How to decide better

What practice changes, and what it does not

Start with the sobering evidence. In 2016 a team led by Daniel Simons reviewed every peer-reviewed study cited by leading brain-training companies. They found extensive evidence that training improves the trained task, less evidence that it improves closely related tasks, and little evidence that it improves distant tasks or everyday cognitive performance. Getting better at a balloon game will make you better at balloon games. It will not, by itself, make you a better manager.

Decision research offers a more hopeful thread, because it trains strategies rather than raw capacity. Carey Morewedge and colleagues gave people a single training session, either an interactive game or an instructional video, on common decision biases such as anchoring and confirmation bias. Bias fell immediately and stayed lower two months later. The interactive version, which gave personalised feedback and practice, cut biases by at least 23.57 per cent after two months, compared with at least 19.20 per cent for the video. The effects also appeared on problems in formats the training never showed.

A later field test is the strongest result. Graduate students in three professional programmes received a one-shot debiasing training either before or after they solved an unannounced business case, modelled on the decision to launch the Space Shuttle Challenger. Those trained first were 19 per cent less likely to choose the inferior, hypothesis-confirming solution.

What about the games themselves? In the 2017 Business Insider report on Unilever's hiring, pymetrics' chief executive said practice would not significantly alter a person's performance on its games. Take that as a claim about style: your broad appetite for risk and trust is fairly stable. It is not a claim that understanding the task is useless. Candidates who misread the rules, panic after one loss or never notice a pattern produce results that describe their confusion, not their personality. Practice removes that noise.

Seven habits that carry across tasks

These are the strategies that decision research and our own scoring agree on. Each game lesson turns them into a concrete method.

1. Price every option before you choose

   Ask what each option pays on average, per round. You rarely need exact numbers. "This deck wins 100 a card but has cost me 1,250 in ten cards, so it averages minus 25" is enough to act on.

2. Judge in blocks, not single outcomes

   One card, one balloon or one partner's return is a tiny sample. Count outcomes over a block of eight or ten before concluding anything, and keep a running tally in your head.

3. Explore early, exploit late

   Spend your testing budget at the start, when there is time to profit from what you learn. By the last third of a run, you should mostly be using your best-known option.

4. Update in proportion

   A single surprise should move your estimate a little. A run of surprises in the same direction should move it a lot. Both overreacting and ignoring feedback cost value.

5. Set a stopping rule before you start

   Decide in advance where you will stop (bank the balloon at this size, stop sampling after these draws) and change the rule only on evidence, not on a feeling in the moment.

6. Read people by what they do

   Invest where your trust is repaid, pull back where it is not, and keep a cheap test running in case someone changes. Punishing unfairness at your own cost pays only if it changes what happens next.

7. Review with numbers

   After each session, look at the value you captured and one descriptive metric. Write one sentence on what you will change. That sentence is where the learning happens.

Why blocks beat single outcomes

The trap habit two protects you from is the one Hertwig and Erev describe: when you learn from experience, rare events carry less weight than they deserve. Watch the ten cards below from a deck that pays well on nine of them.

Ordering exerciseSign in to try it

If you judged that deck on any five consecutive cards that missed the big loss, you would call it excellent. Judged over the whole block, it is the worst deck in the game. That is why the tally, not the last card, should drive your choice.

Why exploration belongs at the start

Wilson and colleagues found people naturally explore more when more choices remain. Use that instinct deliberately: in a 30-draw card run, a single wasted draw in the last five cannot be recovered by anything you learn from it.

Why proportion matters

Our Card Decision report estimates your learning rate from your choices, and our Balloon Pump report shows how much you changed after a burst. Both are descriptive, but both tell you which way you lean.

Judge the decision, not the outcome

Reviewing well is harder than it sounds, because outcomes shout and reasoning whispers. In five studies, Jonathan Baron and John Hershey gave people descriptions of other people's decisions, with all the information the decision maker had, and then told them how things turned out. People rated the same thinking as better, and the decision maker as more competent, when the outcome happened to be good. They did this even though, when asked, they agreed outcomes should not count.

You will do the same to yourself. Bank a balloon at a sensible size and watch the next one of that type burst late, and you will feel you were too timid. Draw from a good deck and hit its one big loss, and you will feel you chose badly. Neither feeling is evidence. The question to ask in review is: given what I knew at the time, was the choice sound? Only a pattern across many outcomes can tell you that the reasoning itself was off.

Carrying the habits into real decisions

The habits only transfer if you use them deliberately outside the games. A few translations that fit student and early-career life:

• Pricing options: before choosing between two offers, write down what each is worth on average over two years, including the chance a promised promotion does not happen. You are estimating an expected value, even if the numbers are rough.
• Blocks, not single outcomes: judge a study method, a supplier or a colleague on ten instances, not on the most recent one.
• Explore early: in your first months in a new team, try several ways of working and ask several people for advice. Later, settle on what works and go deep.
• Stopping rules: decide in advance how many companies you will apply to before you review your approach, or how long you will research a purchase before you buy.
• Reading people: extend reasonable trust to a new teammate on small tasks, increase it when they deliver, and keep checking small things when they do not.
• A decision log: for any choice that matters, write one line on what you expect and why. Compare it with what happened a month later. Over a year, this is the most honest calibration feedback you will ever get.

A four-week plan

The plan below assumes about 20 minutes a day, four or five days a week. Each run of our decision games is one trial, so a session of two or three runs per game is plenty. Review after every session: the value-captured percentage, plus the one descriptive metric named in the plan.

1. Week 1: learn the tasks
   • Read the four game lessons before your first run of each
   • Two runs each of Card Decision and Balloon Pump; note how many draws it took to settle on a deck
   • One run each of Trust Game and Money Exchange; note which partners repaid you
   • Write down your stopping rule for each balloon type before starting
2. Week 2: learning from feedback
   • Three sessions of Card Decision; keep a per-deck tally aloud
   • Aim to settle within the first eight to ten draws
   • Check the report's draws-before-good-decks metric and try to lower it
   • One Balloon Pump session to keep the stopping rule fresh
3. Week 3: risk and stopping
   • Three sessions of Balloon Pump; bank near the middle of each type's range
   • Compare the report's risk appetite with the ideal: close to zero is calibrated
   • Watch the change-after-a-burst metric; aim for small, steady adjustments
   • One Card Decision session as a check
4. Week 4: people and review
   • Two sessions each of Trust Game and Money Exchange
   • In Trust Game, apply the one-third rule and keep a small test running with partners who let you down
   • In Money Exchange, accept every offer and walk your own offers down gradually
   • Finish with one run of each game and compare value captured with week 1

Remember

• Brain training improves the trained task; far transfer is weak.
• Strategy training is different: one debiasing session cut biases for at least two months, and improved a real business-case decision.
• Pymetrics says practice barely changes its game results. Style is stable; misreading a task is not style.
• Price options, judge in blocks, explore early, update in proportion, pre-commit to stopping rules, read people by behaviour, review with numbers.
• Judge a decision by what you knew when you made it; outcome bias makes lucky choices look wise.
• Review each session with value captured plus one descriptive metric, and write one sentence on what to change.

When you are ready, test yourself in the decision making checkpoint.

Sources

1. Simons, D. J., Boot, W. R., Charness, N., Gathercole, S. E., Chabris, C. F., Hambrick, D. Z., & Stine-Morrow, E. A. L. (2016). Do "brain-training" programs work? Psychological Science in the Public Interest, 17(3), 103–186. doi.org/10.1177/1529100616661983
2. Morewedge, C. K., Yoon, H., Scopelliti, I., Symborski, C. W., Korris, J. H., & Kassam, K. S. (2015). Debiasing decisions: Improved decision making with a single training intervention. Policy Insights from the Behavioral and Brain Sciences, 2(1), 129–140. doi.org/10.1177/2372732215600886
3. Sellier, A.-L., Scopelliti, I., & Morewedge, C. K. (2019). Debiasing training improves decision making in the field. Psychological Science, 30(9), 1371–1379. doi.org/10.1177/0956797619861429
4. Feloni, R. (2017, June 28). Consumer-goods giant Unilever has been hiring employees using brain games and artificial intelligence. Business Insider. businessinsider.com
5. Hertwig, R., & Erev, I. (2009). The description–experience gap in risky choice. Trends in Cognitive Sciences, 13(12), 517–523. doi.org/10.1016/j.tics.2009.09.004
6. Wilson, R. C., Geana, A., White, J. M., Ludvig, E. A., & Cohen, J. D. (2014). Humans use directed and random exploration to solve the explore–exploit dilemma. Journal of Experimental Psychology: General, 143(6), 2074–2081. doi.org/10.1037/a0038199
7. Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of Personality and Social Psychology, 54(4), 569–579. doi.org/10.1037/0022-3514.54.4.569
8. Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S. W. (1994). Insensitivity to future consequences following damage to human prefrontal cortex. Cognition, 50(1–3), 7–15. doi.org/10.1016/0010-0277(94)90018-3