GM. This is The Great Unlock, your weekly cheat sheet for the AI revolution. We filter out the noise, red pen fake news, and give out stickers to those deserving.

Today’s deep dive: what Annie Duke’s Thinking in Bets reveals about why we measure the wrong thing in marathons, in classrooms, and in every intervention stack you run.

Let’s unlock it.

A marathon is 26.2 miles. Training is 500+. We celebrate the 26.2. Annie Duke says that is the problem.

This weekend, 50,000 runners will line up at the London Marathon. Some will PR. Some will bonk at mile 20. Some will finish in tears. Some will not finish at all.

And by Monday morning, every one of them will be evaluated by exactly one number: their finish time. Or whether they have one at all.

This is how we evaluate students, too. Test score. Final grade. Pass or fail. The scoreboard is the story.

Annie Duke, former professional poker player and author of Thinking in Bets, has a framework that exposes why this is broken. She calls it resulting: judging the quality of a decision by the quality of its outcome. And she argues it is the single most dangerous habit in high-stakes thinking.

Duke’s framework is a simple 2×2. On one axis: decision quality. On the other: result quality. That gives you four quadrants.

/Every runner and every student lives in one of these four quadrants after any given performance. The scoreboard only reads the top row (result). It never reads the left column (decision).

That is the problem.

Most educators and district leaders focus on the bottom-right: bad decision, bad result. That is where the intervention machine kicks in. Tutoring. Remediation. Extended time. Pull-outs.

But the most dangerous quadrant is bottom-left: bad decision, good result. The student who crammed and passed. The runner who skipped training and still crossed the line. The scoreboard says success. The process was garbage.

This quadrant is dangerous because it reinforces the exact behaviours that will eventually produce catastrophic failure. The student learns that cramming works. Until it does not. The runner learns that skipping long runs is fine. Until mile 22 of a harder course.

The equally important quadrant is top-right: good decision, bad result. The student who prepared well and still bombed the exam. If the only thing we measure is the outcome, this student concludes: “I studied hard and it did not matter. I must not be good at this.” That conclusion is wrong. And it is the system, not the student, that produced it.

Here is where this connects to the work you do every day.

Grades are outcome measures. GPAs are outcome measures. Standardized test scores are outcome measures. They tell you the scoreboard. They tell you nothing about the quality of decisions that produced the scoreboard.

A system that evaluates only outcomes is a system that teaches students to optimize for outcomes. Not for learning. Not for capacity. Not for the quality of their process.

And here is the connection to the overhelping research from our last issue. When we overhelp students, we are doing something very specific inside Duke’s matrix. We are moving them from the bad-decision column to the good-result row without changing the quality of their decisions. The scoreboard improves. Decision-making capacity does not.

The student who was scaffolded into a passing grade did not learn to make better decisions about how to learn. They learned that the system will carry them to a result. That is bottom-left masquerading as top-left. And it compounds.

Duke’s framework suggests three shifts for any system that evaluates performance.

First, separate decision evaluation from outcome evaluation. A student who prepared well and bombed the exam should not receive the same feedback as a student who did not prepare and also bombed the exam. The outcome is identical. The intervention should not be.

Second, make the process visible. Runners log every training run. They know exactly which decisions led to race day. Students rarely have this kind of decision audit trail. The systems that build it — metacognitive journals, process portfolios, self-assessment rubrics — are systematically undervalued compared to test prep.

Third, stop rewarding the bottom-left quadrant. When a student crams and passes, the system currently says: good job. Duke’s framework says: you got lucky. That is not cruelty. That is accuracy. And it is the only way to build students who can sustain performance over time rather than spike it once and collapse.

The marathon is a useful metaphor because it strips the pretense away. No one believes a runner who did not train deserves a medal just because they crossed the line. But we hand diplomas to students every year who were carried across the line by systems designed to ensure they would not fail. And then we wonder why they cannot perform independently.

The unlock is not better scoreboards. It is better decisions. And systems that know the difference.

Not better scores. Better decisions.

Framework: Duke, A. (2018). Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts. Portfolio/Penguin.

Duke’s core argument is that every decision is a bet on a future outcome under uncertainty. Your grading policy is a bet, too. It is a bet that measuring outcomes will produce better learning. The evidence for that bet is thin.

Consider what your current system tells a student who studied for three weeks, used effective strategies, managed their time, sought help when stuck, and then scored 62% on the test because the content was genuinely difficult for them and they had not yet reached mastery.

Your system tells that student: D. Borderline failure. The process was excellent. The outcome was not. And the system cannot distinguish between the two.

Now consider the student who did none of those things, crammed for two hours, and scored 71%. Your system tells that student: C. Passing. Move on.

One of those students is building capacity. One of them is not. Your grading system cannot tell you which is which. That is the design problem.

Not smarter grading. Smarter information.

Every district website in the country mentions growth mindset. Most of them are lying. Not deliberately. Structurally.

Growth mindset, as Carol Dweck originally defined it, means believing that ability is developed through effort and strategy. It means that struggle is information, not verdict. It means that a bad result from a good process is valuable data, not evidence of limitation.

Now look at what happens in most schools that claim to teach growth mindset. A student struggles. The system intervenes to remove the struggle. The student passes. The poster on the wall says “mistakes help us grow.” The gradebook says otherwise.

If your school teaches growth mindset but evaluates only outcomes, you are telling students two contradictory things. The poster says: your effort matters. The transcript says: only your results do.

Students are not stupid. They resolve the contradiction by ignoring the poster.

Not growth mindset culture. Growth mindset architecture.  

Every marathon training plan has a run that nobody wants to do. It is usually a 30-to-35-kilometre run on a Saturday morning in February when it is minus-fifteen and the sidewalks are ice. You do not want to do it. You do it anyway. And three months later, at kilometre 35 of the race, your legs remember.

That run does not show up on the scoreboard. There is no medal for “completed the training run you hated in weather that hated you back.” But every runner knows: that run is the race. The 26.2 miles on race day is just the receipt.

The teachers who understand this are the ones who let students sit in the discomfort of not knowing. Who resist the urge to give the answer when the silence gets uncomfortable. Who grade the annotated draft, the failed attempt with margin notes, the problem set that was wrong but showed clear thinking.

Those teachers are running the system Duke describes. They are evaluating decisions, not outcomes. They are building students who can sustain performance, not students who can spike it once when the conditions are right.

The shout-out this week goes to every teacher who has watched a student struggle and chosen to stay in the room rather than hand over the answer. That is not inaction. It is the hardest form of teaching there is. And it is the form that builds runners who finish, not runners who are carried.

 ________________________________________________________

That's all the unlock for today. Tune in next week.

Stay awesome, you unlockers!

Want more? Order The Great Unlock here.