Does backing the second favourite work?

Of all the systems in the Lab, this is the one that sounds the most reasonable when a punter explains it to you in the pub. The favourite, the story goes, gets too much money. The whole crowd piles onto the shortest price, the bookmaker shortens it further to soak up the bets, and by the off the favourite is poor value, too short for what it really is. So you step one place down the market. You take the second favourite. Nearly the same horse, the reasoning runs, almost as likely to win, but at a fairer price the herd has not trampled.

It feels like the canny middle ground. You are not chasing a 33/1 no-hoper and you are not blindly following the over-bet jolly. You are backing a serious contender, a horse with a real chance, at odds that look a little more generous than the favourite's. You imagine yourself quietly collecting the value everyone else left on the table while they hammered the front of the book.

The second favourite obliges often enough to keep the story alive. It runs into the frame, it places, it is rarely far away, so the bet always feels in the hunt rather than dead on its feet. That sense of being alive in the race is exactly what makes it so easy to believe in.

So we tested it the only honest way: back the second favourite in every single race, flat stakes, settle every bet at the official Starting Price, and count tens of thousands of races. No cherry-picking the meetings that suited, no quietly dropping the days it went wrong. Just the blunt question a punter actually wants answered. If you did this, race after race, all season, would you come out ahead?

Pascal, the Lab's eternally hopeful punter, puts the appeal better than anyone. "Everyone piles onto the favourite," he says, "so its price gets too short to be worth it. The second favourite is nearly as likely to win but pays a bit more, so I'm getting the same kind of horse at a fairer price. Smart, no?"

It is a tidy bit of logic, and that is precisely why it traps people. It borrows a real fact, that favourites are slightly under-priced because the crowd over-backs them, and stretches it into a conclusion that does not follow. Yes, the favourite is shaded a fraction short. But the second favourite is not the leftover value from that. It is simply the next horse along, priced with the very same margin built in.

The other half of the appeal is how the bet feels in real time. The second favourite is a proper contender, not a rag. It wins its share, it places more often than it wins, and it is usually thereabouts at the business end of the race. You are rarely watching your selection toil at the back from the first furlong. That steady drumbeat of near-misses and frame finishes reads as a strategy that is working, a horse that is always knocking on the door.

There is comfort in it too. Backing the second favourite lets you feel cleverer than the mug who blindly follows the jolly, while still siding with a horse the market respects. You get to be contrarian and sensible at once. The trouble is that feeling sensible and being paid are two completely different things, and the second favourite quietly proves the gap between them on every single bet.

The second favourite is not a mispriced horse. It is just the runner the market rates second most likely to win, and the bookmaker bakes the same built-in margin into its price as into every other horse on the card. That margin is the overround. Add up the chances implied by every price in a race and they total well over 100%, around 112% on a typical book. That extra slice is the house edge, and the second favourite pays it in full, exactly like the favourite, exactly like the outsider.

It also sits in a poor spot on the favourite-longshot bias curve. The bias is the long-standing pattern where short prices are slightly under-bet and big prices are heavily over-bet. The favourite, right at the front, is where the bias is least unkind to you. The wild outsiders are over-bet but at least carry the dream. The second favourite gets neither. It is a step longer than the favourite, so it loses a little of that front-of-market cushion, yet it is nowhere near long enough to be one of the over-bet rags with a quirk to exploit. It carries no offsetting edge of any kind.

So here is the real mechanism. The second favourite wins under one race in five, far less often than the favourite. The few extra points of price you take never come close to covering that bigger run of losers, and the overround is skimmed off every bet on top. The result is a flat, mechanical drip. There is no jackpot day to wait for and no losing streak to ride out, just the margin taken race after race until the bank is gone.

The sting in the tail is the comparison. The second favourite loses faster than simply backing the favourite. Picking a sensible-looking runner does not remove the tax built into the odds. It just hides it behind a horse that wins often enough to keep you believing.

The test is deliberately blunt, because a blunt test is the only honest one. We took 27,909 real British races and, in every one, backed the horse the market made second favourite. Flat stakes, the same notional unit on each bet, no progression, no staking plan layered on top. Every bet settled at the official industry Starting Price, the odds returned as the race went off.

The details are where systems usually get flattered, so we were strict with them. Fallers and pulled-up horses are counted as the losing bets they are. A horse that unseats at the second-last or is eased to a stop has cost you your whole stake just as surely as one beaten a length, and any test that quietly drops those non-finishers makes every system look better than it is. We did not drop them.

Joint-favourites were split cleanly so no result could leak in through the back door. Where two horses share favouritism, the genuine second favourite is identified without letting a tie hand the system a free win it did not earn.

A few honesty notes hold for the whole Lab. These figures are to Starting Price with no commission and no allowance for the price drifting against you, so the real world is a touch worse than the numbers say. Starting Price is also the fairest common yardstick, because it is the one odds every punter could actually have had, with no shopping around or early-price luck baked in.

The sample is large on purpose. Twenty thousand-plus races is enough that variance washes out and you are looking at the system's true long-run cost, not a lucky or unlucky season.

Backing the second favourite in every race returned -15.73% to Starting Price across 27,907 settled bets. In plain money, for every £100 you staked you got back about £84.27 over the long run. Stake £10 a race and you hand back roughly £1.57 of every £10, race after race, until the bank is empty.

The strike rate was 20%, so the second favourite won about one race in five. That number matters because it kills the central illusion of the system. The bet feels alive, it places often, it is always thereabouts, but winning one in five is a long way short of the favourite's roughly one in three. The slightly bigger price you take never bridges that gap.

This is not a fragile result balanced on a knife edge. The 95% confidence range runs from -17.7% to -13.8%. Even at the kindest end of that range, the most charitable reading the data allows, you are still losing nearly 14p in every pound. There is no version of the sample where this creeps towards break-even. It is a durable, structural loss, not a run of rough luck.

Now the comparison that punctures the whole "fairer price" story. Backing the favourite on the same races loses about 12.5%. The second favourite loses 15.73%. Stepping one place down the market, the supposedly canny move, costs you roughly three extra pence in the pound. It is also worse than the odds-on favourite at about -7%, the least-bad bet anywhere in the Lab. So every step you take away from the very front of the market makes the result worse, exactly as the favourite-longshot bias predicts.

And this loss is measured to Starting Price with no commission. A real punter, taking real prices and paying a margin to get on, bleeds faster still. The data does not whisper here. It states plainly that the second favourite is a clear, repeatable loser.

So, does backing the second favourite work? No. It is a tidy-sounding way to lose money faster than just backing the favourite: -15.73% to Starting Price over 27,907 real British races, with fallers and pulled-up horses counted as the losers they are. It is a steady leak with no upside to redeem it.

The "fairer price" story is an illusion, and it is worth being clear about why. You are not capturing value the crowd left behind. You are paying the bookmaker's margin on a horse that wins often enough to feel sensible but never often enough to cover the rake. The second favourite is not a mispriced bargain sitting next to the over-bet jolly. It is just the next horse along, carrying the same overround, sitting in a worse spot on the favourite-longshot curve than the favourite it is meant to improve on.

The league table settles the argument. It loses more than the favourite at -12.5%, and far more than the odds-on favourite at -7%. Every step you take away from the very front of the market here makes it worse, not better. There is no corner of the data where this turns a profit, and over short runs a lucky cluster of second-favourite winners is variance, not an edge. It always tightens back to the loss.

It suits nobody as a staking plan. If you take one thing from this experiment, let it be the rule the whole Lab keeps proving: picking a likely winner is not the same as getting paid enough when it wins. A sensible-looking selection does not remove the tax built into the odds, it just hides it behind a horse that runs respectably. The second favourite is the cleanest illustration of that gap we have, which is exactly why we publish it, loss and all.