Is each-way an outsider worth it?

Pascal's pitch is the one almost every punter has heard at the bar. Back a big price each-way and you cannot really lose. If the 12/1 shot wins you collect a fortune, and even if it only plods into third the place money pays you back and then some. It feels like the canny way to chase a big price rather than just gamble on it, a dream bet with a safety net stitched on.

The idea has a long pull because each-way betting was built for exactly this. Split your money in two, half on the horse to win and half on it to place, and the place half is supposed to be the cushion that keeps you in the game when the win does not land. On a short favourite that cushion looks mean, so the reasoning runs that the bigger the price, the bigger the place return, so the more sense an each-way bet makes the longer the horse is. Back the outsider each-way, the story goes, and you get the upside of the longshot with the comfort of a placed return underneath it.

It is the contrarian's bet too. The crowd piles onto favourites, so the value must be hiding at the other end of the book, and each-way is the sensible way to go fishing for it. You are not a mug having a blind punt on a no-hoper, you are a careful player hedging a big-priced fancy two ways at once. That framing is what makes the bet feel clever and safe instead of what it actually is. We took the claim at face value and ran it across tens of thousands of real races to see whether the safety net holds any weight at all.

The appeal lives in the safety-net story, and it is genuinely seductive. An each-way slip on a big price gives you two ways to win and only one obvious way to lose, so it feels like you have hedged your dream rather than chased it. Even when the horse does not win, the place return drops a few quid back into your hand, and that small, frequent payback reads as the system quietly working. You walk away from a losing race with something, which is exactly the feeling a steady leak needs to survive.

Then there is the memory of the one that landed. Everybody who backs outsiders each-way carries a 33/1 shot that romped home, or a 25/1 fancy that snuck into third and paid for the round. Those days are vivid and they stick, while the long, grey line of losing slips that funded them fades out of mind. One big result feels like proof the method pays, because the human brain weighs a single fairytale far more heavily than a hundred quiet defeats.

The contrarian flattery does the rest. Backing the horse nobody fancies feels like spotting value the herd has missed, and dressing it up each-way makes it feel measured rather than reckless. Two losing bets stapled onto one slip are just opaque enough that the maths never quite announces itself, so the bet feels cleverer and safer than a straight punt on the same no-hoper. The truth is the reverse. You are doubling your stake on the worst-value runners on the card, and the part that feels like protection is the part the bookmaker is happiest to sell you. The appeal is real, the arithmetic underneath it is not.

It loses because you are buying the most over-priced part of the book, and then buying it again. A 10/1-plus horse is long because the market has judged it very unlikely to win, and the market is usually close to right, these outsiders win only a tiny share of races. On top of that low chance the bookmaker loads the heaviest margin of all onto big prices. This is the favourite-longshot bias. Punters keep over-backing longshots for the dream of a life-changing payout, so a 10/1 shot is priced shorter than its true chance, and the very biggest prices are wildly so. You are starting from the worst value on the card before the each-way structure even bites.

Then the structure bites twice. Each-way is really two stakes, a win bet and a place bet, and the place half is priced off that same inflated win price at a fixed fraction, usually a fifth of the odds. So the place stake does not escape the longshot margin, it inherits it. On standard one-fifth terms a 10/1 horse pays just 2/1 for placing, and over thousands of runs the placed longshots simply do not come back often enough, at a high enough fraction, to cover the win stakes that lose outright. The win half on its own bleeds around 37 percent because these horses win so rarely, and the place half is far too thin to rescue it.

The third layer is the racing itself. Plenty of these runners fall or are pulled up over jumps, and a horse that does not complete is a dead stake on both halves of the bet, your whole outlay gone, win or place. We count those fallers and pulled-up horses as the losing bets they are, which is exactly why this figure is honest and steep. The drip is relentless rather than dramatic, no single day looks alarming, but a third of your stake gone, every cycle, compounds until the bank is empty.

We did not model coin tosses or cherry-pick a good run, we used real results. The test ran across 27,909 real British races, and for the each-way outsider rule that gives 26,220 qualifying bets, every race that had a 10/1-plus runner to back. The rule was mechanical and textbook. In each qualifying race, back the longest-priced runner each-way, one unit to win and one unit to place, settled at the industry starting price returned when the race went off. No early prices, no shopping around, no exchange commission, just the official SP.

The settling was deliberately honest, which is the whole point of the Lab. Fallers and pulled-up horses were counted as the losing bets they are, because a horse that does not complete loses you both the win stake and, in almost every case, the place stake too. An earlier version of these numbers wrongly dropped the non-finishers, which quietly deleted a pile of losing stakes and flattered every system. We fixed that. Joint-favourites and any tied positions were split so no phantom result could leak in and tilt the figure our way.

Two notes keep the test fair to the bet rather than harsh on it. First, measuring to SP with no commission is generous, a real punter taking real prices on an exchange or after a price drifts against them bleeds faster still. Second, we applied standard place terms across the sample rather than assuming the enhanced, extra-place promotions that only appear on a handful of big-field handicaps each season, because a system has to be judged on the conditions you actually get all year, not on the best day the marketing department offers. With the rule fixed and the settling honest, we let it run and totted up what came back.

The result is blunt. Backing every 10/1-plus runner each-way returned -32.78% to starting price across 26,220 real British races. In plain money, for every £100 of total stake you put through, only about £67 came back and roughly £33 was gone, every cycle. The 95% confidence range runs from -38.0% to -27.3%, so even the kindest reading of the sample is a heavy, durable loss, not a bit of bad luck waiting to even out. This is one of the most expensive single-bet habits on the whole page.

The strike rate tells you why. These outsiders place just 6.9% of the time, fewer than seven races in every hundred, and they win far less often than that. So more than nine slips in ten come back with nothing at all, both halves lost. The place half, the part that is supposed to be your safety net, fires so rarely and pays so little when it does that it cannot begin to cover the win stakes that lose outright. The win half on its own loses around 37 percent, and a place return of barely 2/1 on a one-fifth 10/1 shot, landing under seven percent of the time, never claws that back.

Hold it against the rest of the board and the verdict sharpens. Each-waying the favourite loses about 12.4 percent, and simply backing the favourite to win loses about 12.5 percent. Each-waying an outsider loses 32.8 percent, well over double either of them. The extra price you chase by going to the longest runner on the card does not buy you value, it buys you a deeper hole, exactly as the favourite-longshot bias predicts. The number is steady, it is honest, and it is firmly in the red.

So, is each-way an outsider worth it? No. At -32.78% to SP across 26,220 real British races it is one of the heaviest losers on the entire page, and it should never be treated as a way to make money. It is a slow, two-bets-at-once leak that runs on the very favourite-longshot bias the bookmaker built the price around, and the place half is far too thin to cover a win half that loses around 37 percent. The safety net is the illusion, not the protection. You are doubling your stake on the worst-value runners on the card and handing the house its biggest margin twice over.

The only defensible use is hunting genuinely enhanced place terms in the handful of huge-field handicaps each season, where the firms compete by paying a quarter of the odds across five, six or seven places. That claws back a chunk of the place stake and softens the loss on those specific races. But softening a loss is not a profit. You are still paying the longshot margin on the win half, the terms are short-lived promotions rather than a system, and the figure tightens straight back towards the full -32.78% the moment you drift onto a smaller field or standard place terms. It is loss reduction, never an edge.

The honest takeaway is the one the whole Lab keeps proving. A bet that feels safe is not the same as a bet that pays, and stapling two stakes together does not remove the tax built into the odds, it just lets you pay it twice. The rare 33/1 winner delays the loss, it never reverses it. This is measured to SP with no commission, so real life is worse again, and the bottom line is plain. Each-way an outsider is not a way to beat the bookies.