Recently I have finished reading Daniel Kahneman’s bestseller “Thinking, Fast and Slow” – a book dealing with our common biases and thought flaws that we all as humans are guilty of. Overconfidence, risk-aversion, manipulative framing – you will find it all here. You have certainly come across the title already if you frequent some of the other betting blogs around the web, since it is often being recommended as a must-read for any aspiring sharp punter or trader. The author probably didn’t write the book with sports betting in mind – the book (and Kahneman’s life’s work in general) is rather intended as a challenge to the traditional efficient market hypothesis used as a basis for most classical economic theory. The author himself is a renowned scholar in the field of behaviorial economics and won The Nobel Prize for his contribution to the field.
What’s in it for punters?
The clash between the efficient market hypothesis and alternative economic theories is central to sports betting. The efficiency of the sports betting markets has been a subject of a lot of research. If you are new to the topic I recommend getting started with the Wisdom of the Crowd article of Joseph Buchdahl from football-data.co.uk, which explains what exactly does market efficiency mean and why any edge you might find in the sports betting market is bound to diminish and disappear over time.
However, “Thinking, Fast and Slow” offers an alternative view to the markets – one where we are all blinded by the same biases and do the very same mistakes. As a result the markets we participate in are less than efficient, since we do not always act as the rational agents economic theory expects us to be (referred to by Kahneman as “econs” as opposed to “humans”).
Kahneman explains the main ways in which, according to him and his collaborators (most notably Amos Tversky, his life-long research partner), we act differently than classical economic theory would expect us to act. These are all well-researched and also pretty fascinating, since many of us would find themselves falling victim to the very same biases while reading the book. “Thinking, Fast and Slow” offers a number of lessons very relevant to the field of betting. In general it is a book that can help you improve your decision making process in any area of life quite a bit, but I will focus on its implications for betting.
I have found it hard to summarize the main findings of the book within just one article, so in order not to bore you with a wall of text I will cluster the findings and release them into a series of articles. I will start with 5 rules related to Overconfidence (Part 1), plus some advice on how to mitigate its negative effects. In the follow-up articles I will cover the topics of Extreme Events (Part 2), Intolerance to Risk and Uncertainty (Part 3) and Good Policies and Habits (Part 4).
1. Ignoring the base rate – the average punter is a loser
Behavioural psychologists have found out that people often tend to ignore the base rate when trying to answer a statistical question. The author illustrates the point by asking us whether a shy and withdrawn guy with little interest in people, a need for order and structure and a passion for detail is more likely to be a librarian or to be a farmer. If you are like most people you would have picked librarian and you would be wrong. While the description suits the stereotypical librarian better, most people tend to forget that there are much more farmers in the world than there are librarians and some of the farmers do match this description.
While focusing on provided information we tend to ignore the big picture, which turns out to be much more relevant for the question at hand. There is an important lesson for every punter here. Sometimes we tend to follow a certain promising strategy in our betting activity – be it new software, a tipster or a pattern we have identified in the historical information. We focus on what we have – the (possibly skewed) historical stats presented by the software vendor/tipster or the positive historical returns from our cherry-picked strategy – but we tend to forget that on average the punter is losing. With that in mind we should always be very critical to every betting strategy and think twice before putting money into it. The base rate is against us and we should never forget that.
2. Bayesian statistics – quantify the base rate
That being said, the base rate of losing punters is by far not the only way in which you can integrate base rates in your betting activity. Long term trends such as home advantage are base rates in themselves too. More interestingly, research has continuously shown that people tend to ignore those too and focus more directly on the question at hand (which team will win game X). Backing the home team may or may not deliver some edge, but in general, if you wish to quantify base rates in your models you might want to have a look at Bayesian statistics – it will help you accurately reflect base rates in your models and get an edge in front of the bookie.
3. Rejecting the null hypothesis – always assume that your hypothesis is wrong
This is certainly not Kahneman’s discovery but merely a best practice in the field of science that he is sharing with the readers. It is a practice often forgotten in the world of betting where someone would try to sell us what is claimed to be a profitable system or where we try to full ourselves that we are onto a profitable trend while in fact we are not. We will often try to believe and confirm hypotheses, which we wish were true.
There is a similar problem in the field of science where researchers would try to come up with a shocking discovery in order to get a publication, even if the data at hand does not confirm the findings with any satisfactory level of certainty. Therefore the adopted approach in science is to assume your hypothesis is untrue – the so called null hypothesis. Only when you manage to reject that null hypothesis with sufficient certainty is it reasonable to conclude that you might be up to something. It is highly recommended to follow this approach when testing tipsters and betting strategies. For the technical details you can refer to my article about the t-test to get familiar with the basics.
4. The illusion of pundits – why we put too much trust on “experts”
Experts are often wrong. What’s worse, research has found out that in many areas of life experts are wrong in just about half of their predictions, which makes them as good as a random number generator. CEOs, economists, political experts and, more often than not, betting tipsters, will create around themselves the aura of competence, only to later prove themselves completely incapable of making a forecast worth a dime. The book quotes a study showing that forecasts of political experts attempting to foresee the outcomes of upcoming events are basically a random guess or even a little worse than that. There are also numerous studies showing that the majority of fund managers underperform a broadly diversified low-cost stock index. Finally, in the world of betting it has been proven that the majority of tipsters cannot provide you a steady positive return, meaning their tips are worth nothing.
It appears that in certain areas of prediction every outcome is influenced by way too many factors for an expert to foresee. Kahneman tells the story of a colleague of him with whom they had a heated argument for years regarding the validity of experts’ predictions. Kahneman’s colleague opposed his theory of expert uselessness with the example of leaders of firefighting commands who would normally anticipate a roof falling down in a burning building within seconds, their quick reaction saving many lives throughout the years.
Both researchers joined efforts in collective research to find out why some expert forecasts work and others do not. An agreement was reached that the areas where expert intuition has any merits are those in which the experts have repeatedly faced the same situation, which has helped them develop an inner feeling about it better than most people (think driving a car or similar activities).
Back to betting, and thinking about sports games, an outcome is influenced by a virtually infinite number of factors, many of them unknown to the outside viewer. No two sports games are exactly the same. While you would know that after turning the wheel left the car turns left too without even thinking of it, it is hard to develop the same intuition about football games.
Real experts in the field of sports are hard to come by. As much as we like listening to an entertaining commentator of a football game, he is indeed nothing more than entertainer and can hardly help you predict the result of a game. Same goes for tipsters or any self-proclaimed experts in the field of betting. If someone claims to possess a skill that is next to impossible to master, it would be wise to reject his claim per default until proven wrong.
5. Ignoring the reliability of evidence
Speaking of tipster records – it has been experimentally proven that we humans tend to systematically ignore the reliability of the source of evidence and focus on the presented evidence instead. Funnily enough, we continue doing so even when we receive explicit confirmation that the source of the evidence is unreliable and not to be trusted.
Challenging the source of our information just does not seem to be our thing. If you cannot force yourself to do that, try to adopt a sort of policy to be sceptical about anything you read on the Internet (which is a generally sound advice, not only as it comes to betting) – especially if someone tries to sell you something. If you wish to have reliable records, try to gather them yourself.
That’s all on the topic of Overconfidence as covered by “Thinking, Fast and Slow”. Stay tuned for the next parts of the review dealing with the topics of risk aversion, extreme results and good punter habits. See you soon!