The Art of War, Chapter 8: 九變 (Variation in Tactics)

Chapter 7 is about contesting the advantage and moving fast to get it. Chapter 8 is the check on that. It argues that a commander has to know when the usual rule does not apply, and that the commander’s own weaknesses are among the ways an army is lost.

Like every chapter in this series, it has three parts: the reading, an analysis of what the chapter argues, and how finance professionals can interpret it in their own work. The original text follows the Chinese Wikisource edition of 《孫子兵法》. The Chinese and English explanations are Claude’s own paraphrase.

Reading the chapter

8.1

原文

孫子曰：凡用兵之法，將受命於君，合軍聚眾。圮地無舍，衢地合交，絕地無留，圍地則謀，死地則戰。

中文解說

孫子說：用兵作戰的通則是，將領接受君主的命令，召集並組織軍隊。在難行的險地不要紮營，在四通八達的要地要結交鄰國，在補給斷絕的絕地不要停留，在容易被包圍的地方要設法謀劃，在無路可退的死地就要奮戰。

這五句是對不同地形的簡短處置規則。「九地」的細部定義見第十一章，這裡只列出處置的原則。

English

A commander receives orders from the ruler, then gathers and organises the army. After that comes a set of short rules, one for each kind of ground. Do not camp on difficult ground. On crossroads ground, build alliances. Do not linger on isolated ground. When surrounded, plan. When there is no way out, fight.

These are rules of thumb for different situations. The types of ground are defined in more detail in Chapter 11. Here they set up the point that follows.

8.2

原文

途有所不由，軍有所不擊，城有所不攻，地有所不爭，君命有所不受。

中文解說

有些道路不該走，有些敵軍不該打，有些城池不該攻，有些地方不該爭，有些君主的命令也可以不接受。

English

There are roads not to take, armies not to attack, cities not to besiege and ground not to contest. There are even orders from the ruler that are not to be followed.

The last clause is the strongest. It says that the commander on the spot may set aside an order when it no longer fits the situation. Chapter 3 makes a related point about a ruler who interferes with the army.

8.3

原文

故將通於九變之利者，知用兵矣；將不通於九變之利，雖知地形，不能得地之利矣；治兵不知九變之術，雖知地利，不能得人之用矣。

中文解說

所以，將領如果精通各種變通的好處，就算是懂得用兵了。如果不精通這些變通，即使熟悉地形，也無法取得地形的好處。帶兵而不懂得變通的方法，即使知道五種地利，也無法讓部下發揮作用。

English

A commander who understands the benefit of varying his methods knows how to use an army. One who does not will fail to turn the ground to his advantage, even if he knows the terrain well. One who commands troops without knowing how to vary his methods will not get the best out of his people, even if he knows the five advantages of ground.

九變 is usually read as 'nine variations' or as 'many adjustments'. Commentators disagree on whether the number is literal. The argument does not depend on it. Knowing the facts is not enough, because the facts have to be applied with flexibility.

8.4

原文

是故智者之慮，必雜於利害，雜於利而務可信也，雜於害而患可解也。

中文解說

所以聰明人的考慮，一定同時兼顧利與害。在有利的情況下考慮到不利的因素，事情就可以順利進行；在不利的情況下考慮到有利的因素，禍患就可以化解。

English

A wise person's thinking always mixes gain and harm. When the gain is considered together with the harm, the plan can be trusted to work. When the harm is considered together with the gain, the trouble can be resolved.

The point is to look at both sides of every situation. In a good position, ask what could go wrong. In a bad one, ask what is still available.

8.5

原文

是故屈諸侯者以害，役諸侯者以業，趨諸侯者以利。

中文解說

因此，要使諸侯屈服，就用他們害怕的損害；要使諸侯聽從差遣，就用他們不得不做的事務；要使諸侯奔走趨附，就用他們想要的利益。

English

To make rival rulers yield, threaten them with harm. To keep them occupied, give them tasks. To make them hurry toward you, offer them gain.

The word 業 is read differently by different commentators, as 'tasks', 'enterprises' or 'burdens'. All the readings point to the same idea: other parties can be moved through what they fear, what they have to do and what they want.

8.6

原文

故用兵之法，無恃其不來，恃吾有以待之；無恃其不攻，恃吾有所不可攻也。

中文解說

所以用兵的原則是，不要指望敵人不來，而要依靠自己有準備來應對；不要指望敵人不進攻，而要依靠自己擁有使敵人無法攻破的實力。

English

Do not rely on the enemy not coming. Rely on being ready for them. Do not rely on the enemy not attacking. Rely on having a position that cannot be attacked successfully.

This is a rule about where to place your trust. Safety comes from your own preparation, not from the other side's goodwill or inaction.

8.7

原文

故將有五危︰必死，可殺也﹔必生，可虜也﹔忿速，可侮也﹔廉潔，可辱也﹔愛民，可煩也。凡此五者，將之過也，用兵之災也。覆軍殺將，必以五危，不可不察也。

中文解說

所以將領有五種危險的性格弱點：只知拚死，就可能被殺；只想求生，就可能被俘；暴躁易怒，就可能被激怒而受騙；過分清廉自守，就可能被羞辱而失控；過分愛護百姓，就可能被騷擾而疲於奔命。這五種情況都是將領的過失，也是用兵的災難。軍隊覆滅、將領被殺，一定是由這五種弱點造成的，不可不仔細考察。

English

A commander has five dangerous faults. Recklessness can get him killed. Fear for his own life can get him captured. A quick temper can be provoked into a mistake. A fixation on his own honour can be shamed into acting. Too much concern for the people can wear him down with constant trouble. These are the commander's failings and the disasters of war. When an army is destroyed and its general killed, it will be through these five. They must be examined.

Two of the five are contested. 必生 can mean cowardice or simply a refusal to take any risk. 愛民 can be read as a real fault or as a warning that a commander who cannot bear to see his people suffer can be manipulated by threats to them. The text ends with 不可不察也.

Analysis: what the chapter argues

Rules have conditions. 8.1 lists a rule for each type of ground, and 8.2 immediately lists exceptions. Together they say that a rule is a default, not a command. The commander has to know when the conditions that justify the rule are absent.

The exceptions go all the way up. 8.2 includes orders from the ruler in the list of things that may be set aside. Reading this against Chapter 3, the point is not that orders are unimportant. It is that the person closest to the facts has authority to adapt, and that the system should allow for this. Commentators differ on how far the clause goes, and the text does not set a limit.

Knowledge is not enough. 8.3 says that knowing the terrain and the five advantages without knowing how to vary methods produces no result. Information has to be turned into judgment. This is an argument against mechanical application of any playbook, including this one.

Always look at both sides. 8.4 describes the wise person’s habit: add the harm to the thinking when things look good, and add the gain when things look bad. Both errors, overconfidence and despair, come from seeing only one half.

Other parties respond to harm, tasks and gain. 8.5 lists three levers. The reading of 業 is contested, as 8.5 notes, but the logic holds under each reading.

Rely on your own readiness. 8.6 puts the source of security in what you can control. It is one of the most quoted lines in the book because it is easy to state and hard to follow.

The commander is the main risk. 8.7 turns the lens from the situation to the decision maker. The five faults are not external threats. They are traits that an opponent can see and use. The section ends with 不可不察也, which asks for careful examination, here of oneself.

How finance professionals can read it

This section is Claude’s interpretation. Sun Tzu wrote about armies, and none of the mappings below are his. Chapters 1 to 7 describe how to plan, cost, position and move. This chapter is about the limits of all of that: when the plan should bend, and when the person running it is the weak point.

Investing

A playbook is a default. Position limits, valuation screens and entry rules are the equivalent of the ground rules in 8.1. They work in the conditions they were built for. Write down those conditions so that it is clear when they no longer hold.
Define the exceptions in advance. 8.2 lists things not to do even when they look attractive. Some assets, structures or counterparties are not worth pursuing at any price. Deciding this before a deal is live is easier than deciding it under pressure.
Pair every bull case with a harm case. 8.4 maps to writing the downside next to the upside, and to writing the remaining options next to a loss. This is a habit of thinking, not a formula.
Check the five faults on yourself. Recklessness, loss aversion, anger after a loss, attachment to a reputation for discipline, and being worn down by pressure from stakeholders are all recognisable in an investor. A review after a loss that asks which of these was present is more useful than one that only reviews the numbers.

Raising funds

Do not rely on the market staying open. 8.6 maps to funding. A company that plans to raise only when it needs to is relying on the enemy not coming. Hold enough runway or committed facilities that the company is not forced to raise on bad terms.
Know what moves each counterparty. 8.5 maps to the different motivations of different investors and lenders. Some respond to downside protection, some to a role in the company, some to return. Reading this wrongly wastes time.
Adapt the process to the situation. A standard process suits a standard company. If a company’s situation is unusual, forcing it through the usual template may produce the wrong result.
Keep the founder’s temper out of the room. A founder who reacts to a hard term sheet with anger, or who cannot bear to disappoint staff, is showing 8.7. A trusted adviser can be the check on this.

M&A

Know when the process rules do not fit. Standard timetables, bid procedures and diligence lists are defaults. A deal with an unusual seller, a distressed target or a regulatory deadline may need a different path. Document why the deviation is justified.
Walk away from some deals. 8.2 includes the cities not to attack. Some targets are not worth the cost even at a discount. A pre-set list of conditions under which the team will stop is easier to follow than a decision made after months of effort.
Look for the lever that moves the other side. Per 8.5, the counterparty may respond to a risk, a task or a benefit. Understanding which one matters to them is part of preparing the approach. Anything involving pressure on a counterparty needs a legal and compliance check.
Guard against sunk-cost attachment. A deal team that has invested months is open to the fault in 8.7 of being unable to stop. An independent reviewer who was not part of the work can ask whether the deal still makes sense.
Check the escalation path. 8.2 raises the question of when the person on the spot may depart from an instruction. Deal teams need a clear rule about who can change the plan, and when, with the governance and approvals that apply.

Restructuring and distressed situations

The rules differ by type of ground. 8.1 gives different rules for the difficult, the isolated, the encircled and the hopeless. Distressed situations are also not all alike. A liquidity problem, a solvency problem and an operational problem need different responses.
Plan inside the constraint. When surrounded, 8.1 says to plan. When there is no way out, it says to fight. Both can be read as advice to stop hoping for an easy exit and to act inside the limits that exist.
Be careful of the five faults in the room. Anger at the other side, a wish to be seen as fair, and desperation to avoid failure are all common in restructurings. Any actions involving creditors, directors or insolvency law need legal advice.

Across all three

Write down when the rule stops applying. For every standing rule, record the conditions under which it should be set aside, and who decides.
Read the downside next to the upside, and the upside next to the downside.
Build your own readiness. Count on what you can control, not on the other side holding back.
Review yourself as well as the deal. Ask which of the five faults was present in the last decision that went wrong.

Questions to carry into the next chapter

Chapter 9 (行軍) moves from the commander’s judgment to the practical work of reading signs: how an army camps, marches and interprets what it sees in the enemy.

Which of your standing rules has conditions that you have never written down?
Which of the five faults is most likely to affect your decisions under pressure, and who would tell you?
Where are you relying on the other side not acting, rather than on your own preparation?