Sophisticated Forecasting, Fragile Strategy

Prediction is unavoidable. Every consequential decision requires some judgement about a future that has not yet occurred. Investors form expectations about markets, companies make assumptions about demand and competition, governments anticipate economic and geopolitical conditions, and individuals make decisions about careers, relationships and opportunities whose consequences may unfold over decades. We cannot act without forming some representation of what is likely to happen next. The problem begins when the confidence required to make a decision gradually becomes dependence upon that decision being correct. 

I have increasingly come to think that this distinction reaches beyond the conventional limits of forecasting. Human beings necessarily simplify the world in order to navigate it. Reality contains more information, interactions and possible outcomes than any individual can observe or process, so we construct representations that make it intelligible. Our experience, education, culture, profession, relationships and beliefs contribute to those representations, and the longer they remain consistent with what we observe, the more natural they begin to appear. What began as an interpretation can gradually acquire the psychological force of reality itself. 

This creates a particular difficulty when we attempt to predict the future. Conditions that have persisted for years can begin to look permanent. A successful professional path appears likely to remain valuable, a profitable product appears capable of continuing its trajectory, an established international relationship seems durable, and a market pattern that has repeatedly worked can begin to appear almost structural. The familiarity of the pattern encourages confidence in its continuation. Yet the apparent stability belongs partly to the map through which we understand the world rather than to the world itself. 

Prediction therefore contains a psychological danger in addition to an analytical one. We may become attached not only to what we believe will happen, but to the larger representation of reality that makes that prediction appear sensible. Once a forecast becomes connected with capital, reputation, identity, time and previous decisions, being wrong becomes increasingly costly. At that point, we no longer merely prefer the forecast to prove correct. We may begin to need it to be correct. 

When strategic concentration becomes psychological concentration 

Financial markets make this mechanism unusually visible. A position initially begins as a judgement. The investor observes the available evidence, develops a thesis and commits capital because one outcome appears more probable than the alternatives. Nothing about that process is inherently problematic. Decisions require conviction. The difficulty arises when commitment grows to such an extent that the investor's ability to reassess the thesis becomes compromised by the consequences of abandoning it. 

A losing position can then acquire significance far beyond its financial value. Closing it may require accepting that the original judgement was wrong. It may crystallise a substantial loss, damage a record, contradict a view that has been defended publicly or undermine confidence in an analytical process in which considerable effort has already been invested. What began as financial exposure becomes psychological exposure. The investor is no longer evaluating the position from the same distance that existed when it was first established. 

This helps explain the familiar tendency to double down when reality begins to move in the opposite direction. Increasing commitment can sometimes be entirely rational if new evidence genuinely strengthens the original thesis. But the same behaviour can emerge for a very different reason: the decision-maker increasingly needs the original thesis to recover because too much has already become dependent upon it. The distinction between perseverance and attachment becomes difficult precisely because both can produce persistence. One continues because the evidence remains strong; the other continues because recognising the evidence has become too costly. 

The same mechanism operates outside financial markets. A professional who has devoted twenty years to one career does not confront technological disruption as an abstract change in labour-market conditions. The disruption reaches accumulated expertise, status, relationships and identity. An entrepreneur who has invested a decade building a company may experience deterioration in the business model as something much larger than a strategic problem. A country whose economic or geopolitical position has developed around a particular relationship over generations may find it difficult to recognise that the external environment has fundamentally changed. In each case, commitment has accumulated across several dimensions simultaneously. 

Strategic concentration can therefore become psychological concentration. The more capital, time, reputation, identity and effort become attached to one interpretation of the future, the greater the internal pressure for that interpretation to remain valid. This can reduce precisely the capacity that uncertainty requires most: the ability to observe new evidence and change direction. 

Probability does not become certainty 

The underlying problem is not conviction itself. Serious decisions cannot be made from permanent hesitation. Some outcomes really are more plausible than others, and better analysis should influence how we act. The mistake lies in allowing relative probability to acquire the practical status of certainty. 

Suppose, simply as an illustration, that several future scenarios can be identified and that one appears substantially more likely than each of the others. It may be entirely rational to position for that outcome. Yet its relative superiority tells us only that it is our best judgement among the available alternatives. It does not remove the possibility that another path will materialise. Even if probabilities could be estimated with unusual confidence, which itself is difficult in complex systems, the most probable scenario could still fail. 

There is also a deeper limitation. Some future possibilities can be identified even if their probability remains uncertain. Others may be conceivable but appear remote. Beyond both categories lies the larger problem of developments that were never included in the original map at all. A strategy constructed exclusively around the expected scenario therefore depends upon two demanding assumptions: that we have estimated the probabilities correctly and that we have identified the relevant possibilities in the first place. 

This is why increasingly sophisticated forecasting cannot, by itself, solve the strategic problem. Better analysis can improve judgement and should always be pursued. It can help distinguish stronger from weaker scenarios and improve the allocation of resources. But no refinement of the forecast changes the underlying condition under which the decision is made: action precedes knowledge of the outcome. 

A forecast answers the question of what we currently think is most likely to happen. A strategy must address an additional question: what will we do if something else happens? 

Conviction in the system 

This distinction changes where conviction should ultimately reside. The strongest conviction should not be placed in the infallibility of a particular forecast. It should be placed in the quality of the system through which forecasts are translated into action. 

A robust system can still act decisively. It can commit resources to the scenario that appears most attractive and benefit substantially when the judgement proves correct. Its distinguishing characteristic is that other outcomes remain tolerable. It retains enough capacity, alternatives and room for manoeuvre to respond if the original expectation fails. The system does not require every important decision to be correct in order to remain capable of functioning. 

This is an important extension of what I mean by Strategic Depth. Much of risk management naturally focuses on the capacity to absorb adverse outcomes. Yet that capacity also changes the psychology of decision-making. If being wrong can be absorbed, recognising that one is wrong becomes less threatening. If a decision does not place the entire structure at risk, changing the decision becomes easier. If identity has not been completely concentrated in one role or judgement, revising that judgement does not require dismantling the entire conception of oneself. 

I have experienced a version of this effect in my own professional life. During my earliest media interviews, the possibility of making a serious mistake carried considerable weight. I prepared intensely because I had relatively little accumulated evidence that I could perform well in that environment. A badly expressed argument, an incorrect figure or an interview in which I simply failed to communicate clearly felt consequential because the track record was still small. 

Hundreds of interviews later, the standard I expect from myself has not fallen. I still notice when an explanation could have been clearer, when I have expressed something imperfectly or when an intervention has not reached the quality I wanted. What has changed is the architecture around the individual error. A much larger body of experience, knowledge and previous performance now surrounds any single interview. One imperfect appearance does not invalidate everything that preceded it. 

The accumulated reserve creates room for imperfection without making excellence irrelevant. More importantly, it changes the relationship with error. A mistake can be examined as a mistake rather than experienced as an existential verdict on the entire professional identity. That distinction creates greater freedom to learn from it. 

Freedom from the need to be right 

This may be one of the less obvious benefits of Strategic Depth. Reserves, alternatives and diversified sources of capability are normally understood in terms of resilience. They allow a system to continue functioning when something goes wrong. But they may also preserve something more subtle: intellectual freedom. 

A person who cannot afford to be wrong has strong incentives to defend the original judgement. New evidence becomes threatening because accepting it may carry consequences far beyond changing an opinion. By contrast, a person or institution that has retained sufficient capacity can examine contradictory information with greater independence. The cost of updating the map is lower because the entire system does not collapse when one part of the map proves inaccurate. 

This does not imply indifference to prediction. Nor does it imply weak conviction. The objective is almost the opposite. Conviction becomes more useful when it can remain proportional to what is actually known. One can develop a strong thesis, act upon it and still retain the capacity to say that subsequent evidence has changed the judgement. The strength lies in being able to commit without becoming captive to the commitment. 

The distinction is especially important over long horizons because occasional mistakes are not the relevant test. Any sufficiently long career, investment process, business history or life will contain periods in which several things go wrong. Forecasts can fail consecutively. Decisions that were reasonable given the available information can produce poor outcomes. External circumstances can change faster than expected. A system designed merely to survive one isolated error may still prove fragile when adversity arrives in clusters. 

Strategic Depth should therefore be evaluated partly by the number and variety of errors a system can absorb while preserving meaningful capacity to continue. That requires asking a more demanding question than whether we can survive being wrong once. What happens if several important judgements fail within the same period? Can we still think clearly, revise our assumptions and act, or have previous commitments left us with no viable alternative except hoping that reality eventually reverses in our favour? 

The difference between strategy and gambling becomes increasingly clear at that point. Gambling can contain sophisticated forecasts, elaborate models and enormous conviction. What makes it fragile is the dependence of the outcome upon being right. Strategy accepts that judgement is necessary while refusing to make continued freedom contingent upon its infallibility. 

Prediction remains indispensable. We should continue trying to understand what is likely to happen, improving our analytical methods and forming convictions when the evidence justifies them. But prediction performs only one part of the strategic task. The other part is constructing enough depth around our judgements that we remain capable of learning when they fail. 

The freedom to change one's mind is therefore not simply an intellectual virtue. Under uncertainty, it can become a strategic asset. When the cost of error is survivable, evidence can remain evidence rather than becoming a threat to identity. Conviction can coexist with revision, commitment with optionality and ambition with humility about what can actually be known. 

A strategy worthy of the name should allow us to pursue what we believe is most likely without requiring reality to confirm that belief. Its deepest strength may lie precisely there: in allowing us to hold strong convictions without becoming captive to the need to be right.