Fixed odds betting terminals: the house didn't always win

The government is in the process of conducting a gambling review. As usual with governments, it is very unlikely to lead to any liberalisation but it is unclear which forms of illiberalisation it has in mind.

I found this podcast a useful guide to what's going on from some people who understand the business, including Philip Davies (Conservative MP), Jon Bryan (poker player and gambling writer) and Emmet Kennedy (presenter on The Final Furlong Podcast). They talk about the prospect of the government limiting how much money people can spend on gambling each month, an idea so ludicrous that I see it as a red herring. More likely are advertising bans and heavy restrictions on online gambling.

As I have said before, this all started with the agitation against fixed odds betting terminals (FOBTs). Once the anti-gambling coalition won that battle, it was open season. 

I discovered something interesting about FOBTs recently. The government's rationale for reducing the maximum stake to £2 was that no one had lost more than £1,000 in a session playing at that level in 2015/16 whereas 170,670 sessions at higher stakes had ended with people losing more than £1,000. 

 

Preventing people from losing £1,000 had never been a principle of UK gambling regulation before and the government's own figures showed that losses of this kind occasionally took place on other gambling machines (and happen all the time with retail betting). The reality was that the government was under pressure to reduce the stake to £2 and needed to find a justification for it.

But if you look at the data the government used to make this decision - which was based on 128 million sessions between July 2015 and June 2016 - you'll notice something interesting. Yes, there were 170,760 sessions ending with a loss of over £1,000, but there 209,464 sessions ending with wins of over £1,000. There were 543 sessions ending in losses of over £5,000, but 592 sessions ending in wins of over £5,000.

There were more big payouts than big losses. This holds true at the £500-£1,000 level too. There were 626,897 sessions ending with losses between £500 and £1,000, but 700,063 sessions ended wins of between £500 and £1,000.

At every point below this, the bookies won. There were 3,008,317 sessions ending with losses of between £200 and £500, for example, whereas only 2,628,258 sessions ended with wins of this amount. The most common loss was between 1p and £5. There were 28.4 million of these, but only 6.5 million winnings of up to £5.

 

But when it came to big wins and losses (i.e. over £500), the players won more times than the bookies. None of the people who won more than £5,000 staked less than £20 a spin and most of them staked more than £50. None of the people who won more than £1,000 staked £2 or less and most of them staked at least £40.

It's not immediately obvious what explains this phenomenon, but it isn't a fluke result. You can see exactly the same thing happening in 2014/15. The nature of the machine's fixed odds is that they produce a reliable profit over time (the house margin is around 3%). The only explanation I can come up with is that players exploited the only advantage they have over a machine: they can walk away when they're winning.

Or is there some other reason that I've missed?