[[Start here]] → [[manage the monkey mind|Mindset]] → Simple Rules --- ![[Simple rules beat human judgement.png]] We live in a highly specialised world and experts are everywhere. Scientists, doctors, lawyers, fund managers, policy makers… the list goes on. The media love them. The market pays highly for them. Students study for years to become them. *But how much value do they really add?* Well when it comes to predictions, the true answer is not so much. The evidence is overwhelming that **simple rules beat expert judgement** across a huge range of domains. In the year 2000, William Grove published a round up of 45 years of studies into decision making in his paper “Clinical Versus Mechanical Prediction”. [^1] The study analysed the predictive decisions of experts versus mechanical rules in domains as varied as heart disease diagnosis, criminal probation success, startup failure rates, expected academic performance and more. The results are startling. In only 6% of more than 130 studies was expert prediction found to be more accurate than rules based approaches. In up to half of all studies, mechanical rules “*substantially outperformed*” clinical experts. ![[Clinical_Versus_Mechanical_Prediction_A_Meta-Analy_pdf.png]] Even more startling, when experts were given the mechanical predictions, they still underperformed. As James Montier exclaimed in his excellent write up of the paper[^2]: > The evidence is clear, models provide a ceiling from which we detract performance, rather than a floor on which we can build performance. We tend to overweight our own opinions to those of the models. In many domains they have begun to use simple, rules-based models to improve decision making. Surgeries now use checklists to ensure sanitation and airline pilots use checklists before take-off - two scenarios where simple rules have proven to radically decrease the probability of disaster. Meanwhile, more algorithmic approaches are gradually being accepted in medical diagnoses. [^3] But the investment industry still sticks to the cult of the expert. When celebrity fund managers are heavily promoted, there has often been disappointment, such as Neil Woodford’s fund empire collapse or Anthony Bolton’s poor ventures into Asia. These are case studies of experts believing their own hype, and thus failing to stick to the simple rules that brought their success.[^4] After all, nobody likes being told what to do. Investors are often highly over-confident people, we want agency. When a “system” triggers an action - such as a buy or sell - it removes any sense of personal autonomy. It makes us kick like mules. But even Benjamin Graham, the father of Value Investing, knew that investment success could be founded on just a handful of criteria - even as few as one as long as you used sensible diversification. > To try to buy groups of stocks that meet some simple criterion for being undervalued-regardless of the industry and with very little attention to the individual company. Imagine – there seems to be practically a foolproof way of getting good results out of common stock investment with a minimum of work. It seems too good to be true. But all I can tell you after 60 years of experience, it seems to stand up under any of the tests I could make up. **Ben Graham**[^5] As individual stock pickers, we’re often time poor amateurs. The more information we gather, the more confident we may become, but the more attached we will also be to our holdings. The accuracy of our predictions won’t increase with more information. We are likely to continue to be outperformed by simple rules. The lesson is clear. Understand [[What works in stocks?|what works]], build a checklist of simple rules based upon them, and implement them through regular routines. You may find out that you outperform 90% of expert managed funds by doing so. --- More notes related to this idea: - [[investing rules are best designed around base rates]] - [[a mistake is not following your rules]] [^1]: [[Grove - Clinical Versus Mechanical Prediction - A Meta-Analysis]] [^2]: [[Montier - Painting by Numbers - an Ode to Quant]] [^3]: Deepmind - [Using AI to predict retinal disease progression](https://www.deepmind.com/blog/using-ai-to-predict-retinal-disease-progression) [^4]: Ed Croft - [What Went Wrong at Woodford - A Forensic Analysis](https://www.youtube.com/watch?v=QCeKKnQX-sI&ab_channel=Stockopedia) [^5]: Milne & Kahn - [Ben Graham - the father of financial analysis](https://www.ivey.uwo.ca/media/3065497/ben-graham-father-financial-analysis.pdf)