Allan MacDonald - Decision Making in Sport

Episode 220: In this episode of Accelerate Podcast, host Nicola Graham is joined by Allan MacDonald — a decision-making specialist who has spent 15 years in elite sport before transitioning into government, and who holds a professional doctorate in judgmental forecasting and decision-making.

Allan's career has taken him across high-performance sport, public sector strategy, and academic research — giving him a rare cross-domain perspective on how decisions actually get made under pressure. His doctoral work focuses on the science of forecasting and human judgement, exploring why even experienced professionals fall short when it matters most.

At the centre of the conversation is a question every high-performance environment faces: how do we make better decisions — not just with better data, but with better thinking? Allan unpacks the gap between having information and knowing how to weigh it, and why expertise alone is no defence against the biases that quietly shape our choices.

The discussion also explores how the principles of judgmental forecasting and calibrated confidence translate beyond sport — into business, strategy, and any environment where decisions are made in the face of uncertainty.

Topics Discussed

• Experience as a limiting factor in decision making
• Common decision-making errors in high-performance team environments
• Probabilistic thinking and numerical confidence scoring
• Calibration and the accuracy of subjective confidence in sport
• Data overload and its effect on decision quality
• Outside-in thinking and base rate anchoring
• Pre-mortem methodology for complex and novel decisions
• AI and large language models as decision-making aids

Key Points

• Experience is frequently assumed to be the primary driver of decision-making quality in elite sport, but the evidence does not support this view. Practitioners making decisions all the time assume that repetition automatically produces improvement, yet research shows that experience can actively impair decision making in many contexts. This occurs because experience introduces selective recall — practitioners tend to remember successful situations and apply the same solutions even when conditions have fundamentally changed. Despite making high-stakes decisions daily — often involving athletes worth tens of millions of pounds — the majority of sport practitioners have received no formal training in decision making, and when asked how they improve this skill, virtually none report engaging with the decision-making literature or deliberate practice.

• Decision making exists along a broad continuum, from simple and complicated situations supported by clear evidence and expert consensus, through to complex and chaotic environments where no established solution exists. For decisions in the simple-to-complicated range — such as concussion protocols — pre-written, evidence-based policies allow practitioners to respond consistently under pressure without needing to deliberate. However, complex and novel decisions, such as adjusting training load across multiple time zones, managing surface changes, or making selection calls in high-pressure tournament stages, require a fundamentally different approach involving diversity of perspectives, probabilistic framing, and structured deliberation rather than reliance on policy or individual experience.

• Observational research conducted in professional sport environments including Premier League football, NFL, and NBA settings has identified consistent patterns of decision-making failure. The most prevalent error is dominance, where one or two individuals lead the discussion and others align with their position regardless of independent knowledge. Closely related is contamination, where a more experienced team member speaks to colleagues before or during a meeting, causing others to suppress contradictory information for fear of appearing uninformed. A third common error is consensus-seeking, where groups are asked whether everyone agrees with a proposed course of action, often resulting in agreement driven by fatigue or social pressure rather than genuine assessment of the evidence.

• The assumption that more data leads to better decisions is directly contradicted by research evidence. Studies using bookmakers assessing horse and dog racing outcomes showed that decision accuracy peaked at five data points and did not improve with the provision of additional information. Critically, confidence continued to rise as more data was introduced even as accuracy plateaued — producing a widening gap between subjective certainty and actual performance. In practice, this means practitioners should agree on the specific data criteria to be examined before any decision-making meeting begins. Only after an initial decision has been reached should the group be asked whether any data exists that would fundamentally disprove the proposed course of action.

• Framing decisions in terms of numerical probabilities rather than binary agreement or disagreement is one of the most consistently effective interventions in decision-making research. Each group member independently assigns a confidence value to each available option before the group convenes, and these values are averaged to produce a collective estimate. This approach resolves apparent stalemates — where equal numbers of people support opposing options — by revealing differences in the strength of conviction. Two members who are 90–95% confident in one option will, when combined with two members who are only 45% confident in the opposing option, produce a clearly differentiated outcome. This probabilistic averaging approach has shown improvements across multiple domains, group sizes, and levels of decision complexity.

• Calibration is the degree to which a practitioner's stated confidence in a judgment matches the actual rate of correct outcomes over a series of decisions. A well-calibrated practitioner who expresses 80% confidence should be correct approximately 80% of the time. Research on approximately 150 sport practitioners shows that calibration is reasonably accurate at lower confidence levels (50–70%) but deteriorates substantially at higher levels. When practitioners report 90–100% confidence, they are typically correct only around 54% of the time — barely above chance. This has direct practical implications: very high confidence should function as a deliberate trigger to construct the strongest possible counterargument, seek an independent second opinion, or reassess the decision using outside evidence.

• Sound decision making requires anchoring initial judgments in base rate data before accounting for case-specific variables. This approach — referred to as outside-in thinking — asks what typically happens in situations of this type, based on the broadest available comparison data. This base rate probability is then adjusted using inside information: athlete age, injury history, training surface, environmental conditions, and other contextual factors specific to the situation. The two-stage process guards against decisions being driven entirely by the most recent or most vivid information available. Failure to establish a base rate anchor is a consistent finding in sport practitioner decision making and produces inconsistency across similar decisions made at different times.

• Pre-mortem analysis is a structured decision-making technique grounded in a psychological mechanism called prospective hindsight. Rather than asking what might go wrong before a decision is made, practitioners place themselves imaginatively in a scenario where a specific negative outcome has already occurred — such as elimination from a tournament or a significant athlete injury — and identify what caused it. Research shows this approach generates between 20 and 40% more specific and granular problem identification compared to forward-looking risk assessment. Pre-mortem analysis is most valuable for complex, novel, and high-consequence decisions where no strong precedent or established evidence base exists. It is most effective when combined with independent input from multiple team members.

• Research examining how practitioners currently use AI and large language models in decision making found that these tools are being used primarily for base rate information retrieval, pre-mortem scenario generation, and devil's advocate functions rather than as autonomous decision-making agents. The principal strengths of AI systems lie in processing large volumes of data to identify patterns and trends, and in rapidly generating scenario analyses across multiple possible futures. Human practitioners retain a significant advantage in interpreting information not captured in the data — contextual factors such as altered training demands arising from squad injuries, environmental changes, or subtle movement quality variations that current technology cannot reliably quantify. Research demonstrates that data is not objective; the selection of which metrics to collect and examine introduces practitioner bias at the point of system design.

• A practical first step for any practitioner seeking to improve decision-making quality is to map the decision environment by categorising decisions according to their complexity and the time typically available. This enables appropriate tools and processes to be matched to different decision types rather than applying a single method universally. Calibration training — structured practice designed to help individuals understand how accurately their subjective confidence corresponds to actual outcomes — is a further evidence-based intervention that can be implemented at both individual and team level. Teams that assess and develop calibration awareness, and that incorporate probabilistic thinking into standard meeting processes, produce more accurate outcomes across a series of high-stakes decisions compared with teams relying on experience and consensus alone.

Sponsors

Gameplan Performance, Gameplan Performance is a rehab Project Management & Data Analytics Platform that improves operational & communication efficiency during rehab. Gameplan provides a centralised tool for MDT's to work collaboratively inside a data rich environment.

Hytro: The world's leading Blood Flow Restriction (BFR) wearable, designed to accelerate recovery and maximise athletic potential using Hytro BFR for Professional Sport.

VALD, makers of the Nordbord, Forceframe, ForeDecks and HumanTrak. VALD Performance systems are built with the high-performance practitioner in mind, translating traditionally lab-based technologies into engaging, quick, easy-to-use tools for daily testing, monitoring and training.