An internal analysis may be correct and still lead to the wrong decision. Knowing what you can do does not reveal how an opponent will respond or which conditions favour their position. That is the specific problem of Sun Tzu’s middle scenario.
Ignoring the enemy does not necessarily mean underestimating yourself or being unaware of your own resources. It can coexist with a rigorous internal assessment. In self-knowledge and strategic information, this case shows the limit of a decision built solely on data about the person making it.
The problem of knowing only one part of the relationship
In chapter 3, section 18 of Lionel Giles’s 1910 translation, Sun Tzu contrasts every victory that comes from knowing one’s own side while ignoring the enemy with a defeat. That wording does not establish a 50 per cent rule for victories. It points to the insufficiency of one-sided knowledge; it does not describe a series of measured outcomes.
The full analysis of knowing the enemy and knowing oneself places this case among the other scenarios in the passage. Its particular feature is that self-preparation does not disappear. What is missing is a basis for judging how it will work against specific external conditions.
A strength expresses an ability; an advantage requires a comparison. Being able to act quickly does not prove that speed changes the outcome. It could be insufficient against another response, irrelevant to the objective, or costly in coordination terms. The leap from ‘we can do it’ to ‘doing it works in our favour’ needs additional information.
It is also necessary to distinguish the opponent’s capabilities from their intentions. Knowing that they have a resource does not confirm when they will use it. Observing an action does not necessarily reveal its full purpose either. A lack of information is not solved by attributing to the rival the motives one would have in their place.
For that reason, ignoring the enemy means more than knowing nothing about them. It may involve using outdated information, knowing aspects that do not affect the decision, or taking an untested interpretation as certain. A bulky file can still leave unanswered the specific question on which a plan depends.
A hypothetical case: the same capability, a different decision
Imagine two teams putting forward proposals for organising an organisation’s document archive. It is a fictional contemporary example, not an episode from the treatise. We will assume that both meet the basic requirements and that the organisation has not yet chosen either one.
Team A knows its available hours, has checked the quality of its work, and can prepare a trial in ten days. Team B needs fifteen. On the basis of that difference, A centres its proposal on speed and considers it unnecessary to study what B offers. Its internal data about the deadline may be completely correct.
The reasoning adds, without proving it, that delivering earlier will be the decisive criterion. It also assumes that the services are equivalent in every other respect. Neither of those claims follows from knowing one’s own productivity. They are external conditions introduced into the analysis as though they had already been verified.
Now let us add new information to the example: the organisation’s published criteria give greater importance to continuity of work during the transition than to the date of the trial. In addition, B’s public proposal includes support for that transition. This does not prove that B will be selected; it changes what A needs to examine in order to support its proposal.
Team A can still work in ten days. What loses force is the conclusion that speed alone will be enough. It can check whether its process keeps the organisation’s activity going and explain how it will do so, or recognise that its proposal does not cover that requirement adequately. None of those responses requires pretending to have abilities it does not possess.
The change does not consist of copying the other team automatically. It consists of assessing one’s own capability against a better-defined problem. Information about B provides comparative context; the organisation’s criteria show what that comparison is for. Confusing the two would lead to competing on aspects that nobody has asked for.
To test the example, it is worth asking: is the deadline really the priority, which part of B’s offer is documented, and what information is missing about our transition? It also matters to establish what finding would require the proposal to be revised. A useful question is one whose answer can change a decision, not one that merely confirms that a lot of work was done.
The outcome remains open. A could improve its proposal and still not be chosen, or could be chosen despite having analysed part of the problem badly. No isolated result would demonstrate a 50 per cent probability or turn a lack of information into an advisable method.
Obtaining useful information without confusing it with certainty
Chapter 13 gives an explicit function to human intelligence about the adversary and rejects replacing it with mere calculation. The five classes of spies in chapter 13 explain that approach in its military context. The question of interest here is which piece of data allows a specific assumption to be tested.
Gathering, checking and interpreting are different operations. A message may have been received without being verified; a verified piece of data may admit several explanations. Keeping those differences clear prevents the confidence with which a report is presented from being confused with the confidence of its content.
Information also has a scope. An observation allows us to describe what happened in one place and one moment; extending it to others requires reasons. The fact that a capability has not been shown does not prove that it does not exist, just as announcing a capability does not prove that it is available.
In a business adaptation, open-source competitive intelligence focuses on examining public information. That adaptation does not equate any data-gathering practice with the treatise’s espionage, nor does it justify accessing private communications or other people’s secrets.
There can also be resistance to revising a conclusion once it has been defended publicly. The risks of deciding in anger and ego address that difficulty. Receiving outside data is of little use when the criterion for acceptance is that it must not contradict the image one wants to preserve.
There is no need to know everything about the rival in order to abandon an unjustified assumption. The information available may be enough to narrow the scope of a decision or keep an alternative open. What it does not allow is presenting a lack of data as evidence that no obstacle exists.
Before attributing an outcome to a lack of information, identify what was unknown and how knowing it would have changed the decision. That is how a specific limitation is distinguished from an explanation constructed after the outcome.
View all articles in Know yourself: self-knowledge and information in Sun Tzu
