Race Filter Lays: How to Filter Races for Lay Betting

A race filter for lays is a set of fixed rules that cuts a day’s racing down to the few races, and the few runners, worth laying. Instead of scanning every card by feel, you decide in advance which race types, field sizes, prices and form figures qualify, then lay only what passes. Done well, a filter removes guesswork. Done badly, it only dresses up a hunch with numbers. This guide covers how filters work, what to filter on, how to test one honestly and how to keep every lay loss capped.

TL;DR

• A race filter is a written set of rules applied the same way every day: race type, field size, price range, form and market signals.
• Laying a horse at 9.0 against a £10 backer’s stake risks £80 to win £10, so the filter has to be tested against liability, not just strike rate.
• A strike rate on its own proves little. Judge a filter on its full results history, priced at the odds you could really get, with losses included.
• Profits from UK betting are tax-free for customers, but the Gambling Commission still expects you to stay in control of time and money.
• Our own signals are the output of a filter like this, published before the race and tracked in public.

Table of Contents

• What a race filter means in lay betting
• Why filter at all
• What you can filter on
• How to read a filter’s results
• A worked example of building a filter
• Our view on filters
• Cap the damage: liability and staking
• Limits and risks of lay filters
• Using filters with live signals
• A simple daily routine
• FAQ
• Sources

What a race filter means in lay betting

Lay betting means betting that a selection will not win. On a betting exchange you act as the bookmaker: if the horse loses, you keep the backer’s stake. If it wins, you pay out. Betfair’s own help page on lay bets describes it in exactly those terms. If you are new to the idea, our plain-English guide to lay betting walks through it from scratch.

A filter answers one question before you stake anything: which races and which runners are we even willing to look at? It works in two layers.

1. Race-level filters decide whether the race is worth your attention at all. Examples are the type of race, the number of runners, the going and the course.
2. Runner-level filters pick the individual horse to oppose. Examples are its price band, recent form, its record at the distance and what the market is doing.

The point of writing the rules down is consistency. A tired layer at 4pm will talk himself into a lay he would have rejected at 10am. A filter does not get tired.

Why filter at all

Laying has an awkward shape. Most lays win, because most horses lose: in a field of ten, nine lose. But the wins are small and the losses are large. That asymmetry is why laying without discipline goes wrong.

Betfair’s worked example lays a horse at odds of 9 against a £10 backer’s stake. The potential profit is £10. The potential loss is £80. One lay going wrong takes back eight winning lays of the same size.

That is the arithmetic every filter must respect. A filter is not just a way to find losers. It is a way to find losers more reliably than the market already assumes, and to avoid the horses where one upset does the damage.

Three practical reasons to use one:

• It cuts volume. A busy day can have dozens of UK, Australian and US races. Most carry no edge for a layer.
• It makes results testable. Rules written down can be checked against past races. A gut feeling cannot.
• It keeps liability predictable. If the filter limits you to a price band, your worst-case loss per lay is known before you start.

What you can filter on

There is no single correct list. These are the categories most layers work with, and what each one is really telling you.

Race conditions

• Field size. Bigger fields mean more losers per race, but also more chances for a rank outsider to run into a place of its own. Very small fields can leave a favourite so short that laying it carries little reward.
• Race type. Handicaps, maidens, novice hurdles and conditions races behave differently. Handicaps are built to bunch the field, which makes any single horse harder to oppose with confidence.
• Going and distance. A horse with no form on soft ground, running on soft ground, is a different lay from one with proven form on it.
• Course and country. A filter built on UK data should not be assumed to work on Australian or US racing. Track layouts, field sizes and the way markets form all differ.

Form figures

• Recent form and days since last run. A long absence and a poor last run are common reasons to oppose a short-priced horse, though layoffs also hide fitness the market has noticed.
• Course, distance and going record. Does the horse have a win or a place at today’s conditions, or is it guessing?
• Strike rate. The share of a horse’s runs, or a trainer’s or jockey’s, that end in a win. It is the base statistic most layers build on, though it needs a decent sample to mean anything.

Market signals

• Price band. Most layers set a ceiling on the lay odds. The higher the odds, the higher the liability for the same stake, so a ceiling is your first line of defence.
• Market movement. A horse drifting out in price tells you something different from one being backed in. Neither is a signal on its own, but both are context.
• Liquidity. If there is little money on offer at your price, you will not get matched. A good filter only includes markets deep enough to trade.

How to read a filter’s results

This is where most filters fail. Finding a rule set that would have worked last year is easy. Finding one that will keep working is not.

Strike rate is only half the story. A filter that wins 85 lays in 100 sounds excellent. But if the 15 losers are at odds of 10.0 or more, the maths changes. Take an illustration with round numbers: you lay at 6.0 with a £10 backer’s stake each time. Each win earns £10 and each loss costs £50. Over 100 lays, 85 wins give £850 and 15 losses cost £750. That leaves £100 before commission. The same 85% at 9.0 odds means 15 losses of £80, which is £1,200, against £850 of wins. A loss overall. Same strike rate, opposite outcome. These figures are an illustration of the arithmetic, not results from any filter.

So always read three numbers together:

1. Strike rate: how often the lay wins.
2. Average lay odds: which sets the size of each loss.
3. Overall profit and loss: after losses, in units of stake or liability.

Use the price you could have had. Many results lists use the price at the off. If your lay would have been matched earlier at a different price, your real record differs. Check whether results are priced at published odds or at hindsight odds.

Count every signal. A results page that quietly drops losers is worthless. The honest version shows every selection, wins and losses, with dates.

Beware small samples. Twenty winning lays tell you very little. Variance in lay betting is lopsided, and a long run of wins can end with one costly loser. Judge a filter over hundreds of selections, not dozens.

Watch for over-fitting. If you keep adding conditions until the past results look perfect, you have described last year’s races, not a method. A sound filter has few rules, each with a plain reason behind it.

A worked example of building a filter

Here is how a sensible layer might build a simple filter. The numbers are for illustration, not a recommendation.

Then apply it to historical races and record what happened, listing every qualifier and not just the pretty ones. Split the results into halves. Build the rules on the first half and test them on the second, which the rules have never seen. If the second half collapses, the filter was fitted to noise.

Only after that should you stake real money, and then at small size.

Our view on filters

We built DonkeyRadar on one belief: a lay signal is only worth having if you can check it. Our process uses historical strike rates and live market prices to identify the weakest runner in a race, across UK, Australian and US racing. The selection is data-driven, with tested statistical parameters rather than opinion.

Three things matter to us more than any headline figure.

• Signals go out before the race. Every selection is published ahead of the off, so nobody can claim a result in hindsight.
• Results stay public. Wins and losses are recorded and kept. We report a strike rate of over 85%, and we also report the losers. It happens. Transparency is the whole point.
• Strike rate is not the target. Liability is. A filter that wins often but loses big is not a good filter, which is why our signals come with staking tier grading.

If you build your own filter, hold it to the same standard. Write the rules before the race, record every result and look at profit and loss, not just the win count.

— Donkey

Cap the damage: liability and staking

A filter picks the lay. Staking decides whether one bad result hurts or ruins you.

On an exchange, your liability is your maximum loss on a lay. Betfair states it as the backer’s stake multiplied by the odds minus one. Lay a £10 stake at 5.0 and your liability is £40. That money is held from your balance until the market settles, as the exchange guide from RaceWatch also explains.

The cleanest approach is to fix the liability rather than the stake. Choose a small share of your bank, commonly 1 to 2%, as the most you will lose on any one lay, then work backwards to the stake that gives it. A £1,000 bank with a 2% cap means £20 liability per lay. At odds of 5.0 that is a £5 backer’s stake. At 9.0 it is £2.50. Higher odds, smaller stake, same worst-case loss. Our guides on fixed liability staking and capping lay losses at 1 to 2% go through the method step by step, and the Betfair liability explainer covers how the figure is calculated and held.

A few practical points:

• Minimums apply. Betfair’s Starting Price lays carry a minimum £10 liability, and unmatched SP lays below that are cancelled. Small banks need to plan around this.
• Use a calculator. Our lay betting calculator shows liability, profit and break-even before you place the bet.
• Do not chase. A filter has losing days built in. Raising stakes to recover is how a capped system becomes an uncapped one.

Limits and risks of lay filters

Be straight with yourself about what a filter cannot do.

• It does not remove variance. A strong strike rate still contains losers, and the losers are bigger than the wins.
• Markets adapt. If a pattern is widely known, prices shorten to reflect it and the edge fades.
• Data can be wrong or thin. Form figures for a first-time runner or a rarely seen course mean little.
• Late changes matter. Non-runners, going changes and market moves after you have set up your lay can change the whole picture. Check before the off.
• In-play is a different game. Prices move fast once a race starts. The Gambling Commission’s guidance on in-play betting notes that in-play is fast-paced with prices amended continuously. If you lay in running, read our in-play lay betting guide first.

Tax and staying in control

For UK customers, gambling winnings are tax-free regardless of the amount won, according to the Gambling Commission. Our UK betting tax guide covers what that means for exchange layers. Tax-free does not mean risk-free, though. The same Commission guidance expects operators to watch for signs of harm, including changes in how often people bet and how much time or money they spend. Apply the same test to yourself. If you are betting more often, staking more than you planned or chasing losses, stop. If gambling is causing you worry, talk to your operator about deposit limits or self-exclusion, or contact the National Gambling Helpline. Betting is for over 18s only.

Using filters with live signals

Building and maintaining your own filter takes time: data, testing and constant upkeep. Many layers prefer to start from tested output and work backwards.

DonkeyRadar publishes lay signals for UK, Australian and US horse racing, each one chosen by our statistical filter from historical strike rates and live market prices. The free tier gives you the daily signals. DonkeyRadar Pro costs £29 per month after a seven-day trial and adds real-time alerts by email and Telegram, the full results history and API access for trading software. The dashboard links straight to the Betfair Exchange market, and each signal carries a staking tier grade so you can size the liability sensibly.

You can see how the signals work on today’s lay tips page, read about the method in our Betfair lay betting strategy guide, or start with DonkeyRadar Free. If you want to see how to place the lay once you have a selection, our step-by-step guide on how to lay a horse on Betfair takes you through it.

A simple daily routine

Whether you use your own filter or our signals, a repeatable routine beats improvisation.

1. Check the filter output the evening before or in the morning. Do not add selections later because a horse looks wrong.
2. Confirm the market. Check for non-runners, going changes and enough liquidity at your price.
3. Set the liability, then the stake. Use your fixed cap and a calculator.
4. Place the lay and note the price you got. Record the odds actually matched.
5. Log the result. Wins and losses, same day, same format.
6. Review monthly, not daily. One bad day tells you nothing. A hundred selections tell you something.

For broader ideas on pairing selections with the best available prices, see our notes on odds comparison for UK lay bettors, and for more on the approach behind laying the field, lay the field with liability caps.

FAQ

What is a race filter in lay betting?

It is a fixed set of rules that decides which races and which runners you will consider laying. The rules can cover race type, field size, price band, form and market signals. Because they are written down, you can test them on past races and apply them the same way every day.

Is a high strike rate enough to trust a lay filter?

No. In laying, the losses are larger than the wins, so a high strike rate can still lose money if the losing lays are at long odds. Look at the strike rate, the average lay odds and the overall profit and loss together, across a large sample with every selection included.

How much should I stake on each lay?

Fix the liability, not the stake. Many layers cap each lay at 1 to 2% of their bank, then work out the backer’s stake from the odds. Remember that Betfair SP lays need a minimum £10 liability, so check this against your bank size.

Are lay betting profits taxable in the UK?

The Gambling Commission states that UK customers’ gambling winnings are tax-free regardless of the amount won. Our betting tax guide explains how that applies to exchange layers. Tax rules elsewhere differ, so Australian and US bettors should check their own country’s position.

Can I use a filter on Australian and US racing too?

Yes, but build and test it separately. Field sizes, track types and market behaviour differ by country, so a rule set that works on UK racing should not be assumed to carry over. Our signals cover all three countries, each assessed on its own data.

Sources

• Betfair Support: what does the term Lay mean
• Betfair Support: Starting Price (SP)
• Betfair help: lay liability
• RaceWatch: Betting Exchange Explained
• Gambling Commission: In-play or in-running betting
• Gambling Commission: Winnings from gambling for customers

Recommended

• 5 Defence First Questions for Race Selection When Laying Horses
• 100 Bets to Test Odds Filter Lays for Exchange Bettors
• Horse racing data analysis for lay bettors
• 1–2% Bankroll Caps: Liability First Lays for Handicap Race Favourites