I was standing just twenty feet away when Frankie Dettori launched into his flying dismount after the seventh of his seven winners at Ascot on the 28th September 1996. I can’t say I was in a hugely celebratory mood having backed seven losers but I felt a lot happier than some. Moreover I’d learnt an important lesson that day: don’t go racing in cheap new shoes. It transpires girlfriends are not impressed by anguished hobbling.
Subsequent analyses have suggested there may be other lessons to learn; the behaviour of the on-course bookies before the seventh race has received particular attention. The wisdom of their decision to aggressively lay Dettori’s mount, Fujiyama Crest, many of them far exceeding the liabilities they would usually be prepared to countenance on one horse, is still a matter of debate. The bookies stand by their actions, as one rails bookmaker put it “If a bookmaker hasn’t lost a fortune this afternoon, then they just don’t lay horses properly”; while others, not so close to the action, have characterised the decision as illogical and reckless.
I can certainly accept that some bookies were reckless; illogical, however, seems unfair. This was a once in a lifetime opportunity. In their wildest dreams gamblers would be hard put to imagine a 1/12 shot drifting to 1/2 for wholly transparent reasons that had virtually nothing to do with its actual chances, but that was the situation facing on-course bookmakers on Dettori Day. To them it would have been clear that the money coming for Fujiyama Crest was not clever money but a reflection of the enormous liabilities the off-course bookies had on Dettori’s mount in the last.
The ‘Frankie through the card’ multiples placed in the betting shops that morning, as they were every morning, threatened to do Ladbrokes and the like so much damage that their on-course representatives poured money on to Fujiyama Crest in an attempt both to claw back some money should the horse win and to make the starting price (SP) as short as possible. Frankie’s mounts in the fifth and sixth races had returned at considerably shorter than expected SPs for the same reason and there was some small consolation for the off-course bookies when Fujiyama Crest won – an SP accumulator on Frankie’s seven mounts paid barely a tenth as much as an accumulator placed at the early prices. They still lost an estimated £40 million but this certainly softened the blow.
So, while a shortening price would normally suggest the initial 12/1 was too big, that clearly wasn’t the case here. There was no reason to believe Fujiyama Crest had become a better horse in the course of the afternoon or that its rivals had become worse, all that had changed was that its jockey had ridden six straight winners. The only question for the bookies was how much did that affect the horse’s chances?
I doubt there’s a bookie or a professional gambler in the world who believes a top jockey being ‘on form’ going into a race is worth more than a pittance. Certainly there is a small adjustment to be made if a mediocre jockey is replaced by a champion but the 12/1 was available when Dettori was already due to ride Fujiyama Crest. Whichever way one looks at it there is no way Dettori’s winning streak could legitimately change the horse’s actual chance of winning from 1 in 13 to 1 in 9, let alone 1 in 3. The only real worry for the bookmakers – and one their actions would suggest they gave little credence to (before the race at least) – was that the race might be in some way fixed, in particular that the other jockeys, wanting to be part of history, might not try as hard as they usually would.
Clearly there was an overwhelming case for laying Fujiyama Crest for value reasons – and indeed none of the quotes I have seen from bookies state they laid the horse for any reason other than value – but that hasn’t stopped repeated suggestions that the gambler’s fallacy played a big part in their behaviour, that they thought Frankie was ‘due’ a loss, that the ‘law of averages’ meant his previous six winners made the seventh less likely. The bookies, according to this interpretation, were just as foolish as the punters backing Fujiyama Crest at 2/1 because Frankie was on a roll, both failing to appreciate race seven was a new event independent and unconnected to the previous six. It is right to stress the independence of events as the key to the bookies’ willingness to lay Frankie’s final mount, Fujiyama Crest, but it is the bookie’s awareness of that independence, how Fujiyama Crest’s chances had not significantly changed, that caused their behaviour not an erroneous assumption to the contrary.
Let’s put ourselves in the position of the bookmaker – how much should we be willing to risk? Offered 1/2 about a true 1/12 shot – the spin of an unbiased roulette wheel being 32 or less (stakes returned if it’s zero), say – there would be a good case for seriously exceeded our customary stake. Arguably – if we apply the Kelly Criterion, a mathematical formula designed to maximise bankroll growth while minimising the chance of going broke – we’d be justified in risking three-quarters of our entire bank roll; not only would we be getting amazing value we’d also have a very high chance of success. A roulette wheel, however, can’t want to be part of history, a roulette wheel’s odds are determined by mathematics not fallible human judgement – we can’t have the same confidence we’re betting a genuine 1/12 shot (by laying a genuine 12/1 shot) on the racecourse.
But these aren’t the only reasons that should curb our staking: those words ‘once in a lifetime opportunity’, that intuitively seem to urge us to stake more aggressively, should actually be urging us towards caution. The essence of value betting is repeatability; it works because in the long run the luck evens out. With the knowledge that we’ll get 1/2 about a 1/12 shot numerous times we could afford to be very bullish as we could rely on repeat bets to dig us out of any trouble caused by bad luck. Without repeatability, though the hole we’re in if we lose is no deeper, the tools we have to extricate ourselves from it are a lot less effective. When we gamble we risk not just the money itself but also the money that money might earn us in the future. It might make sense to put most of your eggs in this one very attractive basket if you don’t think you have much of a long-term edge: this is your big chance to get a big payout. But a bookie, one would think, has a good long-term edge, so why risk an all but assured bright future for a big payout that will come to you anyway with time and hard work?
On the day, however, I doubt these were the sort of thoughts going through the bookies’ heads. Swept up in the extraordinary events no amount of experience could really prepare them for, looking to win back the money lost in the earlier races, wanting to make this bet of a lifetime count, they just got carried away, some losing all sense of proportion. Not many bookies went home without a financial headache that evening; one in particular, Gary Wiltshire, famously laid the horse to lose considerably more than 100% of his assets, let alone his bankroll, and spent years working his way back to liquidity. Barry Dennis lost £23k on the race – £20k to the off-course bookies, just £3k to ordinary race-goers – having never lost more than £5k on a whole meeting before.
To their critics the bookies were gambling, having a punt when they should have been making a book, but this is a rather naive view of how bookies, particularly on-course bookies, operate. Bookmakers gamble all the time, they don’t come to the race meeting looking for a balanced risk-free book on every race, they come with opinions, horses they want to oppose at the expected prices, horses they want to avoid laying. And in the words of Gary Wiltshire “The price of the last one was ridiculous – if you can’t lay that, you shouldn’t be a bookie. It was the best gamble of my life.”
It’s easy to ‘find’ fault in the bookies’ actions in retrospect, knowing what we know, with the emotion taken out and time to consider our options, but perhaps the bookies did made the mistake of ignoring the possibility that the other jockeys might go easy on Frankie. It’s not that one shouldn’t bet if the integrity of the race is in question but one has to factor it in. Just a 25% chance the other jockeys would let Frankie win drastically improves Fujiyama Crest’s chances – even without the possibility he could win an unfixed race his chances rise to 1 in 4.
After the fact there are many who believe Dettori’s victory that day did owe a lot to the other jockeys’ attitudes, they let him run his own race and only Pat Eddery seemed to be really trying his best to beat him on the home straight, but who knows. What probably irks the likes of Gary Wiltshire above all is the knowledge that had it been clear the jockeys were going to try their hardest the bookies still could have laid the 2/1 til the cows came home. Had Frankie been less popular with his fellow jockeys maybe the bookies genuinely could have laid 2/1 about a 12/1 shot but that’s not to say the horse wouldn’t still have won.