Targeting and disappearance is described in Software Development : Security. What regards to the rest, then I do not know practically anything about it. This text here reflects purely my subjective, may be extremely stupid, preferences in a situation, where I should be involved in some quarrel like that.
- Strategy (that little that I know of it)
- Tactics (that little that I know of it)
- Ideas for Weapon Systems
Security Team ("Army")
To minimize the amount of communication and to eliminate the opportunity of enemies to kill "VIP-targets", there is no hierarchy and there is no real-time central decision-making. Job of a classical general is obsolete. The battle plan is implemented in an agent oriented programming fashion with an understanding that information sources are not reliable (File:The Byzantine Generals Problem by Leslie Lamport and Robert Shostak and Marshall Pease.pdf). Members of the security team decide themselves, if, when, how they participate and whom they trust, with whom and to which extent they cooperate.
Classical security companies might sell their own company specific "war bonds" to keep some of their war capabilities available. Just like a zoo gets sponsors for feeding an elephant, a security firm might sell "war bonds" to maintain tanks, helicopters, drones, etc.
Online games have demonstrated that if enemy players that are of most skillful rank overwhelm the most skillful players of one's team, defeat is probabilistically guaranteed. Therefore, to have any chance at all against a bigger army, there is no point of spending resources and time on advancing the skills of weapons that the bigger army has. The only thing that can help is better strategy and better weapons. The effort and resources should go to weapons development and the training for advancing the skill to use the new weapons. The smaller the team, the harder it is to find and spy on it.
Every second weekend there might be a gathering of security firms and other military equipment owners (tanks, helicopters, etc.) for war practice/exercises/competitions. Those are also events, where trust bonds are created and its a marketing opportunity for security firms to demonstrate, what their "war bonds" are worth.
The attacking side has always the advantage of being prepared. The attacking side can place attack units at area exits, prepare its activity, get to know the area of the dogfight. Attack unit, where a human is bundled with a sword or a gun, is outdated and will be out of order the moment the human is hit with a single bullet. Besides, humans are physically big targets, if compared to drones, droids, turrets.
In my opinion the only reasonable goal of the party, who did not initiate the dogfight, is to avoid being shot/slashed, escape and get the dogfight initiator off its tail. All efforts should be placed on the escape effort. Offensive action should only be used to support the escape. Generally, if the other side already knows the physical area, where the defensive party resides, it's game over with all of the rest of the activity and escape is the only reasonable thing to do.
To avoid being attacked in the future, an offense must be carried out later. To carry out the offense, automated information gathering must take place to determine the attacking party.
If the defensive side consists of more than one human, then the defending parties should disperse, escape to different directions. If there are a lot of people involved at the defense side, they should disperse in pairs to make it possible for one of the pair members to carry out the other, should the other get wounded.
Preparation. Preparation. Preparation. One should never reside at the area, where the action takes place. If there's not an outright war going on, the offensive party must stay anonymous from its victim.
Given the fact that everyone has some area, where they will physically appear repeatedly, an office, home, place, where family members live, some club, groceries, it is fair to say that the moment the identity or the occupation of the person is known, it's only a matter of surveillance to find his/her location.