivers may have noticed that despite driving a new route
several times while guided by a satnav, they cannot recall which directions to take at junctions.
This disconnection between human and IT is ignored by some system designs that are satisfied with
ensuring that information reaches only as far as the operator's terminal. Extra traffic to acknowledge
receipt, and correction protocols to rebroadcast, all add to the communications load and are largely
irrelevant; it is the receiver that needs to acknowledge receipt, timely enough for the sender.
6.3 Decision Anti-support
Some information systems encourage dangerous decisions by not providing information when they
fail. “Combat ID” is an attempt to electronically label all friendly units, so that remote firers such as
attack helicopters can check they are not firing on friendly troops. This is meant to be a safety check to
reduce friendly fire fatalities, but it is the wrong one. Equipment that fails, is lost, or is not supplied has
no signature, and the enemy has no signature. So a lack of friendly ID merely indicates a lack of
6.4 Augmented Reality
Efforts have been made to provide suitable parts of the tactical information model to the soldier
through augmented reality hardware, typically goggles, monocles or weapon sights with displays
projected onto the line of sight through them. The technology for doing so reliably, without too much
extra weight or power requirements, and without succumbing to the harsh environment, is improving.
However some outstanding issues remain: what should be displayed to avoid overloading the view or
the mind, what training is required to avoid it becoming a distraction, how can the user interact with it
to clear it or switch it, and so on. .
Information management in counter insurgency operations combines all the ordinary information
management challenges with a heavily constrained environment where humans are significant sensor
and computing components.
Existing systems are in place and being adapted to suit current operations. Further improvements can
be made, in three overall areas:
Understanding the information management & exploitation requirements of counter-
Systematically monitoring the effectiveness of changes to those systems.
Distributing the results of that monitoring through the military
We offer this fascinating field as an extreme environment to test the effectiveness of information
management techniques and technologies, and invite experts to contribute their work to it.
Storr, Jim (2009): “Real Role of Small Arms in Combat” [online], RUSI, UK
DEVIL, Task Force (2003) “Modern Warrior's Combat Load”, US Army Center for Army Lessons
Hall and McChrystal (2009) “ISAF Commander's Counterinsurgency Guidance” NATO,
Gant, Jim (2009) “One Tribe at a Time” (Small Wars Journal) http://blog.stevenpressfield.com/wp-
Tuker, Lt Gen Sir Francis (1948), “The Pattern of War”, Cassell, London
Hill, M (2009) “Distributed Operational Pictures”, CIBM, UK (forthcoming)
Love, Tom (1993) “Object Lessons”, SIGS Publications, New York
Simon, H.A. (1971), "Designing Organizations for an Information-Rich World", Computers,
Communication, and the Public Interest, The Johns Hopkins Press, Baltimore.
Biddle, Stephen (2004) “Military Power”, Princeton University Press, Princeton
ARRSE 148286 (2009): “The Reality of Herrick”
Potts, Malcom & Hayden, Thomas (2008) “Sex and War” Benbella Books, Dallas
Kilcullen, David (2009) “The Accidental Guerrilla” Oxford University Press