How I pick the horses

I start at the top of the betting

Every horse on my list is near the front of the market. That's where the form is honest and the money is deep, and it's why 6 in every 10 of my bets win.

Backing favourites isn't the hard part, though. Anyone can find the horse at the top of the betting. The hard part is knowing which ones to leave alone, because a lot of them are short for the wrong reasons.

Then I cross most of them out

A horse is short for one of two reasons. Either it's the best horse in the race, or everyone remembers its last run. Only the first kind pays.

A finishing position only tells you so much. A horse might have travelled well and faded late. It might have been slowly away, or never found a gap. You can always find an excuse for the last run, and so can everyone else, so by the next race the excuse is already in the price.

An excuse explains the last race. What I want to know is what's changed for this one.

Maybe it's dropping back in trip after fading. Maybe it meets a weaker field. Maybe it has first time cheekpieces from a yard in form. If something has changed for the better, it can go on your list. If all I've got is a hard luck story, you never hear its name.

So I ask the same questions of every horse near the front of the market. Was it finishing in the first three before its last run? How far was it getting beaten? Has the trip changed? How long since it ran? And everyone checks that last one, but I also check how long since it won, and where it's rated today against the mark it won off. A horse that's drifted well below its last winning mark isn't automatically a bet, whatever the forums say. It's a question, and the rest of the form has to answer it.

A horse turned out again in six days used to be a warning sign. Trainers know a lot more about each horse's fitness than they did twenty years ago, so I don't treat a quick return or a long break as good or bad on its own. I look at what that horse has done after a gap like it before.

A horse that fell and a horse that was pulled up both failed to finish, but they're not telling you the same story. The first may have been going well. The second was usually being beaten. Both get a letter in the form book, and both get avoided. I want to know which story it is.

And the price still has to do its job

A horse can be the likeliest winner in the race and still be too short to bother with, because everyone else has spotted the same thing. Those get left out too. Being right about the horse isn't enough. The price has to leave something in it for you.

I judge every part of what I do by what it returns after commission, not by how many winners it finds. This autumn I took a set of place bets out of my extras. They were landing most of the time. They also weren't paying for themselves, and once they were gone the return on every pound staked went up and the worst dip got smaller. A reassuring run of winners isn't a reason to keep a bet. The return is. And I didn't fill the gap with other bets just to keep the numbers up.

Most horses go on twice: to win, and to place

The win bet is the loud one. My win bets land 43% of the time at an average price of 3.19, and that's where most of the +190 points has come from.

The place bet is the quiet one. It lands nearly 8 times in 10, at an average price around 1.37, so it keeps your bank steady through the weeks when the winners are slow.

Three kinds of bet

• The main shortlistMost of what you get. It's the core of the whole thing, and it has made money in every year on record.
• The extrasOn busier weeks there are a few more win bets from a second set of my rules. They only go on when they add something the main list doesn't already have.
• The comeback betA jumps horse that fell, unseated or pulled up last time, is back within 60 days at a price no bigger than 5, and whose form before the fall says it had the ability. These go on at half stakes, and only since 2026, because they're the newest part of what I do. Why I back horses that fell last time.

The results pick the method, not me

I don't run one method and hope it holds. I run several side by side and keep score on all of them.

Every three months I look at which has made the most over the last two years, with the recent months counting double, and that one gets the biggest stake for the quarter ahead. The decision is always made on bets that had already run. I've binned ideas that sounded clever because the figures didn't move. The method that gets the biggest stake each quarter is the one that earned it.

The main list never has more than four points on in a day, and the extras and comeback bets share a limit of four more. In practice the most that has gone on in one day is just over six. No single bet ever decides your month.

Putting the bets on

Everything goes on at the Betfair Exchange, at starting price. You only need the one account, and it takes a couple of minutes in the morning.

1. Open the race on the Betfair Exchange. Use the "Win" market for win bets and the "To Be Placed" market for place bets.
2. Click the horse's back price and put in your stake.
3. Tick "Take SP" on the bet slip. You get the starting price, the same as my results, and you can forget about it until the evening.
4. Check how you got on in your members' log in that evening.

How much to stake

My results are in points, and you decide what a point is worth. Win bets are 1 point, and in some quarters half. Comeback bets (horses that fell last time out) are half, and place bets a quarter or a half, with the stake on every tip.

Keep a bank of 40 points. The worst dip in the record took 14 of them, so with 40 a bad spell never worries you.

A point is worth	Your bank	A 1 point win bet
£5	£200	£5
£10	£400	£10
£25	£1,000	£25

Only ever use money you can afford to lose. Safer gambling.

See exactly what I do for two weeks.
For only £5.

Every bet I make, by email each racing morning, with the stake and my line on why. Pay £5 today, then £49 for the winter pass before day 15.

Start your trial Or get the free guide

If your first month on the £49 loses, I refund it. Cancel with one email. Nothing renews after 31 March. Full details.