You can be making progress and getting closer to failure at the same time.

Imagine a game character facing twelve enemies. You use your strongest attacks and kill four. Your health is almost full. Everything looks under control.

But your shield has broken. Your speed boost is fading. Your healing is already used. The numbers look better, while your ability to handle the next surprise has got worse.

This is a scenario from Sebastian Marshall’s Roguelike. It asks a useful question: what options do you still have?

There’s a staircase one move away. Marshall’s answer is to go upstairs, recover your abilities, and return to fewer enemies.

Why leave while you’re doing well? Because the next stage of the fight depends on the resources you have left, not the success you’ve already had.

In this example, retreat is available and recovery is cheap. That matters. This isn’t a rule to abandon every difficult situation. It’s a reminder to examine your remaining responses before the visible score tells you something is wrong.

Here’s a workplace analogy of our own. A small team accepts several exciting projects. Every hour is booked. Then one task runs late, and there’s no room to absorb it. The busy calendar was hiding a fragile plan.

Marshall tells a related story about his friend Kai. During a visit to London, Kai skips a board game to finish consulting work several days early. Later, unexpected conversations and events fill the trip, and he can join in.

The point isn’t to work every spare minute. It’s that resolving one obligation can buy freedom to respond later.

A second example shows why the most impressive upgrade may be the wrong one.

Marshall imagines equipment built around a powerful chain: daze an enemy, land critical hits, trigger extra attacks. The pieces reinforce each other.

Then a ghost dragon appears. It can’t be critically hit. The condition that makes the combination work has disappeared.

A cloak that adds more power to the usual combination doesn’t solve that problem. A modest light-damage attack can help against the ghost. It doesn’t guarantee victory. It gives the character a response where the main strategy breaks down.

Recognising danger is easier than acting on it while you feel successful. That’s why Marshall recommends defensive policies.

Imagine you have three major heals. Instead of waiting until all three are gone, decide to retreat after the first or second. Which one depends on how dangerous it will be to get away.

The structure is simple: notice a specific event, take a specific action, and keep enough reserve to carry it out.

For our imaginary team, that could mean reviewing new commitments when the agreed schedule buffer disappears. The right threshold depends on the situation.

A plan also has to fit the person using it. Marshall calls this piloting.

A dozen powerful abilities may be better in experienced hands. For a beginner under pressure, a simple sequence can be easier to use well.

That doesn’t make complicated strategies bad. It means a plan’s demands on attention and judgment are part of its cost.

Policies can reduce that burden. They also need room for deliberate exceptions when circumstances change. The aim is reliable action, not obedience to a rule that no longer fits.

But life gives us choices a game may leave out. We can negotiate, cooperate, ask for help or work to change the rules.

And understanding an advantage doesn’t mean you have to take it. Marshall explicitly allows refusing something on ethical grounds, accepting the cost or choosing another environment.

His inventor doesn’t have to build a successful company. They could learn the business side, work with others, or invent for the pleasure of it.

Before you make a strategy more effective, decide what you actually want it to achieve.

One last limit: good decisions can still lose. Marshall says that even excellent play can be defeated by chance. Preparation changes some risks; it doesn’t make every outcome controllable.

So these stories are useful prompts, not proof that people cause every setback they experience.

Our synthesis is three questions: What is the goal? What could break the plan? And what options would remain?

Sometimes the answer is to keep going. Sometimes it’s to cover a weakness. Sometimes it’s to take the stairs while you still can.
