A few months ago I talked a bit about the main findings of my MSc Psychology dissertation. I got a good grade for my research, but it wasn’t a top grade, mostly because I tried to do too much. That bothered me at the time but in hindsight it was a fair criticism, because now I’ve happily rustled up another 1200 words on the subject. Sorry!
In the first instalment of academic self-aggrandisement, I explained how we bet efficiently but stupidly by using mental shortcuts. Today I’m going to talk about two secondary themes, which I think are actually more interesting.
The Near Miss Effect
If you’ve ever played a lottery scratch card, then you’ve experienced the Near Miss Effect. Typically, there is a game where you have to match three pictures or prize amounts to win, and the first two uncovered are jackpots. You take a deep breath and slowly scuff away the last bit of silver… and lose.
Similarly, if you’ve spent much time playing slots (or fruit machines) you will have experienced countless spins where the first couple of reels line up for a big win, then the third stops just short or long of where you want it. So near yet so far!
It happens routinely for normal wins but is even more noticeable when it comes to starting a bonus feature. More advanced modern machines slow down the critical moment and play suspenseful sound effects to build the tension… before shattering your dreams.
In both cases, the anticipation of a big win causes a physiological rush of excitement. Although it is disappointing to subsequently lose, the subconscious thrill of the almost-win provides a big incentive to continue playing. Research suggests the compulsion can be just as strong as the effect of an actual win.
I decided to test this effect in my study by using a spinner to determine wins or losses. You can see an example of the wheels I used below. The layout was chosen deliberately so that I could measure how close a participant came to winning or losing. The speed of the spinner was set to run long and slow, precisely to build tension.
Just observing the players, it was obvious that close proximity to the win line caused much more animation/agitation. When there was uncertainty until the last moment, there was invariably a physical response. Conversely, when it was clear that the spinner would land comfortably in a win or loss segment, the participant was usually serene and unflustered.
Unfortunately, capturing data on the physical response of my volunteers was not an option. As much as I would have loved to hook them up to a heart monitor, or have them play in an fMRI machine, budget and logistics did not allow it.
Instead, I used the self-reported enjoyment of the game to determine whether exposure to the Near Miss Effect made a difference. That’s a fairly crude measure, but it worked. The most reliable predictor of participant enjoyment (more than winning or losing, or general progress in the game) was whether they experienced a near miss.
It was this finding that has led me to consistently ponder the role of suspense in our enjoyment. I believe that having a bet that makes you sweat is the most enjoyable element of gambling – when it is a fair and natural part of the game.
The reason I started this segment with examples from scratch cards and slots, is that they are usually engineered to cause the effect. The operators for these games have learned about near misses and are rigging their games to make you feel a false high even when you lose, because they know it improves your chances or trying again. To me, that is despicable.
Heuristics and Favouritism
Firstly, a short recap of the experiment. I asked participants to bet £3 on one of the five warriors below to win a fictional tournament. You get to see the appearance of the characters, their prize money, and their probability of winning. Who would you go for?
The main idea was to see if the players would make an easy choice based on the pictures and prizes, or a harder one based on the maths. Humans are lazy and the easy option won in a landslide (88%). In a separate article, I further discussed how the bettors didn’t really care about losing the £3, because the amount was small and I put up the money. No loss aversion found.
But there was more. If you’re in the 12% that took the time to study the five characters, you noticed that there were two pairs that had the same prize money and probability – the only difference between them was the meaningless picture. This was a sneaky test of subconscious bias.
One pair was set up to test for gender preference in a betting situation. The second and fifth images above were chosen to look similarly capable of winning, so that gender would be the only determining factor. If you had settled on an £18 prize or the ~13% chance of winning, then it came down to a simple preference between male or female.
Academic research suggests that people can be strongly biased on gender, with both sexes showing varying degree of favouritism depending on the context of the situation. Under these conditions, the female cohort had a slight preference for the male character, although statistically speaking it was not significant. In short, they were balanced in their choice.
The male cohort were not. Of the eighteen males that picked from the gender bias pair, every single one chose the male warrior. If you had a coin that only landed on heads for eighteen flips in a row, you would be rightly suspicious.
I would like to see if this finding holds up in video gaming. How often do male players pick Peach in Mario Kart? According to a study last year she was the joint most popular character alongside Mario (11%), but no gender breakdown was provided. There must be masses of data on this, but I couldn’t dig up anything that addressed the subject.
The other matched pairing was a test of the attractiveness bias. The science shows we have a general preference towards more attractive people. For example, better looking actors tend to be paid more for their parts and hotter hospitality staff usually to get better tips. With that in mind picture 1 was picked to be hansom and picture 4 was chosen for being a minger.
This test didn’t yield such a clear-cut result. There was a small overall preference for the better-looking character, but not one that was significant. Perhaps attractiveness isn’t a big consideration when you’re betting on warriors in a tournament, or maybe the ugly character was just too ugly. There was a notable outpouring of pity for him among his selectors.
My results certainly don’t disprove the existence of this bias. The book and movie of true story Moneyball is based on the idea that people overrate the attractive option. The Oakland A’s built a championship winning team out of supposed misfits who were markedly undervalued based on the metrics that mattered.
It’s perfectly feasible that we do the same with our betting, but I couldn’t prove it here. If you had access to the mountains of data the bookies have, you could easily put it to the test. And if it is the case, it follows that the bookies would adjust their prices, or design their games, accordingly.
If people are backing pretty-boy Jack Grealish to score goals partly based on his appearance (even subconsciously) then you may as well reduce the price a little and boost your profit margins. As a former gambling industry scumbag, that’s the sort of speculative edge I’d be investigating.
This week I’ve just been wasting free promotional fivers on football matches, but I will be betting on the Ryder Cup at the weekend. I don’t really need to, because it’s so engaging, but I can’t help but try to pick off loose lines (even though I’m a rank amateur in matchplay golf betting).