Last updated: 22 June 2026
A forecast bet is one of the most tempting tickets on the card. You are not just picking the winner, you are calling the first two home in the right order, and the payout reflects that. A few quid returns a number that makes you sit up. So the casual backer scribbles down the two horses he likes the look of, hands over his stake, and waits for the photo. Most of the time the second favourite ruins it by running third.
The seduction of a forecast is the size of the return. The problem with a forecast is that you now have to be right twice, and the dividend has a fat margin built into it before you start. Done carelessly it is one of the fastest ways to give your money back. Done with a genuine read on the shape of a race, it is one of the few bets where a sharp opinion is properly rewarded.
What is a forecast bet in horse racing?
A forecast bet means naming the first and second horses in the correct finishing order. A reverse forecast covers the same two in either order for a doubled stake. The dividend is set by the computer straight forecast, which carries a heavy margin, so forecasts only pay you over time when you have a real read on the top two, not just a hunch.
What a forecast bet actually is
There are three things hiding under the word forecast, and it helps to keep them straight. A straight forecast is a bet on two named horses to finish first and second in that exact order. A reverse forecast is the same two horses to finish first and second in either order, which is really two straight forecasts, so it costs twice the stake. A tricast extends the idea to the first three home in order, and gets harder again.
The payout is not a fixed price you take when you bet. It is settled by the computer straight forecast, the CSF, a formula that prices the result after the race using the starting prices of the runners. Two outsiders filling the frame returns a small fortune. The favourite and second favourite finishing as the market expected returns very little. The CSF is a side market with its own margin, which is the part most backers never stop to think about. It sits in the same family as the other side markets you can reach for when the straight win bet does not appeal.
Why forecasts are harder than they look
A win bet asks one question: which horse crosses the line first? A forecast asks two questions at once and ties them together: which horse is first, and which of the others is second. Getting the winner right and then watching your second pick finish third is the standard forecast heartbreak, and it happens far more often than beginners expect because the gap between second and third is usually tiny.
On top of that you are paying the CSF margin to play. That margin is the forecast equivalent of the over-round on a win market, and it is not small. This is why a forecast is not a casual add-on to a day’s betting. It is a bet that has to clear a higher hurdle than a single win bet, so it only earns its place when your read on the race is strong enough to clear it. The honest test, as with any angle, is whether your forecast selections beat actual versus expected over a real sample at Betfair SP. If they do not, the margin is quietly eating you.
The only races worth a forecast: small, top-heavy fields
Forecasts are a tool for a specific kind of race, not a bet to force into every one. You want a race where two horses clearly stand out from the rest and the remainder look exposed. A field where any of eight could place is hopeless for a forecast, because the second slot is a lottery. A field where two are a class apart and the rest are makeweights is the one you are hunting for.
So the filter comes before the selection. Start by asking how competitive the race really is. If you cannot confidently reduce it to a small group of genuine contenders, there is no forecast here and you move on to the next race. Most cards will give you only a handful of races that qualify, and that is the point. The discipline of passing the races that do not suit is what protects the bet, the same way it protects every other betting system that works.
How to find the two horses that fill the frame
Once you have a race with no more than five genuine contenders, you need a way to separate the strongest two from the rest. The method I have always used for this is a simple overlay comparison. Rate each contender out of ten in four areas that decide most races: class, speed, current form, and connections. You have already done the hard analysis to get to five contenders, so this is just compressing that work into four honest scores per horse.
Then compare the contenders against each other in pairs. Take any two and ask how often one would beat the other across ten runnings of the race. If you make a horse a 7.5 out of 10 against a rival, you are saying you expect it to win that private duel three times in four. Work through every pair and a clear order falls out: the two horses that beat the rest of the field most consistently are the two you use in your forecast, with the stronger of them on top for a straight forecast, or both ways round for a reverse forecast if you genuinely cannot split them.
That pairwise comparison is a hand-built version of what a model does in an instant. Race Advisor’s Race Predictor weighs hundreds of factors at once and outputs a win and place probability for every runner, which is exactly the read you are trying to draw by hand: who is most likely to fill first and second, and by how much. Used for forecasts, the two horses with the highest combined probability of filling the frame are your pair, and the gap between them and the third horse tells you how safe the second slot really is.
Across nearly 141,000 UK races between 2013 and 2026, the Race Predictor’s top two named horses fill the forecast frame, finishing first and second in either order, 14.1% of the time, and land the exact straight forecast order in 7.8% of races. That sounds modest until you split it by field size, which is where the whole forecast case lives. In races of six runners or fewer the top two fill the frame 27.2% of the time, better than one race in four. By fields of nine to twelve it drops to 9.7%, and in a thirteen-runner-plus cavalry charge it falls to 6.5%. The small, top-heavy race is not a preference, it is the entire edge.
You can sharpen the read further by stacking ratings. Where Race Advisor’s PR Rank and VDW Rank both make the same horse their number one, which happens in roughly 44% of races across more than 71,000 runners, that horse wins 34.2% of the time and makes the frame 53.7% of the time, the strongest single pointer in the data to who fills first or second. But here is the catch that decides whether you bet at all: its A/E to Betfair SP, the ratio of actual wins to the wins its price implied, sits at 0.99. Agreement tells you who is most likely to fill the frame. It does not tell you the forecast is value. That second question is settled by the price.
Price the forecast before you back it
Finding the right two horses is only half the bet. The other half is deciding whether the forecast is worth taking at all, and that is a price question. Turn your read into a rough probability that the pair fills the frame in your chosen order, then ask whether the likely CSF dividend pays you enough for that probability once its margin is stripped out. If your two are the favourite and second favourite, the dividend will be skinny and the value usually is not there, however confident you feel.
The forecasts worth backing tend to involve at least one horse the market has underrated, so the dividend is generous relative to the real chance of it landing. This is the same discipline as setting a minimum price on a win bet: you decide what the bet is worth before the market tells you what it pays, and you pass when the number is short. Liking two horses is not a reason to bet a forecast. Being paid more than the result is truly worth is.
So treat the forecast as what it is: a sharp tool for the few races where two horses dominate and the market has not fully noticed. Find those races, separate the top two with a clear method, and only strike when the dividend pays you for the risk. See how today’s runners price up by win and place probability on today’s racecards, and pick one qualifying race to price a forecast on by hand this week before you back it.
