Be confident — but not too confident

Do you lack confidence? Are you afraid to set up a new business, embark on a new career, commit to a relationship or take up hang gliding?

Don’t worry too much about it. You might be making the right decisions. Being too confident can be worse than not being confident enough.

But how can you tell? Turn to Don A. Moore’s new book Perfectly Confident. It’s about making the best decisions.

            Moore says most popular treatments assume that more confidence is better. People just need to overcome their fears and jump in. This is true for some people and some decisions. But it can also be disastrous.

When you see a sports star making a seemingly brash prediction of winning, you might imagine that being really confident is necessary for success. After all, if you’re not confident, how can you do your best? Not so quick, says Moore. There’s actually little evidence that super-confidence improves performance. Those sports stars have worked hard and long, and may be making reasonable judgements about their chances of victory.

Overconfidence is potentially dangerous and can lead you to take unwarranted risks. If you’ve never tried base jumping, it’s better to be very cautious and prepare carefully before your first jump. Most new small businesses fail within the first year. Perhaps their owners were overconfident.

            There is evidence that most people overestimate how good they are at things. In a classic survey, 93% of US drivers said they ranked in the top half. Most young people think they are more honest than average and better than average at relationships. The reason is that people think, “I’m honest most of the time, so I’m better than average” but don’t stop to think that most other people may think the same way. Moore says the way to fix your perception of superiority is to be more specific. For example, if being a good driver is specified as never having had an accident or a ticket, then fewer people will overestimate their abilities.

There’s another side to people’s thinking about their own capabilities. When it comes to an uncommon skill, like riding a unicycle or subtracting large numbers in your head, most people underestimate their abilities. You might think, “I wouldn’t last three seconds on a unicycle” and forget to think that most other people might have the same difficulty.


Could you unicycle across China?

            One of the methods Moore recommends is to think probabilistically. Consider all possible outcomes of your decision. Consider the new business. You might guess that there’s a 10% chance of making a lot of money, 40% of making a little, 30% of losing a little and 20% of losing a lot. Just writing down the possibilities can be sobering. Overconfident people never stop to think of failure and hence can make unwise decisions. Assigning probabilities also helps in overcoming the tendency to think in terms of yes or no, success or no success.

You also need to weigh up the benefits against the costs. In setting up the business, you might be working 90-hour weeks. This can be exhilarating but it might also be exhausting. You should factor these possibilities into your decision. Vital here is the idea of opportunity cost. All that money and those hours of effort might be invested in some other activity. Thinking in terms of different possible outcomes and opportunity costs can help counter overconfidence.

A confident scholar?

Many times in my career as an academic I’ve had to make decisions about whether to write an article or a book and then, after writing it, where to submit it. When I was first starting out, I’d write an article and then try to figure out where to submit it. Before long, I learned this was not a good strategy, because sometimes there was no suitable outlet. Moore would say I was overconfident and needed to consider the possibility of wasting effort, at least for the purpose of publication, which is crucial for aspiring academics.

These days, before writing an article, I think about where I plan to send it, and the likelihood of it being accepted. Sometimes there is a high-prestige journal that I think could be worth trying. I might estimate the chance of acceptance as 5 percent, one out of twenty. I have to weigh up the effort of tailoring the article to this journal and going through the admission process, along with associated delays, against the 95% chance of rejection.

In many cases, I decide not to bother with the high-status journal and go straight to one where the odds are better. This points to another factor to consider when writing an article: are there fall-back options should my first-choice outlet reject my submission?

Another decision is whether to undertake a PhD. When I did my own PhD, aeons ago, I didn’t think about failure. I took a risk without considering the full range of outcomes. Now, as a potential PhD supervisor, I regularly talk to prospective students. They need to make several decisions: whether to pursue a PhD, what university to attend, what topic and what supervisor. It’s a big decision because writing a PhD thesis requires years of effort. Although about three quarters of students who’ve started with me as their supervisor have graduated, the cost for those who don’t finish can be large: they could have been doing something else with their time and energy. On the other hand, a student can acquire skills and obtain satisfactions along the way, a sort of consolation prize for non-finishers.

            Therefore, in advising prospective students, I point to the large and sustained commitment required and note that most PhD graduates do not obtain academic posts. After reading Moore’s book, in future I’ll recommend that prospective students assign probabilities to different outcomes. That will help counter overconfidence.

For students who are part way through their theses, a more common problem is under confidence. The challenge seems enormous. It can be helpful to have the courage to continue, knowing that most students, including most of those who finish, go through periods of self-doubt.

A confident whistleblower?

Another area where Moore’s recommendations are relevant is whistleblowing. Thinking from the point of view of managers in organisations, he says that being results-oriented is not necessarily a good thing. Being results-oriented often means rewarding employees for success and penalising them for failure.

This sounds logical but it misses an important consideration: sometimes it is wise to take risks even though some of them don’t pan out. If developing a new app costs $1 million and has only a 10% chance of success, it’s still a good bet if success means a return of $100 million. But when employees are penalised for failure, they won’t take risks like this. Apple never would have developed spectacularly profitable devices if it hadn’t supported risk-taking with positive expected returns.

Imagine being a manager and one of your employees reports possibly fraudulent activities in the organisation. You investigate and discover your employee is wrong. Does this warrant a penalty? Moore would say that this whistleblower should be encouraged even if the report was wrong, at least if there’s a reasonable chance it might have been right.

In practice, employees who make allegations of wrongdoing are often penalised even when they’re right, especially when the wrongdoing implicates higher management for being involved or for tolerating it. That’s another story.

            The whistleblower, treated badly, then turns to a watchdog agency such as an ombudsman or anti-corruption agency or court. A good idea? Moore’s advice would be to consider all possible outcomes and assign them probabilities, and also to consider other options. Few whistleblowers do this. They want vindication and assume that some higher authority will provide it. They do not investigate the success rate of previous whistleblowers, which can be abysmal. Because they know they are right, they do not consider the possibility that justice will not be done, and that many previous whistleblowers also knew they were right but failed in their efforts to be vindicated.

Moore recommends learning from experience. When you have a decision to make, assign probabilities to potential outcomes and consider alternative courses of action. When you learn the outcome, go back to your probabilities and figure out whether you may have been too confident or not confident enough. Gradually, over time, you can improve your skill in predicting outcomes.


Don’t just jump in! Learn to predict outcomes.

            This is good advice for many purposes. However, when you’re faced with a decision that is likely to be made just once in a lifetime — like doing a PhD or blowing the whistle — then it’s sensible to learn as much as possible about what others have done in the same situation. Why make your own mistakes when you can learn from others’ mistakes? By undertaking this sort of investigation, you minimise the risk of making a wrong decision. And when things don’t work out, remember that you still might have made the right decision. If you are successful in everything you try, you probably aren’t taking enough risks!

Brian Martin
bmartin@uow.edu.au