9. Foresight

Foresight is a set of methods to:

• Improve our anticipation of what might otherwise be unexpected
• Enable us to rationally assess the credibility, and the desirability, of various possible future outcomes
• Prepare us to respond better if these outcomes (or something like them) transpire
• Help us understand which actions would be most effective in altering the likelihood of various scenarios.

Resources providing an overall introduction to foresight:

• The Foresight Guide by John M. Smart and colleagues
• The Teach The Future library by Peter Bishop and colleagues
• The School for the Future of Innovation in Society at ASU (Arizona State University)
• (More pending)

9.1 Forecasting

9.1.1 When forecasts were mistaken

Many forecasts in the past turned out to be incorrect: what was forecast did not actually happen.

However, a review of these past failures shows that, provided we avoid various traps, foresight remains a skill that is valuable and critically important.

Examples of actual (or perceived) forecasting failures include:

• Forecasts of an ongoing Malthusian tragedy, with populations continuing to be subject to famine and plague – forecasts repeated in the 1960s and 1970s by Paul Ehrlich
• Forecasts that various outbreaks of flu would become dangerous global pandemics – whereas they had only limited impact
• Forecasts of an impending mini ice age – seemingly replaced by forecasts of impending global warming
• Forecasts that stock prices (and/or house prices) would inevitably continue to rise
• Forecasts of technology rendering large numbers of people unemployable
• Failures of election polls to correctly predict what were “shock” results
• Failures to foresee important military attacks, coupled with delusions of military impregnability – such as the Maginot Line that would supposedly defend France from invasion in any World War Two
• Forecasts of worsening terrorist attacks, with access to weapons of mass destruction

(Video to be added here, taking material from Chapter 3 of Vital Foresight.)

(More to be added here)

9.1.2 When forecasts proved invaluable

(Examples to be added here of the forecasts that turned out to be the most valuable, and where the success cannot be ascribed to “luck” or “statistics”.)

9.1.3 Forecasters and superforecasters

This section covers the main findings of Philip Tetlock, as covered in, for example, the 2015 book Superforecasting.

The research of Philip Tetlock is informed by a large set of predictive challenges about events that might, or might not, happen a short time into the future. Some examples:

• “In the next year, will any country withdraw from the eurozone?”
• “Will India or Brazil become a permanent member of the UN Security Council in the next two years?”
• “When will the NYC public schools resume in-person instruction for at least some students?”
• “When will enough doses of FDA-approved Covid-19 vaccine(s) to inoculate 200 million people be distributed in the United States?”

A group of several thousand volunteers provide answers on a regular basis to these questions. These forecasters are asked to provide in each case, not only their predictions, but also the confidence with which they make these predictions. For example, if a challenge question offers three possible date ranges as answers, a forecaster might assign 5% probability to the first date range, 15% probability to the second, and 80% to the third.

The accuracy of an individual forecast can be calculated by something known as the “Brier score”. It’s the average of the squares of the errors. In the above example, if the correct answer turns out to be the third option, the accuracy is calculated as (0.05-0)²+(0.15-0)²+(0.8-1)², divided by 3, that is 0.022. If the correct answer had been the second option, the accuracy would have been (0.05-0)²+(0.15-1)²+(0.8-0)², divided by 3, that is 0.755. Evidently, lower scores indicate a better forecast. The smallest possible score is 0, and the largest is 1.

In this setup, forecasters are able – and indeed are encouraged – to revise their forecasts in the light of new information received. For example, a forecast about the likelihood of a country exiting the eurozone could be updated in the light of comments from a country’s finance minister. The assessment of a forecaster’s performance takes into account their predictions throughout the period of time the question remained open.

Some predictive challenges turn out to be easier than others. What’s of particular interest is how well the scores of a forecaster compare with the median (average) score of the overall community of forecasters. This is known as their relative Brier score.

The first important finding from this project is that some forecasters are systematically better than the crowd. They’re not just accurate on occasion; they tend to make better forecasts nearly all of the time. Tetlock calls these individuals “superforecasters”.

The second important finding – actually a set of findings – is that these superforecasters are characterised by a number of personal attributes. These are attributes which can be studied and learned – which is good news for those of us who aspire to improve our forecasting abilities.

Tetlock summarises this finding as follows:

Broadly speaking, superforecasting demands thinking that is open-minded, careful, curious, and – above all – self-critical. It also demands focus. The kind of thinking that produces superior judgment does not come effortlessly. Only the determined can deliver it reasonably consistently, which is why our analyses have consistently found commitment to self-improvement to be the strongest predictor of performance.

Here are some of the characteristic attributes of superforecasters:

1. Rather than insisting on seeing things in binary (“black and white”) terms, superforecasters are comfortable with probabilistic reasoning
2. They prefer precise numerical language over vague terms (such as “might”) whose apparent meaning can change with hindsight
3. Rather than “fate” or “inevitability”, they can accept contingency, such as the chaotic amplification of the flapping of a seagull’s wing, as often being the factor responsible for an outcome
4. They consider hypotheses about the potential influences on outcomes, and then seek out evidence that would either confirm or refute these hypotheses
5. They are ready to incrementally adjust their forecasts in the light of new evidence, and they avoid investing emotional energy in defence of a prior forecast out of some misplaced sense of loyalty
6. Critically, they prioritise looking for evidence that would disprove their current favoured hypothesis, rather than just trying to accumulate more data points that are compatible with it
7. Rather than accepting an argument merely on the say-so of a fellow forecaster who has a good track record, they subject that argument to their own scrutiny, thereby avoiding the perils of groupthink
8. They believe their forecasting abilities are by no means fixed or innate, but can be improved via a combination of practice and review
9. Accordingly, they regularly reflect on what they can learn from their past forecasting experiences – both successes and failures.

One more characteristic attribute of superforecasters deserves particular attention. This attribute is whether someone is a big idea “hedgehog” thinker or a pragmatist “fox” thinker. This terminology, by the way, was borrowed from the philosopher Isaiah Berlin, who had in turn adapted it from this stanza by the Greek lyric-poet Archilochus:

The fox knows many things; the hedgehog one big thing.

The “big idea” of someone in the hedgehog category was a core belief around which they sought to organise all the rest of their thinking. The belief in question varied from hedgehog to hedgehog. It might be socialism, with its preference for state control of the economy. It might be free-market fundamentalism, with its desire to reduce regulation. It might be a pessimism about the environment, or an optimism about the potential of technology to fix every problem, or so on. Members of this category tended to disregard as irrelevant potential evidence that lay outside their preferred filters. The best understanding, they assumed, would come from remaining focused on the most important principles.

In contrast, forecasters in the fox category were each willing to embrace a wide variety of different analytical styles, conceptual tools, and procedural methods. Evidence from multiple sources was equally welcome. They had no “sacred dogmas” that constrained their thinking. Rather than being ideological purists, they were open-minded about the best approach to each individual forecast challenge.

The evidence from Philip Tetlock’s research is emphatic. Forecasting foxes consistently outperformed forecasting hedgehogs. For better forecasts, ideology is a hindrance, rather than an enabler. Being self-assured may win you a crowd of followers who are impressed by what they see as charisma, but it will hinder your ability to pay proper attention to countervailing evidence or to reach a deeper understanding of particular circumstances. Having simple answers ready to deploy in all situations may result in repeat invitations to talk shows or keynote presentations, but the predictions you make at such events will likely turn out to be wrong – or to be couched in such vague language as to be without value.

9.1.4 Metaphors: Animals and beyond

(About “black swans”, “grey rhinos”, “the elephant in the room”, etc)

9.2 Trend analysis

9.2.1 Identifying trends that have the potential to cause disruption

9.2.2 The Gartner Hype Cycle, and why it often misleads

9.2.3 The factors behind potential exponential change

9.2.4 Identifying potential accelerators – and brakes – for trends

9.2.5 Establishing “canary signals” for advance warning of tipping points

9.2.6 Noticing “signals” that don’t match previous assumptions

9.3 Scenarios

9.3.1 The purpose of scenarios

Analysis of the purpose of scenarios:

The example of the Mont Fleur scenarios project, South Africa, 1992:

9.3.2 How trends can combine to form breakthrough scenarios

9.3.3 PESTEL classification of types of trend

9.3.4 Examples of interconnections producing unforeseen scenarios

9.3.5 Using imagination to anticipate novel scenarios

9.3.6 Assessing the credibility of imagined scenarios

9.3.7 Assessing the desirability of imagined scenarios

9.4 Transformational foresight

9.4.1 Assessing actions to influence the actual course of scenarios

9.4.2 Precautionary and proactionary approaches to risk and opportunity

9.5 Deepening foresight skills

9.5.1 Agile futurism

Regular exercises to develop and deepen, in stages, what can be called “foresight muscles”:

• Foresight of possible threats and opportunities
• Foresight of possible responses (new tech, biz model change…)
• Ability to handle unexpected changes more quickly
• Become “inoculated against future shock”

In each sprint (2-4 weeks, perhaps), select a small, focused task

• E.g. “understand trend T”
  • Identify potential accelerators, brakes, and canary signals
• Or “explore scenario S”
  • With a small number of trends
• At the end of the sprint, present the findings to peers
• Review feedback
  • Pay particular attention to blockages that hindered progress
• Iterate and/or Pivot
  • Decide which trend(s) and/or scenario(s) to feature in future sprints

Need to be able to learn from successes and from failures

• The meanings of “Fail fast”, “Fail forward”, “Fail smart”
• The cultivation of emotional resilience
  • Not to fear failure
  • Not to penalise failure

The framework for agile futurism

• Determine metrics that value the growth of relevant knowledge and skills
• Resist being dominated by the metrics applicable to the “business as usual” parts of organisations
• Regularly rotate people from the wider organisation into and out of the agile foresight team

9.5.2 Collaborative futurism

There are too many skills relevant to foresight for any one person (or team) to master. These include:

• Trend factors: Political, Economic, Social, Tech, Environmental, Legal
• How to usefully imagine new possibilities, “outside the box”, subtractions…
• How to reduce uncertainty, cumulatively, one sprint at a time
• Maintaining sense of urgency, over multiple time periods

It is therefore particularly important to be able to collaborate well:

• Identifying partners, communities, mentors
• Knowing when to break (or deprioritise) relationships that were previously important
• Identifying independent knowledge repositories
• Being sure to grow internal knowledge and skills too, rather than being overly dependent on external resources

Foresight methods that utilise collaborative intelligence include:

• Delphi Questionnaires – when different respondents can view each others’ submissions (answers and explanations), and adjust their own submissions in response

9.6 Constructive future-oriented fiction

There are many drawbacks with most fiction about the future:

• Writers prefer sensational plotlines to credible ones
• Lots of fiction has the purpose to provide commentary on present-day circumstances, rather than to suggest truly plausible future scenarios
• Most fiction about the future changes only a few aspects of the human condition, whereas the likelihood is that a more comprehensive set of changes will have taken place
• Future scenarios that feature abundance and other desirable attributes often convey no credible idea of how major present-day problems were solved on route to these better futures.

However, in the absence of constructive, credible future-oriented fiction, people will inevitably be influenced by fiction that is destructive or incredible.

Accordingly, this section of the Vital Syllabus seeks to collect information about constructive future-oriented fiction.

Strengths and weakness of Hopepunk

See https://en.wikipedia.org/wiki/Hopepunk