Filter+: build, save and track your own horse racing filters

What is Filter+ in RaceWatch?

Filter+ is the RaceWatch member filter builder. You combine race flags, from a list of 25+, into a filter that produces a shortlist, save the filter, and follow its profit and loss to a £1 SP stake.

Filter+ exists to turn an idea you already have about racing into something you can apply consistently and measure. Instead of reading every card in the same loose way, you write the idea down as a set of flags, let Filter+ build the shortlist, and then read the runners on that shortlist properly on the racecard.

The tracking is what makes a filter worth keeping. Each saved filter has its own profit and loss figure at a £1 stake at starting price, so over time you can see whether the idea behind it holds up, rather than relying on memory of the winners it found.

How do race flags build a shortlist?

Each race flag is a condition, and a filter lists only what meets every flag you have chosen. One flag gives a broad list; each extra flag makes the list shorter.

The simplest way to build a filter is to start with the single flag that matters most to your idea and see how long the shortlist is. Then add a second flag and watch how far the list falls. If one extra flag cuts the list to almost nothing, it is doing a lot of work, and it is worth asking whether that flag really belongs to the idea or is just trimming the list.

Check the flag list inside Filter+ for the exact options on offer, and write a short note on why each flag is in your filter. A filter you can explain in one sentence is easier to trust, easier to adjust and easier to judge when its record starts to build.

Why should I save a filter?

Saving a filter keeps its flags fixed and gives it its own £1 SP profit and loss record, so you judge the idea on results rather than on a handful of remembered winners.

A filter you rebuild from scratch each day tends to drift: a flag gets added after a loser, another gets dropped after a quiet week, and the idea you started with is lost. A saved filter stays the same, which is what makes its record meaningful.

Saving also lets you keep several ideas side by side. Two filters that differ by a single flag, each with its own record, show you whether that flag adds anything. Give each saved filter a clear name that describes the idea, so the list stays easy to read as it grows.

How does £1 SP profit and loss work?

Profit and loss to a £1 SP stake assumes £1 on every runner the filter finds, settled at the official starting price. A winner adds its SP odds, and a loser takes away £1.

Using the same £1 stake on every runner is called a level stake, and it is the fairest way to compare one filter with another, because no runner counts for more than any other. Starting price is used because it is a single, published price for every runner, so every filter is measured on the same terms.

The worked example shows the arithmetic for five runners. Two winners at 4/1 and 2/1 add £4 and £2, and three losers take away £1 each, so the running total is +£3.00. A filter with more losers than winners can still show a profit if its winners come at bigger prices, and the reverse is also true.

How should I read a filter's profit and loss record?

Read the number of runners first, then the profit and loss. A short record can look spectacular or dreadful by chance, and one long priced winner can carry a small sample on its own.

Suppose a filter has found twenty runners and shows a healthy profit. If most of that profit came from a single 20/1 winner, the record is really telling you about one race. Take that winner away and ask what is left. A record spread across many winners, horses and meetings is far more convincing than one built on a single result.

Past profit and loss describes what happened; it does not promise what comes next. Treat a filter's record as evidence about an idea, keep reading the shortlist on the racecard, and expect losing runs along the way, because every filter has them.

• How many runners are in the record?
• How many winners, and at what prices?
• Does one big priced winner carry the total?
• Is the record spread across different horses and meetings?
• Has the filter held up since you saved it?

How do I avoid a filter that only fits the past?

Keep filters simple, make every flag earn its place and give more weight to results after you saved the filter than to anything before it.

With enough flags, almost any set of past races can be cut into a filter that looks brilliant. The trouble is that each extra flag fits the past a little more tightly and makes the filter less likely to repeat. This is known as overfitting, and it is the most common mistake people make when building their own rules.

Two habits help. First, write down the reason for each flag before you add it, and drop any flag you cannot explain. Second, once a filter is saved, leave it alone for a good while and let its profit and loss record build, because results that arrive after you built it are the only honest test of the idea.

Keep reading