When to Stop

Tower has one question and it is not how high can you climb. It is enough?

Every survived row gives you the same option: take the multiplier you earned, or risk it for a bigger one. The math of that choice is brutal in a quiet way, and almost nobody looks at it before they sit down.

Look at it now.

The compounding problem

Say you are on Medium difficulty — three cells per row, two safe, one mine. Each step is roughly a 67% chance of survival.

A six-row climb on Medium fails about 91 times in 100. The multiplier is bigger because the survival rate is smaller. The tower is not being mean. It is being honest. The math has to balance.

The instinct, mid-climb, is to think I have made it this far, the next one will probably hold. The next one will probably hold. The next one is independent of every step before it. The 33% mine chance you faced on row one is the 33% mine chance you face on row six. The tower has no memory of how far you have already come. It does not care that you "deserve" to keep going.

This is the gambler's fallacy, dressed up. Resist it.

House edge, plainly

Every Tower payout is set just slightly below what perfectly fair odds would pay. The difference, expressed as a percentage of the bet, is the house edge. On Tower it sits in the low single digits — historically documented around 3%, with the exact value varying by difficulty and the depth of the row you are climbing.

Whatever the precise number, the principle is the same: a small slice of every wager goes to the house, and over enough rounds that slice accumulates. Over a short session, the edge is mostly noise — variance dominates, and you are more likely to walk away dramatically up or dramatically down than you are to land near the average.

Two things follow:

1. Short sessions are unpredictable in your favor or against you. Variance is high.
2. Long sessions trend, on average, in the house's favor. The edge accumulates.

Both are true. Most players forget the second one.

The "double or nothing" trap

The most expensive mistake in Tower — and in every game like it — is climbing one more row to prove the previous climb was not luck.

It was luck. All of it is. The next climb is also luck. Doubling-or-nothing on a hot streak is the same expected value as starting fresh. The streak does not exist. The hot you feel is the brain's pattern-finding doing what the brain does.

If you are about to make a bet specifically to confirm or rescue a previous one, do not. Cash out. Stand up. Drink water. Come back if you want.

Pre-deciding

The strongest tool against in-the-moment greed is pre-deciding. Before you start a session, write down two numbers:

• Win cap. The multiplier or sat amount at which you will stop, no matter what.
• Loss cap. The amount you are willing to lose this session, no matter what.

The win cap is harder to honor than the loss cap. People expect to enforce the loss cap and almost never expect to enforce the win cap. Both matter equally. A session that hits the win cap and keeps going is a session that gives back the win.

Pre-decide. Out loud, in writing, somewhere you can see. The tower is patient. You do not have to be.

Bankroll, in one paragraph

Use a fraction of your spending money you can afford to lose entirely. Not your savings. Not your rent. Not money you need to grow. The math means you will sometimes win and you will sometimes lose, and over enough sessions the house edge says you will lose a small slice. That is the cost of the entertainment. Price it accordingly.

What good climbing looks like

The climbers who do well in Tower are not the ones who go highest. They are the ones who hear the question and answer it honestly.

Cash out earlier than feels satisfying. Take the win that is in your hand. Let the streak go.