The Endowment Effect

Second, I decided I needed a model to be able to statistically assess the relative form of players.  I painstakingly cut-and-pasted historic results into a spreadsheet and (without any of the guidance that anyone repeating this feat today would easily find on the internet) built an Elo ratings system.  I had played a lot of chess as a youngster, so Elo seemed a natural choice.

Third, with the help of a statistician work-colleague, I learned about Monte Carlo simulations and how they could be used to convert ratings into percentage chances.  With a lot of difficulty, using less sophisticated versions of Excel than are available today, I cobbled together a way of pricing-up individual matches and knock-out tournaments.

Fourth, I worked out a staking system, based on the Kelly Criterion, which I’d read was, mathematically, the optimum way to maximise profitability when betting.

I’d started this work over the Christmas holidays with the aim of entering the New Year as a New Punter, with my New System ready to go by the time the World Snooker Championships started in April.

With hours to spare, I hit the deadline: the betting bank was primed, bookmaker accounts were set up, by the end of the tournament I fully expected to be well in profit.

By the time Ronnie O’Sullivan lifted the trophy on 7^th May I had lost most of my money.  I had backed him for the win, along with a couple of other players, but the return was paltry when compared with the string of losses in match bets that I’d suffered in each round.

“I was unlucky, I told myself.  I reminded myself that long-term profitability is not about short-term gain and that, statistically-speaking, a string of losses was not significant.” 

My models were the result of hours of painstaking work, and that hard work would pay off.

It didn’t, of course.  The snooker bets continued to lose, and desperate to repair my betting self-esteem, the self-imposed specialisation rule was soon abandoned: I was back to betting on all sorts.  It was a wildly unprofitable year.

Later that year, in an uncharacteristic (for any punter) moment of honesty with my statistically-minded colleague, I confided my crimes.  He offered to look at the models.  He asked a series of simple questions. “How do you know how many points each player should contribute to the pot for any match?”  “Why do you give that much weighting to more recent matches?” “Why have you built a match’s winning margin into your model?”

The answers to his questions were all the same.  “I just did”.  It soon became apparent that a significant step in the process towards profitability had been missed: testing.

He pointed out that, rather than arbitrarily making assumptions within the model – like, for example, regarding what percentage of a player’s Elo points would be up for grabs in each match – it would be better to optimise these assumptions by testing them on past data.  And even when these assumptions had been optimised, he suggested, wouldn’t it be wise to “paper-trade” for a while, or at least place minimal bets, to work out any kinks in the models and processes?

So, over the course of the next few months, I went back to work.  And like all good stories, the ending was a happy one.  The work was fruitful, and my snooker betting became profitable.  It was never better than marginally so, though, and it took a few years to recoup those early losses.  And that wasn’t really the end of the story.

When significant changes to the professional game were brought in for the 2011/12 season – seeing a greatly expanded fixture-list – I probably should have taken a sabbatical from betting on the sport, assessing the impact of the new schedule on results and updating my snooker model to accommodate them.  I didn’t, though.

Given how disciplined my punting was by that stage – well-tested models, applied to a narrow range of events, in only a few sports, with regular tracking of performance pinpointing where improvements were needed – that decision was, by then, out-of-character.

The thing is, that snooker system was my baby: it started the whole journey towards a different (and profitable) approach to punting.  On some level I knew it was the weakest part of what I did, and

“I knew it would be unable to cope with the challenges the sports’ changes would bring, but I was somehow unable to have that honest conversation with myself.” 

After a few tournaments of disastrous snooker betting, though, I eventually had the conversation.  After a decade betting seriously on the sport, my profit and loss hovered just below zero.  I didn’t place another snooker bet for a couple of years.

Elsewhere on this site I’ve written about the fallacy of storytelling, and how we should be sceptical about any narrative, especially one as neatly instructive as the one above.  Nonetheless, it’s time to conclude the fable by highlighting the lesson it provides.

I will never regret the time I spent betting on snooker.  To some, a decade of work for no return would be frustrating, but the skills I learnt during that period underpinned more profitable betting on other sports in later years.  That’s how I justify it to myself, at least.

Most importantly, though, the experience taught me to distance myself from my betting.

In journalism, there is an oft-used saying that writers must learn to “kill their babies”, meaning they must be comfortable shortening their work to fit the demands of the publication, even when they have tirelessly and lovingly slogged over the perfect crafting of the deleted passages.

Would-be successful punters must learn this lesson, too.  Unfortunately, like all humans, we find this difficult, as our brains seem hard-wired to value things we own, or have created, more highly than we should

In 1990, Daniel Kahneman, Jack L. Knetsch, and Richard H. Thaler conducted experiments where participants were given various objects with varying values.  These ranged from tokens to coffee mugs to chocolate bars.  In different scenarios, subjects consistently seemed to place more value on things that were already in their ownership than things that were not.  In follow-up studies, Kahneman and others have consistently demonstrated that humans (and, in one experiment, capuchin monkeys) have this “loss aversion”: they prefer to hold on to something they own rather than gain something in exchange.

Running concurrently to our desire to avoid loss is a tendency to overvalue things we have created ourselves.  In 2011, Michael Norton, Daniel Mochon and Dan Ariely published the results of three studies which demonstrated what they termed the Ikea effect: whether building self-assembly furniture, or folding origami animal figures, subjects were prepared to pay significantly more for items that they had created themselves; with the price they would pay for an item increasing the longer they had spent on its construction.

It’s little surprise, then, that I found it so hard to kill the baby that was my snooker betting model.  An objective outsider would have realised the need to test it thoroughly before using it in action in 2001.  They would also have recognised that, even after improvements, its marginal profitability was a sign that it still wasn’t optimised.  And with tournament schedule changes in the offing in 2011, they would have realised the urgent need for a sabbatical.

I recognised none of these things.  But then this was my baby, I owned it, I’d spent hours of my life making it, and my cognitive hard-wiring meant that I would instinctively fight to avoid the loss of it.

Eventually, though, the profit-and-loss account was hard to ignore, and I let it go.  Around that time, I wrote, in large capital letters, the word “HUMILITY”, and stuck it on the noticeboard above my desk.  It is pinned there to this day.  It’s a constant reminder to question the things I create and the decisions I make.

“They may be mine.  I may have made them.  But it doesn’t make them good.”

Any punter who wants to increase their profitability would be wise to find their own quick way of reminding themselves of the grave betting dangers of the endowment effect.  The other option is 10 unprofitable years where you learn the lesson the hard way.

About The Author – Jack Houghton

As a passionate sports’ fan and punter, Jack has written about sports and betting for over a decade, winning the Martin Wills Award for racing journalism in 2002 and writing Winning on Betfair for Dummies, first published in 2006 and now in its second edition, having sold over 35,000 copies in two languages