Memory Styles and the Limitations of Memory

Recounting Events

In a recent article on the dangers of insider information, I recounted the story of the ultimate punt-gone-wrong, which found me in an illegal betting parlour in Manhattan, watching as Silvano, the carrier of all the money I had in the world (and some that I didn’t), failed to beat Northerly in the 2001 Cox Plate.

It’s one of the most vivid memories I possess: over 17 years later I can still see Silvano getting boxed on the inside, before making a run up the rails in the final metres, before just faltering as the fatigue of his international travels overwhelmed him. I can see Northerly swooping late on the outside, hampering Silvano, pushing him tight to the rails and impeding his run. As the images flash through my brain, they continue to affect me viscerally: the gut still wrenches at the magnitude of the loss and the stupidity of it all.

It was an exercise in self-flagellation, then, when earlier this week I decided to watch a recording of that fateful race. It wasn’t quite as I remembered it, though. Partly that was due to the commentary, courtesy of a Japanese bootleg on YouTube, but the race seemed different, too. For starters, Silvano didn’t get boxed in. Neither did he make any kind of bid for victory.

And whilst Northerly did make a late, wide run, he in no way impeded Silvano: in fact, there were two other horses between them. Silvano’s race was not the heroic and unjust defeat of my memory; it was a limp and lacklustre loss which saw him with no chance of victory with 800 metres still to run.

How could a memory so vivid be so inaccurate?

The problem, it seems, is that memory does not work in the way we perhaps think it does.  Ask most people to explain how memories are created and they will likely describe a process whereby an event is stored in its entirety at the moment it happens, imprinted on something that is likened to a document, microfiche, or film; then kept in a vault, where it lies, ready to be retrieved later.

One of the reasons we get frustrated when we can’t recall information is that we have a sense that it is floating around in our heads somewhere, if only we had the ability to find it.

The reality is that memories – as neuro-scientists with access to ever-more detailed and accurate brain scans have begun to learn – are not created as something happens.  Instead, we first generate a full mental representation of something when we next try to call the event to mind.

The limitation of this process is that our first recollection of an event is likely to be skewed by all the biases and cognitive shortcomings that cloud our everyday thinking.  In my case, as I walked back to my hotel in New York in the early hours of that fateful night in 2001, trying to come to terms with a crippling financial loss, it is clear that I imprinted a version of events that reduced my feelings of stupidity and guilt.

Silvano had been unfortunate, defeated by the race schedule imposed by connections, and felled by an enemy who hadn’t played fairly. And just as Silvano had been a powerless victim of these circumstances, I had too.

We remember events in a biased way

The neat trickery my brain played on itself, distorting the memory to a more palatable form, must have helped my emotional recovery – it allowed me to retain enough self-esteem to carry on – but it highlights a danger that punters face as they make betting decisions.

When assessing what we think might happen in the future, we will naturally recall what happened in the past; it’s all we have to inform our choices.  But if those recollections are as widely inaccurate as mine are of the 2001 Cox Plate, we’re in trouble.

And just as our recollections may be clouded by our cognitive biases, they can be distorted further by our memory style. Neuroscientists have long been aware of a condition called aphantasia, the incapability to create mental images.

However recent work from Rotman Research Institute suggests that, rather than this being a rare trait affecting a minority of people, it is instead evidence of a memory-style spectrum, which finds some people better at recalling facts (semantic memories) and others better at recalling the sights, sounds, smells, tastes and feelings (episodic memory) of an event.

Few people will exist who operate entirely at one end of the spectrum – like those who have aphantasia – but all of us will likely remember events in a way that is biased to our own memory-trait blend.

Memory Bias

Although poorly, it turned out, I felt sure I had a clear mental representation of the events of the Cox Plate that night.  I can still bring to mind the fellow inhabitants of that betting parlour: what they were wearing, what they were smoking, what they shouted at the screens.

But I have limited recollections of the facts of the affair.  I had no memory of the details of my bets (except that the losses were crushing), or even Silvano’s finishing position: my clear mental picture had him finishing second, not fourth.

The race winner, Northerly, alluded my memory, too.  These factual details I’ve had to look up in my betting record and on the internet, suggesting that my memory style likely favours the episodic over the semantic.

Memory traits, then, further complicate the matter for punters.  Not only are our memories skewed by our biases at the point when we form them on first recall, but our brains may be imbalanced in the kinds of details that those memories include.

This creates a problem for any punter who is relying solely on their memory to inform punting choices.  If your assessment of an AFL match is based on what you recall of previous matches, it is almost certain that the information you are basing your decision on will be flawed in some way.

The Solution

The solution – as with many of the cognitive shortcomings we’ve explored – is one requiring humility. Successful punters ought to recognise the limitations of their memories, relying instead on more incontrovertible data sources. Whilst results databases and ratings systems have their limitations – they are only as good as the data which is inputted and the assumptions on which they are based – they at least overcome the shortcomings of our fallible memory.

I’ve pinned the result of the 2001 Cox Plate on to my noticeboard now.  It will serve as a constant reminder that my memory – in all its rich vividness – is as much fantasy as it is fact, and it will hopefully stop me placing any bets before I check out more reliable versions of history.