Read the Game Before It Happens

I spent years playing volleyball in Switzerland’s National League A and on the Swiss Beach Tour. When people watched a good defender, they usually saw the final movement: the sprint, the dive, the ball somehow returning over the net.

What they did not see was the decision made a fraction of a second earlier.

The defender had already noticed the hitter’s shoulder, the position of the elbow, the speed of the approach and the quality of the set. The movement looked fast because the recognition happened early.

Karch Kiraly explains this beautifully in a USA Volleyball lesson on reading the game. He calls it perceptual expertise. In volleyball, we might call it court sense or volleyball IQ.

The name matters less than the mechanism.

You see a cue. You compare it with patterns stored through experience. You eliminate outcomes that no longer fit. Then you move before the answer is obvious.

That is also a useful description of strategy.

Reaction is already late

Most organisations talk about becoming more agile. This usually means they want to react faster after something has happened.

Sales dropped. A competitor launched. A model improved. A regulator published a rule. A customer left.

Then the dashboards turn red and everybody becomes surprisingly interested in speed.

Useful, certainly. Early, no.

Kiraly’s lesson is more demanding. Do not wait for the result. Learn to read the pattern that produces the result.

He argues that great readers build their advantage through repeated exposure, creating a “visual encyclopedia.” Experts also learn which cues deserve attention. A novice follows the ball. An experienced player watches the structure around the ball.

The strategic equivalent is simple.

Do not stare at the quarterly number. Read what is creating it.

Learn the baseline before chasing the anomaly

One of Kiraly’s most useful distinctions is between baseline and “hinky.”

Baseline is the normal pattern. A hitter’s usual approach. A server’s regular motion. A setter’s familiar body position.

Hinky is the deviation. The elbow sits differently. The ball is in another position. The body is too upright. Something small suggests that the normal outcome may not be coming.

This is how experienced operators think.

When I was responsible for a channel ecosystem spanning EMEA and Latin America, with roughly 30,000 resellers, I could not inspect every transaction or attend every meeting. The work depended on knowing what normal looked like.

Normal pipeline movement. Normal partner behaviour. Normal regional variation. Normal delay, noise and optimism, because sales forecasts have never suffered from a shortage of optimism.

Once the baseline was clear, deviations became useful. A change in deal mix, partner activity or conversion gave me a specific place to look before drawing a conclusion.

That is an important distinction. An anomaly invites investigation before any conclusion.

AI systems need the same discipline. An anomaly detector without a trustworthy baseline is a machine for producing anxiety at scale.

Technology: build a pattern library, not a tool collection

Companies currently approach AI like enthusiastic tourists at an electronics market. They collect tools.

One assistant for writing. Another for meetings. Three for research. An agent that promises to run the company before lunch. Nobody is entirely sure which version has access to customer data, but the demo was excellent.

Tools matter. Patterns matter more.

At humAIne, we start by understanding the workflow well enough to see where intelligence changes the economics of the work. Placing AI beside the existing process and calling it transformation misses the point.

What is the normal path from customer question to answer? Where does judgment enter? Which handover causes delay? Which error repeats? What information is missing when a decision is made?

That is the baseline.

Only then can we see the meaningful deviation: the point where an AI agent reduces ten steps to two, where automation creates a new risk, or where a human becomes more important because the machine has made the ordinary part cheap.

Dear executive committee: buying an AI licence gives you a receipt. Perceptual expertise comes later.

Strategy: use time slices

Kiraly teaches players to compare frames from the same movement. Early in a serve, two outcomes may look identical. A few frames later, the shoulder, torso or ball position begins to diverge.

This is a powerful way to think about strategy.

Take time slices of the business.

Compare the same customer journey six months apart. Compare the same product metric before and after a model change. Compare how a competitor describes its market across several earnings calls. Compare regulation as proposal, guidance and enforcement. Compare the cost curve at equal stages of scale.

Time slices reveal the first frame in which the future became different. Admiring the past is beside the point.

Consultants can make almost anything look inevitable after it happened. The useful skill is noticing divergence while the outcomes still look similar.

For an investor, this might be a change in capital allocation before it appears in earnings. For a technology company, it might be a shift from experiments to production workloads. For a market, it might be customers changing behaviour before the category receives a fashionable new name.

One reliable clue does not guarantee the prediction. It tells you where to update the probability.

That is what good strategy does.

Business: freeze before the outcome

Another Kiraly exercise is wonderfully simple. Pause the video before the setter touches the ball. Predict where the set will go. Then play the clip and score the answer.

Business teams should do more of this.

Before the launch, record what you believe customers will do.

Before the acquisition, write down where the value will come from.

Before the AI deployment, predict the bottleneck, error rate and human intervention required.

Before the investment, state what evidence would prove the thesis wrong.

Then reveal the outcome.

Most companies perform post-mortems. Far fewer preserve the pre-mortem in a form that can be compared with reality. Memory is a generous editor. It removes our uncertainty and upgrades old guesses into convictions we apparently held all along.

Write the prediction down.

Scoring it is uncomfortable. That is why it works.

Real play teaches what slides cannot

Kiraly ends with the strongest training method: play real volleyball.

Demonstration helps. Time slices help. Slow motion helps. Freeze-and-predict helps. But players need representative repetitions in the real game, where perception and action remain connected.

This matters enormously in business.

In cultivated seafood, a promising laboratory result is necessary. The business begins when the pattern survives scale, cost, regulation, manufacturing, customer acceptance and time. Biology is not impressed by the presentation deck.

AI has the same problem. A controlled demo proves that the demo worked. A benchmark reports performance on its test. A pilot run by its inventors tells us little about broad adoption.

Real play means putting the system into consequential work with boundaries, monitoring and human accountability. It means observing what people do, including the inventive ways they work around a process that looked perfectly reasonable in the workshop.

Strategy becomes useful when it touches reality.

Five ways to train an organisation to read

Kiraly’s five methods transfer remarkably well.

1. Demonstrate

Show the normal pattern and the deviation side by side. Do not tell people merely to “be more strategic.” Teach them what to look for.

2. Use time slices

Compare the same system, customer, competitor or market at successive moments. Find the earliest meaningful difference.

3. Slow it down

Reconstruct important decisions. Separate signal, assumption and action. Look for the cue that was available before the outcome.

4. Freeze and predict

Pause before the reveal. Record the expected outcome and the reasoning. Then score it without rewriting history.

5. Play the real game

Use live customers, live workflows and live constraints. Keep the experiment safe, but keep it representative.

Together, these methods build organisational court sense.

The result is a better pattern library and a faster update when reality changes. Clairvoyance remains unavailable.

The ten-decision challenge

Here is the exercise I would use with a management team.

Choose ten past decisions with enough information to reconstruct the moment before the outcome became obvious. Hide the final result. Give the team the information available at the time.

For each case, ask four questions:

• What is the baseline?
• What looks unusual?
• Which outcomes can we now rule out?
• What would we do next?

Write the answers. Reveal the outcome. Score the prediction and, more importantly, the reasoning.

Then update the pattern library.

Use the exercise to improve what the organisation notices next time. Cleverness is abundant in rooms where everybody already knows what happened.

Technology will keep accelerating. AI will make more answers faster and cheaper. That increases the value of recognising which situation is developing, which cue matters and when a familiar pattern has changed.

The machines will process the play.

Humans still have to read the game.

---

Source: USA Volleyball — “READING on the VOLLEYBALL Court”, presented by Karch Kiraly; see also USA Volleyball’s official Coach Academy summary.