Challenging Our Punting Assumptions

When it comes to punting, I pride myself on being rational.  I’m quite smug about it, if I’m honest. Being asked by a friend to pick up a winning bet from a betting shop earlier this year, then, was a self-congratulatory experience: surrounded by such irrationality, it was hard not to be conceited.

But it was depressing, too.  Folk shuffling back-and-forth between the counter and their chairs: a steady stream of money-they-couldn’t-afford being handed to the Customer Services Manager.


What's Worse?

Most depressing, though, were the conversations of my fellow inmates, every bet preceded with long explanations – to anyone who was listening – as to the wisdom of the investment.  These explanations would have been laughable if they weren’t delivered so earnestly.

“The favourites are winning today – watch this one go in.”

“It’s about time trap six had some luck.”

“[Insert name of jockey] wouldn’t be riding this if it wasn’t a certainty.”

“The odds on this one are tumbling, all the big money’s on it.”

“[Insert name of trainer] always does well in maidens.”

They were a litany of the worst, and most unprofitable, kinds of punting absurdities.

It was embarrassingly humbling, then, when analysing the men’s French Open tennis tournament that same day, that my own irrationality became evident.


Challenging Our Beliefs

For several years, I’ve maintained Elo ratings, allowing profitable trade on the men’s game.  Unlike many punters who use similar rating systems, though, I’d invested a lot of time when first creating mine to ensure that I could produce different ratings for the same player that adjusted themselves according to playing surface.  It was this, I assumed, that gave me the edge on the market: when assessing matches at Wimbledon, for example, I was able to focus my attention on a player’s previous performances on grass, providing a more precise understanding of their likely performance than when using a rating that including the less-relevant outcomes of clay-court matches.

The edge I had on the market had been steadily eroding over the last few seasons, though.  I was still profitable – comfortably so – but each season the plus-figure was smaller.  The phenomenon of decreasing margins is nothing new.  As markets become ever-more competitive, with the number of savvier punters increasing, making a profit takes a bit more work.  That was my explanation for my waning tennis-punting fortunes.

On the day I visited the betting shop, though, my irrationality-radar no doubt primed, I found myself struggling to answer a question that had come to mind.  If the playing surface is so important, why have the same small group of players dominated the men’s game – on all surfaces – for so many years?

It’s not like I hadn’t considered the question before.  It’s just that the answer I’d always given myself – the likes of Federer, Nadal and Djokovic have a versatility that allows them to excel on different surfaces – no longer felt satisfactory.  The search for a better answer to that question soon led to the death of my surface-specific Elo ratings.

The first step was to steal some analysis of a fellow tennis punter.  She had painstakingly analysed the rally length at the four majors over time and found that there had been a rapid convergence.  Whereas in the 1990s a rally at Wimbledon lasted, on average, around two shots, the French Open provided much better viewer-value: with average rallies lasting more than six shots.


Applying This To Today

Today, however, there is little difference between the four tournaments.  Things are a little punchier at Wimbledon, with rallies lasting an average of around four shots compared to the five-and-a-half of the other tournaments, but the differences are not especially significant.

There is a debate to be had about what has caused this convergence – did the players mentioned above cause this change in the game, or have they simply profited from it – but, whatever the outcome of that debate, what became clear is that surface-specific Elo ratings seemed to be an increasing irrelevance.

To triangulate this conclusion…

I had a look at the performance of my own system, comparing the performance of my raw, all-surface Elo rating, to the surface-specific Elo rating that I had thought provided my competitive edge.

Using an historic eight-year data-set of match results, I compared what my ratings said should happen in matches, to what did happen.  For example, I took all the instances where the ratings predicted a player had a 55% chance of winning, and calculated whether they went on to win 55% of the time.

The result?  The surface-specific Elo rating performed adequately, explaining why I’d been able to stay profitable over the years.  Slightly depressingly, though, the all-surface Elo rating performed better.  It seems that, had I worked less hard, taking a simpler approach, I would have been more profitable.

Beating myself up about that lost profit is not going to be productive.  What’s more important is drawing a lesson from it to ensure that something similar isn’t happening elsewhere in my punting, and won’t happen in the future.

At the heart of the error I made was an assumption based on a long-held belief.  A teenager in the 1980s and 1990s, I grew up with a tennis world that saw uber-specialists concentrating on specific surfaces.  It was perhaps unsurprising that when first creating my Elo ratings, then, I assumed I had to have a way of assessing a surface bias a player may have.  As it turned out, though, it was a moment of irrationality just as flawed as the thinking on offer in the betting shop that day.

There is a wealth of evidence that our childhood experiences shape our adult thinking.  For example, experiencing trauma when young puts us at greater risk of suffering a plethora of mental-health conditions when older.  We also know that other, less-dramatic experiences, determine our later outlook: if we have fewer contacts with outside groups when young, for example, we are much more likely to be xenophobic grownups.

I’m unaware of a study been conducted into the role of observing certain sporting trends as youths and how this might govern our later punting beliefs (I’m guessing it’s hard to get research grants for that kind of invaluable research), but it’s a fair to assume they do, as the downfall of my surface-specific Elo ratings has demonstrated.


Lessons To Learn

There are three lessons to take from this sorry episode.

First, being a profitable punter requires us to constantly assess, and reassess, the assumptions upon which we base our bets.

Second, we need to find quantitative ways to test those assumptions, accepting that sporting conditions change over time, and that these changes may require us to alter our approach.

And third, that a rigorous way of recording our bets, which tracks profitability, provides us with an invaluable early-warning system that something might be wrong in our thinking.

The trick is not to ignore the noise when the alarm bells start ringing.  I did for too long, and it cost me dearly.


Related Articles

Are you a better than an average punter?

You have agreed to take part in a questionnaire.  Please answer the following questions.

Illusion of Knowledge

Jack Houghton discusses the Illusion of Knowledge.

Framing Effect Bias

Published author, Jack Houghton, is our resident expert on Betting Psychology and is on hand to discuss the framing ...