It looks like your browser is not up to date.
This means our website may not look and work as you would expect. Read more about browsers and how to update them here.

Skip to main content
  • Register
  • Help
  • Contact us

Superforecasters: how to make better predictions

We take a closer look at the science behind forecasting and how investors can become better forecasters.

Important notes

This article isn’t personal advice. If you’re not sure whether an investment is right for you please seek advice. If you choose to invest the value of your investment will rise and fall, so you could get back less than you put in.

In 2011 the Intelligence Advanced Research Projects Activity (IARPA) ran a competition to try and improve the forecasting ability of US intelligence services. Different teams gave forecasts which could be rigorously tested, and the results were compared to see which methods worked best.

Philip Tetlock, who had been researching forecasts for decades, entered with the Good Judgement Project – and won so convincingly that the competition was abandoned after two years.

Tetlock went on to write a book called Superforecasters about the science and skill of forecasting. His research suggests that prediction is a skill, and that some people are measurably better at it than others. However, Tetlock’s research also shows that superforecasters, as he calls them, have a lot in common. This means we can study the habits of accurate forecasters and improve our own ability to try and predict the future.

Investing is always, to some degree, about making forecasts. Whether we’re expecting a company to grow or the economy to recover after a shock. So, learning a bit about the science of forecasting should help make us better investors.

In this article I’ve pulled some of the lessons out of Tetlock’s book, but I do recommend the whole thing if you find this subject interesting. There’s a lot more in there than I could fit into this article.

This article isn’t personal advice. If you’re not sure if a course of action is right for you then you should seek advice.

Guts

Superforecasters tend to do well on questions like the three below. Try them before reading on to see how you do.

  1. A bat and a ball cost £1.10 in total. The bat costs £1.00 more than the ball. How much does the ball cost?
  2. If it takes 5 machines 5 minutes to make 5 widgets, how long would it take 100 machines to make 100 widgets?
  3. In a lake, there is a patch of lily pads. Every day, the patch doubles in size. If it takes 48 days for the patch to cover the entire lake, how long would it take for the patch to cover half of the lake?

What ties these questions together is that they all have an answer which, to misquote H.L. Mencken, is simple, obvious and wrong. Go back and try them again to see if any caught you out. (The answers are at the bottom of this article if you’re stuck.)

These questions are part of what psychologists call a Cognitive Reflection Test, which tests a person’s tendency to slow down and think instead of latching onto the first intuition that springs to mind. The questions are actually pretty straightforward, but you probably found that an obvious wrong answer jumped off the screen. That’s normal, but it’s the mark of a great forecaster to ignore these gut instincts and really think about a problem.

We all make quick intuitive judgements, and while sometimes our gut can be a good guide, forecasting is not one of those times. The lesson is clear: when making predictions or investment decisions, slow down and think carefully. Do not trust your gut.

Three settings

People aren’t good at thinking probabilistically. Lots of people only have two settings when thinking about the future – yes or no. Will Chelsea beat Man City in the Champions League Final? Yes. Will England win the Euros? No.

More sophisticated thinkers add a third possibility – maybe. Will the Bank of England raise interest rates this year? Maybe.

But even three settings – yes, no, and maybe – are pretty poor. They leave out all the possible gradations between ‘almost certain’ and ‘vanishingly unlikely’. Superforecasters don’t think like that.

Superforecasters think in shades of grey and make finely calibrated judgements. Will it rain tomorrow? 18% chance.

Unfortunately, even professional analysts sometimes operate with a three setting dial – and lots of equity research houses often offer just three recommendations: buy, sell or hold.

As investors, we should strive to add more degrees of subtlety to our thinking. Instead of confidently predicting that something is going to happen, we should learn to distinguish between the somewhat likely and the highly likely – the 60/40 chance and the 80/20.

Superforecasters routinely give estimates as fine grained as an 18% chance, and when researchers rounded these to the nearest 10% their accuracy fell. So, always try to be as precise as you can and avoid being vague.

How to make a forecast

Step 1. Pick something that’s actually predictable

It’s probably fair to say that in the summer of 2015 we all would have given a wrong answer to the interview question ‘where do you see yourself in five years’ time?’.

The world contains a lot of randomness, and the further out we try to forecast, the greater the opportunity for randomness to intervene. Furthermore, some things are inherently more predictable than others. For example, weather forecasts are pretty good over a few days but virtually useless over a few weeks.

And some subjects make better fodder than others. We can all predict that the sun will rise tomorrow, but some things, like currency markets, are notoriously unpredictable.

So, choose carefully and recognise that the further away something is, the less reliable your forecast is likely to be.

Step 2. Break the problem up into chunks

Always break your forecast into as many smaller chunks as you can.

For example, rather than estimating a company’s profits, estimate sales and costs. Then you can subtract one from the other to get to profits. You could even break this down further by estimating sales from different product lines and different types of costs.

Breaking the problem into chunks forces you to think carefully about every assumption you make, and ultimately makes for better forecasts.

Step 3. Take the outside view first

Tetlock distinguishes between the ”outside view”, which means looking at an event as part of a broader class, and the “inside view”, which means examining the characteristics of a specific case.

For example, say you wanted to forecast the likelihood that the UK market will fall in 2022. It would be tempting to look at the market’s valuation and read the news about vaccines and Indian variants, but that would be the inside view because these facts are specific to 2021. Instead, what you want to know first is – how often does the market fall in general?

It turns out that, before dividends, the UK market has fallen in 31% of all calendar years since 1963. That’s the outside view, and it should usually be your starting point.

Once we’ve established the outside view, in this case 31%, we can then adjust this up and down based on the specific conditions of this year – the inside view.

Remember past performance isn’t a guide to future returns. All investment can fall as well as rise in value, so you could get back less than you invest.

Step 4. Get more opinions.

As part of the Good Judgement Project, Tetlock found that teams make more accurate forecasts than individuals. It seems that the average of several forecasts is likely to be better than any individual one.

Now, most of us don’t invest in teams, but we can discuss our investments and strategy with others. We can also read widely to get more perspectives, especially from people we disagree with. In fact, looking for evidence that contradicts your forecast is probably a lot more effective than looking for evidence that confirms it.

Step 5. Keep up to date.

One of the hallmarks of a good forecaster is a willingness to update your beliefs with new information. For investors, this means reading company announcements and staying abreast of the news. If something changes, then you need to be able to flexibly change your views and possibly your investment strategy. But, you also need to avoid swinging about like a weathervane.

Tetlock’s research suggests that small, frequent updates to your views are better than large but rare course changes. As investors we must also be careful not to rack up unnecessary fees by overtrading and should always focus on the long term – by long term, we mean at least five years. But the principle of keeping an open mind, checking in on your investments reasonably frequently and staying on top of the news is a sound one.

Register for the latest share research

Answers

  1. The ball costs 5p, and the bat costs £1.05 – £1.00 more than the ball. But you may have jumped straight to 10p.
  2. 5 minutes. Each machine makes a widget every five minutes, so the answer is the same regardless of how many machines you have. But you might have rushed to 100 minutes.
  3. 47 days. The patch doubles in size every day, so it would have covered half the lake on day 47 and then doubled to cover the whole lake on day 48. But you might have quickly halved the number of days and got 24.

Editor’s choice: our weekly email

Sign up to receive the week’s top investment stories from Hargreaves Lansdown

Please correct the following errors before you continue:

    Existing client? Please log in to your account to automatically fill in the details below.

    Loading

    Your postcode ends:

    Not your postcode? Enter your full address.

    Loading

    Hargreaves Lansdown PLC group companies will usually send you further information by post and/or email about our products and services. If you would prefer not to receive this, please do let us know. We will not sell or trade your personal data.


    What did you think of this article?

    Important notes

    This article isn’t personal advice. If you’re not sure whether an investment is right for you please seek advice. If you choose to invest the value of your investment will rise and fall, so you could get back less than you put in.

    Editor's choice – our weekly email

    Sign up to receive the week's top investment stories from Hargreaves Lansdown. Including:

    • Latest comment on economies and markets
    • Expert investment research
    • Financial planning tips
    Sign up

    Related articles

    Category: Investing and saving

    Inflation, interest rates and government spending – what’s the latest?

    Following the latest G7 meeting, George Trefgarne looks at the latest on inflation, interest rates and government spending, and what they could mean for investors.

    George Trefgarne

    23 Jun 2021 4 min read

    Category: Investing and saving

    3 stocks that don’t rely on the digital revolution

    While the digital revolution is well underway, not everything is going digital. Here’s a look at three stocks that could benefit from enduring old-school behaviour.

    Sophie Lund Yates

    23 Jun 2021 8 min read

    Category: Investing and saving

    How spring cleaning your savings could pocket you £80 more

    Saving with a high street bank? Here’s how you could get a better interest rate on your savings and qualify for cashback as a thank you at the same time.

    Ryan Kenny

    23 Jun 2021 8 min read

    Category: Investing and saving

    What we’ve learnt about money by working in financial services

    Part 1 of our ‘What I Learned’ series – part of our Financially Fearless initiative for women. We share the money tips we’ve learnt by working in financial services that could help other women like us.

    Hannah Miles

    18 Jun 2021 6 min read