If You're Tired Of Backing Losers And Following Tipsters Who Disappear When The Results Turn Bad...

This Is The True Story Of How I Created AI Betting Software That Turned £25 Into £3,387.50 In 7 Days...

My software has now processed over 80,000 UK race records. It removes human emotion entirely and sends you 2 to 3 selections every morning before your alarm clock goes off

80,000+RACE RECORDS
UNDER 5MINUTES EACH MORNING
2 TO 3SELECTIONS A DAY

To the punter reading this,

If this page has found its way to you, read the next part slowly, because it concerns your money...

Today's 3 Selections Could Return Up To £317.45


This page isn't advertised, and it only goes out through a few partners I trust...

Before you reach the end, I'll tell you who passed it on to you and why.

First, though, the part that matters to your wallet...

By teatime today you could have profit sitting in your account, ready to withdraw.

Back the 3 winning selections Bet Detector Pro sends you this morning, and...

That's between £159.75 and £317.45 in instant cash


If you've been burned by bold promises before, that figure probably makes you wince.

Fair enough, so instead of asking for trust, I'll show you the workings and let the people already using it speak for themselves.

Stick with the selections for 7 days and...

Up To £3,387.50 Over 7 Days, Ready To Withdraw


Where does a number like that come from?

From 10 people who've been on the selections for the past week.

They found this page the same way you did, 7 days ago, and were every bit as doubtful.

I asked each of them to log every bet, the stake, the price and the result, and send me the totals at the end.

Winners and losers went in the log alike, with nothing rounded up and nothing tidied away.

Three of them wrote back with this after their first 7 days...

What The First Week Looked Like


“I put £50 on every pick and didn't touch a thing beyond that, no skipping, no adding my own fancies. Seven days later my Betfred balance had gone from £400 to £2,190 and I still can't quite work out how a piece of software spotted horses I'd have walked straight past.”

Malcolm Foster, Doncaster

“I only paid the £20 so I could prove to my brother-in-law it was rubbish, and by the Thursday I owed him a pint and an apology. Five winners from my first eight bets, £760 in front, and the only regret I've got is the fifteen years I wasted doing it my own way.”

Sandra Walkden, Carlisle

“I drive a delivery round and I'm out the door by half six, so I read the email at the traffic lights and have the bets on before I've finished my first flask of tea. The money side speaks for itself at £905 up in a week, but what really sold me was never once having to sit there wading through form like I used to.”

Neil Ashworth, Stoke-on-Trent

When the replies came in, I noticed how alike their backgrounds were...

Not one of them had ever made betting pay


Several had paid for tipping services on and off for years...

Renewing each time in the hope that next month would be the one...

Few had ever tried to pick horses with a method behind it...

Fewer still would know how to test whether a method worked...

Not one of them works with numbers for a living...

And most have jobs and families that leave them about 10 minutes free before work.

That's not a criticism, because it describes most people who've ever had a bet.

All 10 Finished The Week In Front


Luck had very little to do with it.

Bet Detector Pro picked horses whose prices didn't reflect their real chances, which is something you can't spot by eye at 7 in the morning.

And for once, each of them followed a method instead of a mood.


2026 £0k £2k £4k £6k JAN 2026: £3,690.55 £3,690.55 JAN FEB 2026: £5,120.30 £5,120.30 FEB MAR 2026: £5,540.85 £5,540.85 MAR APR 2026: £6,024.60 £6,024.60 APR MAY 2026: £5,310.45 £5,310.45 MAY JUN 2026: £5,880.70 £5,880.70 JUN JUL 2026: £6,240.90 £6,240.90 JUL AUG 2026: £5,427.85 £5,427.85 AUG SEP 2026: £6,106.45 £6,106.45 SEP OCT 2026: £6,750.35 £6,750.35 OCT

Where The Average Punter's Money Goes


Racing can be beaten, but not by the usual routine


The usual routine is backing a name you like at a price everybody else has already taken.

Bookmakers understand that routine much better than the people following it, and their adverts are built around it.

You'll notice those adverts show famous faces and big-race glamour, and almost never a price.

Add a tipping service that only ever emails about its winners, and the routine is complete.

It's a habit, and a costly one, which the bookmakers are happy for you to keep.

Try this: think about your last 10 bets and count how many you could explain to a stranger without mentioning a feeling.

Most people get to 2 or 3 before they run out of reasons.

Nobody should feel bad about that, though it goes a long way to explaining the losses.

Bet Detector Pro takes feelings out of it entirely


Every selection comes from a model working through prices and probabilities each morning, with no opinions involved.

I could fill the next 20 pages with how it all works...

Every factor and every test behind it...

I'll spare you most of it...

You want winners you can back quickly, not a lecture.

Your mornings are probably full enough already...

And the maths behind it won't make your bets any easier to place...

So the short version is this, a handful of horses in your inbox...

Placed in a couple of minutes, then forgotten about


And by the evening, a fair chance of having winners to show for it.

Nobody sensible expects a betting service to make them rich in a month...

And nothing wins every race...

Losing days will happen, the same as they do for every professional...

What you want is a service that wins often enough to matter...

Expect 2 Or 3 Winning Bets A Day


That's what Bet Detector Pro is built to give you...

A betting balance that grows week by week...

Money you can move to your bank whenever a bill lands...

And the rest left in place to keep growing while you're at work.

There's nothing unreasonable about wanting that...

Yet most services never deliver it


You've probably paid for one or two of them yourself...

The kind that email you in capital letters when a short-priced favourite obliges...

And say nothing at all about the run of losers that came before it.

I've Paid For Those Services Too


And I've wasted more money on them than I'd like to say.

One sent me a horse at 9am, watched it finish tailed off, and emailed at teatime to blame the ground.

Another was full of swagger all summer and impossible to reach by October.

The worst one edited its own results page whenever a losing day spoiled the look of it.

I stopped paying any of them


Not one was worth the subscription.

What I wanted was something that behaved the same on a wet Monday at Southwell as it did on a big Saturday at Ascot.

Selections built on evidence, not on whoever was in form that week.

My guess is you're here for the same reason...

You'd like your betting to finally pay its way...

You're tired of handing money to tipsters who don't deliver...

And tired of losing winnings on hunches and badly thought-out picks.

So let me put it plainly...

Your £50 Stakes Could Grow To £3,387.50 This Week


There's no work for you to do...

A few spare minutes each morning is the full extent of it...

I've set everything up so that making money takes as little effort as possible.

You won't install anything or study a single chart.

If you can read an email and tap a button on a betting app, you're already qualified.

What Happens Before You Wake Up


I promised I'd keep the science short, and I will.

A few of you will still want to see what goes on in the early hours, so I'll walk you round it briefly.

It starts with a lesson from my trading days


Every price a bookmaker offers is somebody's opinion, and opinions can be wrong.

Each runner has a real chance of winning and each runner has a price, and the two seldom line up.

Back a horse when its price is longer than its chance deserves and you've got a bet worth having, and when the price is shorter, you leave it.

My software spends each morning measuring how far apart those two numbers sit.


MY PRICE
THE BOOKIES' PRICE
THE BET

The difficult bit is putting a believable number on a horse's real chance.

Reading a price off an app takes a second, while pricing a horse with any accuracy took me the best part of 2 years to get right.

By the time the model runs, the declarations are in from the previous day and the clerk of the course has updated the going.

As soon as the first prices appear, Bet Detector Pro starts work.


1 80,000+ 2 £10 AT 6/1 = £60 3 BEFORE 8AM

Step 1: Pricing Every Runner From Scratch


Before the markets open, the model takes every runner declared for the day.

Each one is checked against more than 80,000 past race records and scored on a long list of weighted factors.

Some factors are the obvious ones


Recent trainer and jockey form over the past 14 days, split by this course, this trip and this ground.

The going, judged against what each horse has shown it can handle rather than what connections hope it can.

Course form, since some horses come alive at one track and never run a race anywhere else.

Days since the last run, compared with the break its own yard usually leaves before a winner.

Movement in class, and how big the step up or down is.

Field size, and whether this horse copes with a big, rough handicap or prefers a small, tactical one.

The draw, and how that stall has tended to play here on similar ground.

Then the ones most punters skip


The likely pace, because a hold-up horse in a slowly run race starts with a handicap of its own.

Today's weight compared with what it carried in its last big run.

First-time headgear, or a jockey booking that tells you the yard means business.

Even the miles travelled to the course, which take more out of some horses than any form book records.

And one factor I'm especially pleased with, because hardly anyone else measures it.

The rail that changes the race


After rain, or to save worn ground, the running rail gets moved out, and a race billed as a mile can end up a good few yards longer.

For years that extra distance never showed up in the form, so a lot of horses were marked down for failing to stay a trip they'd never been asked to run.

My model uses the distance that was really run, which puts those horses back where they belong.

The difference on one race is small, but across 80,000 records it changes how hundreds of horses are rated.

Each factor earns a score, and none is ever allowed to decide a race on its own.

How much each one counts depends on how much it has mattered in similar races in the past.

Getting those weights right took the longest


The draw can count for a lot over five furlongs on a tight, turning track and barely at all over three miles on a galloping one, so the weights shift from race to race.

So a sprint and a staying chase never get scored with the same recipe.

Treating a two-year-old maiden on the all-weather like a three-mile chase in the mud is one of the commonest mistakes in racing.

It borrows a trick from the City


Each horse is priced against the rivals it actually faces that day, because a horse's chance depends on the opposition as much as on its own ability.

A decent sort can look a certainty in a weak field and an also-ran in a strong one, without anything about the horse changing.

Above that sits the AI layer, a machine-learning model trained to find factors that work together.

That's how it picks up combinations a fixed set of rules would miss.

Ask a tipster which factors matter most at Chester in May and you'll get an answer, though rarely one he can back up with figures.

The model keeps every factor in play for every runner, and the last race on the card gets the same attention as the first.

The morning run takes minutes


It covers every declared runner at every British meeting, tens of thousands of small calculations, and hands me a shortlist before my second coffee.

I once worked out how long the same job would take me by hand, and the answer was somewhere around the following Tuesday.

Every runner ends up with one number


That number is its percentage chance of winning, and the figures for every runner in the field must add up to 100%.

So if the favourite's chance goes up, another runner's chance has to come down, which stops the model rating every horse it likes.

Then it's basic arithmetic, because a horse rated a 25% chance is worth backing at anything bigger than 3/1, and not a penny shorter.

If a bookmaker offers 9/2 about it, the price is too generous, and those are the horses the model is looking for.

If the best price is 2/1, the horse gets left alone, however good it looks in the paddock.

Sticking to that rule without exception is behind every profitable month you're about to see.


25% · £10 2/1 2/1: £20 profit £20 3/1 3/1: £30 profit £30 9/2 9/2: £45 profit £45

When a horse scores well enough and its price is clearly bigger than its chance, it goes on the list.

That horse becomes your selection


Some races produce nothing at all.

Some mornings there are 3 selections and some mornings just 1, and both are the model working as intended.

Sending more bets just to look busy would only cost you money.

One last check happens before the email goes out.

The model looks at the day's selections together, the way a risk manager looks over a trading book, to make sure they don't all depend on the same assumption.

If two horses only qualify because rain is forecast, and the rain never comes, both are in trouble at once.

So it caps how much of a morning can ride on any one idea, whether that's the going or one trainer's runners.

On the trading floor, stacking every position on the same outcome without realising it was the quickest way to lose your job.

Who Sets The Price, And Why It's Often Wrong


Most punters never ask where a price comes from


A bookmaker's first job is protecting their book, and forecasting the race comes a long way second.

The opening show is a rough starting point, and after that the price moves with the public's money.

When thousands pile onto a famous trainer's runner, the price shortens because the bookmaker's liability has grown, while the horse's chance hasn't changed at all.

Meanwhile an unfashionable horse from a small yard can drift out unnoticed, purely because nobody is backing it.

Then there's the bookmaker's margin, a few percent built into every market so the house wins over time.

Add up the chances implied by every price in a typical race and the total comes to well over 100%, and the extra is the bookmaker's cut.

Bet Detector Pro removes that margin first, so it's comparing like with like.

Ordinary punters have been losing for years because they back the same popular names as everyone else, at prices shaped by everyone else's enthusiasm.

Bet Detector Pro goes the other way


It works out its own price for every runner and only then compares it with the bookmaker's.

Usually the two are close, and when they are, the race is left alone.

Now and then the public's love for one horse leaves another priced well beyond its chance, and that's the moment the model has been waiting for.

It has no interest in beating the bookmaker in every race, only in the handful a day where the two prices sit far enough apart.

Waiting for those races is what separates a professional from a punter.

Combinations A Form Book Can't Show You


The model really earns its keep when factors start working together.

Look at a 40-day break by itself and the form book won't tell you much.

Look at a drop of one class by itself and you'll learn about as little.

But one particular yard's horses, back from a 40 to 60-day break, dropping in class, on softer ground than last time, might win a lot more often than their form suggests, and nobody notices, since nobody is tracking that particular mix across tens of thousands of races.

Those are the patterns the model is built to find.

It ignores what's already obvious and priced in, and goes after three and four-way overlaps you can only see when you look at everything at once.

Testing turned up dozens of them, a few of which I'd have dismissed if the results hadn't kept backing them up month after month.

There's nothing mysterious about any of them


They're relationships in the data that would take one person years to find, and no busy punter reading the paper over breakfast is going to stumble on them.

The model also drops a pattern as soon as it stops working.

A pattern has to keep paying its way, and the ones that fade are retired straight away.

Why One Man's Opinion Isn't Enough


Even a good tipster is one man's opinion, well presented.

He watches a few tracks, follows a handful of yards he trusts, and backs what he fancies on the day.

When those yards are flying he looks brilliant, and when they go off the boil he's got nothing else to lean on.

My model has no favourite yards, because it doesn't guess.

It reads every race on the card, including the ones nobody else finds interesting, and it has never backed a horse for its name.

Nor does it have an ego to protect.

A tipster who has backed a horse in public will often stick with it long after the evidence turns, because changing his mind feels like losing face.

My model changes its mind the moment the numbers change.

Step 2: Winners At Prices That Pay


You won't find Bet Detector Pro propping up its record with odds-on shots.

Its winners come right across the price range, including horses at 5/1, 8/1 and bigger.

That's the big difference from most tipping services.

Most of them keep their winner count high by sending short-priced favourites, which earns tiny returns at best and an ugly losing run at worst when those favourites get turned over.

Bet Detector Pro does the reverse


It goes looking for horses whose price is clearly bigger than the chance the model gives them.

Landing winners regularly at those prices is what produces the monthly profits you'll see shortly.

This is where good selections become real money in your account.

When winners keep landing at 4/1, 7/2 and bigger instead of odds-on, even a small balance starts to snowball.

A loser costs you one stake, while an 8/1 winner hands you eight back, so a couple of those a week covers the losers and leaves a tidy profit on top.

That's how members betting level stakes finished their first week hundreds of pounds up without needing a big-priced miracle.

The arithmetic looks after itself, as long as you keep placing the bets.


JAN - MAR: £14,351.70 APR - JUN: £17,215.75 JUL - SEP: £17,775.20 OCT: £6,750.35 2026 £56,093.00 JAN - MAR £14,351.70 APR - JUN £17,215.75 JUL - SEP £17,775.20 OCT £6,750.35

Bigger prices mean smaller stakes still pay


Someone staking £10 a selection sees a single 6/1 winner add £60 profit, which is why so many members start small and build up as their balance grows.

The stake suggested in each email is worked out too.

The wider the difference between the model's price and the bookmaker's, the more of your betting bank it suggests putting behind the horse.

Every bet has a firm ceiling however tempting the price, because the traders who last are the ones who never let one result sink them.

If you'd rather stake the same amount on everything, as several of the test group did, that works perfectly well too.

Step 3: Five Minutes, Then Get On With Your Day


There's no need to understand the model.

You won't need to study form or sit refreshing prices on your phone.

Before 8am UK time, an email arrives with the horse, the bookmaker and the suggested stake.

Placing the bets takes under 5 minutes, and the rest of the day is yours.

Check your account in the evening to see what's landed.

Members with full-time jobs, early shifts and young families tell me it's the easiest thing they do all morning.

The self-control is built into the model, so you never have to find it yourself.

Here's what lands in your inbox


The horse and the race time come first, so you can find it in seconds on any app.

Next is the bookmaker with the best price when the selection was made, and that price.

Then the suggested stake, and the shortest price at which the bet is still worth taking.

No essays on breeding and no long race previews, just what you need to get the bet on.

Over 80,000 UK race records behind every selection.

Winners right across the price range, not just favourites.

2 to 3 selections in your inbox each morning before 8am.

Less than 5 minutes a day to place your bets.

15 profitable months in a row, and counting.


15 AUG: £4,180.60 £4.2k AUG SEP: £3,920.40 £3.9k SEP OCT: £4,510.75 £4.5k OCT NOV: £4,760.20 £4.8k NOV DEC: £3,845.90 £3.8k DEC JAN: £3,690.55 £3.7k JAN FEB: £5,120.30 £5.1k FEB MAR: £5,540.85 £5.5k MAR APR: £6,024.60 £6.0k APR MAY: £5,310.45 £5.3k MAY JUN: £5,880.70 £5.9k JUN JUL: £6,240.90 £6.2k JUL AUG: £5,427.85 £5.4k AUG SEP: £6,106.45 £6.1k SEP OCT: £6,750.35 £6.8k OCT 2025 2026

What This Won't Do For You


If you're hoping to turn £50 into £50,000 by next weekend, this isn't for you.

Bet Detector Pro makes no promise like that.

It's a data-driven way of winning more bets than you lose, at prices that turn those wins into profit over time.

You'll have losing bets and quiet days, as every real bettor does, and any service claiming otherwise is lying.

What the model aims for is that over any rolling 7-day or 30-day stretch, the winnings comfortably outweigh the losses.

The month-by-month figures I'll share shortly show how that has played out.

Not one losing month, and that's down to an advantage that adds up over time rather than every bet winning.

All it needs from you is patience


Back every selection as sent, without adding your own bets or skipping the ones you don't like.

The model doesn't need your opinion, only your bets.

The members who do best treat it like a long-term plan, not a scratchcard.

A winning Tuesday doesn't excite them much and a losing Wednesday doesn't worry them, because the month is what counts.

'Bet Detector Pro' finds winners week in, week out...

Using a tested set of selection rules...

Plus a handful of parameters most people would never think to use...

How A Trader Ended Up Building A Racing Model


It took years before the software made a penny...

But I kept at it, and once it finally worked...

Things started to move quickly...

In the right direction, for once.

First, though, a little about me...

My name is Brian Lewis, married with 3 children...

I spent over 30 years in the financial sector.

The hours were long and the lessons were longer


Most of all about using AI and predictive models to get results.

I started at 18 on the floor of the London futures exchange, in a runner's jacket two sizes too big.

It's a funny job title to look back on now that I spend my mornings with racehorses, but a runner carried dealing slips between desks and fetched coffee for whoever shouted loudest.

It was the best education going, as I spent 5 years watching up close how the people around me made money and lost it.

That was back when a trading floor was loud, before screens took over.

The ones who blew up their accounts were nearly always the bright ones who trusted their gut over the numbers in front of them.

The ones who lasted were the quiet ones who did what the model said, even when every instinct told them otherwise.

An old trader on our desk had a saying he repeated most mornings, that the number doesn't care what you had for breakfast.

It took me years to see he was right


By my thirties I'd moved off the floor to a desk that built and ran pricing models, working out what something should be worth before the market made up its mind.

The job was spotting when a market price had wandered away from the evidence, and acting before it wandered back.


2026: £3,690.55 to £6,750.35 £3,690.55 JAN £6,750.35 OCT £0k £7k

Predictive software made most of my trading decisions.

It beat anything I could manage on my own by a distance.

It could get through more data in an afternoon than I could in a lifetime.

Give it the right parameters and it would find the best outcome available.

Rules First, Feelings Never


That approach made the firm a fortune over the years.

Bonuses were healthy, and we beat the market more years than not.

What stays with me most from that time is the morning meeting.

At 7 o'clock every day, before the markets opened, the desk gathered to go through whatever the models had flagged overnight.

You could only bring an opinion to that meeting if you had numbers to back it, and anyone arguing with the model on a hunch got short shrift.

Once the trades were agreed and placed, nobody touched them for the rest of the day.

Bet Detector Pro runs that same meeting for horse racing, on your behalf, before 8am each day.

The Embarrassing Bit


All week I trusted models with other people's millions, then on Saturday I'd walk into the bookies and back a horse because I liked its name or the jockey's colours.

I lost for years doing that, using the one part of my judgement my whole career had taught me to ignore.

I'd pick a horse on a Friday night because a radio pundit said it had been working nicely at home.

I'd back the jockey who'd won the big race the week before, as though winning were catching.

And when a hunch came in I told everyone, while the losers always had an excuse that had nothing to do with me.

It took me an embarrassingly long time to see what was under my nose.

After three decades in the industry, I understood predictive models inside out.

I'd seen colleagues build systems that pulled profit out of sloppy, lazy markets year after year.

The principle never changes


Find a market where prices are often wrong, then only bet when your own figures disagree with it by enough to matter.

Why I Chose Horses Over Everything Else


People ask why I picked horses after a lifetime in money markets.

Racing is one of the last big markets still driven by feeling rather than fact.

Millions of pounds are staked every day on famous names, lucky colours, pub tips and jockeys someone once saw ride a winner.

In the markets I came from, that kind of emotional money was squeezed out decades ago, but at the racecourse it's alive and well and helping to pay my mortgage.

Wherever feelings set the price, someone with a cool head and the right tools can make a profit.

What finally pushed me to do something about it was a Saturday afternoon about 4 years ago.

I'd lost on every race I backed, and on the drive home I did something I'd always avoided, I added up what I'd given the bookmakers over the previous 12 months.

I won't share the figure, but on the trading floor it would have got me escorted out of the building.

That number stayed with me for weeks


I'd never have accepted it from a junior trader, so why was I accepting it from myself?

So I decided to fix it


I'd build my own AI betting software and treat racing like any other market I'd worked in.

If the approach worked on a trading desk, I couldn't see why it wouldn't work on my bets.

It was a long slog, and I leaned heavily on an old colleague, a softly spoken systems man called Iain.

Iain spent 20 years keeping our pricing models running, which mostly meant finding out why they'd broken at 3am and fixing them before the open.

He'd retired to the Welsh borders and spent his days walking the dogs and complaining that his brain had nothing to chew on.

When I rang and explained the idea, there was a long pause before he asked how soon I could send the data.

We did nearly all of it on video calls, me at home and Iain in his study, sharing a screen late into the night.

The first 6 months went on cleaning and organising data, with nothing glamorous about it at all.

Racing records are a mess


They're full of contradictions and missing values, plus patterns that look meaningful for a fortnight and then vanish.

Distances were logged in furlongs in one place and yards in another, and going reports changed mid-afternoon with no record of when.

Some older results didn't even add up, with more finishing positions listed than horses that ran.

Trustworthy had to come before clever


Every race had to be laid out the same way, every course, trip and going description, going back years.

Anything we couldn't trust went in the bin, because a model fed rubbish will hand you rubbish back with total confidence.

We also had to stamp out a problem we'd both seen wreck models in the City, known as look-ahead bias.

It happens when a model is accidentally shown information that wouldn't have existed when the decision was made.

In racing, the obvious example is the starting price, which nobody knows until the race is off, yet it sits right there in the old records.

Let a model peek at it during training and it looks brilliant in testing, then falls apart the first morning it's used for real.

So every record was time-stamped, and the model could only use what it would have known at 7am on the day.

It cost us weeks, and without it nothing that followed would have been worth a penny.


£56,093.00 JAN - JUL: £37,808.35 AUG - OCT: £18,284.65 £37,808.35 £18,284.65 JAN - JUL AUG - OCT

The early versions were only so-so


Better than guessing, but nowhere near good enough to survive a losing run.

So we changed the question.

Rather than picking over individual losers, we grouped thousands of runners by the chance the model gave them and checked how often each group won.

In the City that's called calibration, and no pricing model gets a tougher test.

If the model calls a group of horses one-in-five chances, about one in five should win, not one in three and not one in ten.

Wherever the results drifted from the model's numbers, it was misreading something, and we went looking for it.

What we found was humbling


It was ignoring how much the pace of a race affected hold-up horses, and it treated a soft-ground specialist the same on quick ground as in the mud.

It also got carried away by big wins in weak races and was far too harsh on narrow defeats in strong ones.

And it had no idea that some trainers bring a horse to form slowly while others have it ready first time out, which mattered hugely after a break.

The rail correction was Iain's idea, after a run of strange misreadings at two northern tracks that made no sense until he dug through the course reports.

Every fix was then tested on years of races the model had never seen, so we knew we were finding real patterns and not flattering ourselves with hindsight.

That matters more than anything else, because building a model that explains the past perfectly and then fails on Saturday is very easy.

The real danger is a pattern that isn't there


Give a computer enough numbers and it'll find a link between winning favourites and the colour of the trainer's cap, purely by chance.

Iain called these ghosts, patterns that look solid in old data and disappear once real money is on them.

We spent nearly as long getting rid of ghosts as building the model.

Every rule that survived had to prove itself again and again on unseen races before it got near a betting slip.

Iain had one more test


He'd ask if there was a sensible racing reason for a pattern, one a trainer or jockey would recognise if you explained it over a pint.

If there was, it stayed, and if the only explanation was the numbers, it was treated with deep suspicion until it had proved itself several times.

After hundreds of hours of that, the results started climbing, slowly at first and then with real conviction.

Once the improvement held through 3 months of live running, I finally believed we'd built something that worked.

A few more months of testing and tweaking, and it was ready...

I switched it on for real and the results started arriving...

£250 Weeks, Then £500 Weeks


Hardly any losing stretches worth mentioning...

Then my first full month in profit...

Bet Detector Pro picked 26 winners from 44 bets...

£2,725.30 Profit In 30 Days



£250 WEEK £250 WEEK £500 WEEK £500 WEEK £2,725.30 30 DAYS £2,725.30 30 DAYS

Early on, betting my own money, the model went 9 bets without a winner.

Every instinct I had told me to switch it off and go back to my own picks.

I didn't, because I'd seen that shape a hundred times in testing and knew it was bad luck running its course.

Over the next fortnight it landed 14 winners from 19 and wiped out the losing run.

That was when I stopped being a punter who owned some software and started trusting it like I'd trusted the models at work for thirty years.

It's the hardest lesson in betting, and my whole career had been preparing me for it.

Knowing a losing run is normal is one thing, and sitting through one with your own money at stake is another.

After that, the months fell into a rhythm


Quieter months still cleared three and a half to four thousand pounds, and strong ones went well past six.

A busy racing calendar gives the model more prices to work with, but the pattern holds whatever the time of year.

What never changed was the shape of it, more winners than losers, at prices big enough to leave a healthy profit after the losers were paid for.

I stopped celebrating individual winners long ago, because when the method is right, winners are what you expect.

And the winners kept coming, month after month...

I'd get my bets on first thing at the best odds going and follow the results on my phone through the afternoon.

For the past year I've been taking anywhere between £3,500 and well over £6,000 a month out of my betting.

How 2026 Has Gone


No table here, just the year as I'd describe it to an old colleague.

Every figure comes from my own betting accounts, recorded at the time.

No months missing and nothing left out.

2025 ended well, with November at £4,760.20 and December at £3,845.90, but this year is the one worth looking at.

January was the quietest month of 2026 and still made £3,690.55 profit.

February jumped to £5,120.30, and March beat it with £5,540.85.

April was the best month of the spring at £6,024.60.

May came in at £5,310.45, a terrific month by most standards and an ordinary one by this year's.


£0k £20k £40k £60k JAN: £3,690.55 JAN FEB: £8,810.85 FEB MAR: £14,351.70 MAR £14,351.70 APR: £20,376.30 APR MAY: £25,686.75 MAY JUN: £31,567.45 JUN JUL: £37,808.35 JUL £37,808.35 AUG: £43,236.20 AUG SEP: £49,342.65 SEP OCT: £56,093.00 OCT £56,093.00

July: Over £6,000 From 64 Winners In 96 Bets


June had been just as strong, with 61 winners from 92 bets returning £5,880.70.

The first seven months of the year added up to £37,808.35 in profit.

August kept things going with £5,427.85, and September went higher still at £6,106.45.

Then October Topped Them All At £6,750.35


October's £6,750.35 is the biggest monthly profit Bet Detector Pro has produced so far.

It took the total for the first ten months of 2026 to £56,093.00.

October catches a lot of punters out, because the Flat form starts to tail off before the winter form has settled.

That's when a model really helps, since it works from evidence while everyone else guesses which horses have gone over the top.

November has started the same way, with £1,204.50 profit in the first week and £1,375.40 in the second.

That's £2,579.90 in the first fortnight, and the third week is already under way.


NOV · £2,579.90 WEEK 1 WEEK 1: £1,204.50 £1,204.50 WEEK 2 WEEK 2: £1,375.40 £1,375.40

That's what a normal month looks like now


I haven't had a losing month in the past year...

A long way from the mediocre results I used to get before the software existed.

And it's never finished


Each completed race is added to the records the next morning, so the model always works from the latest form.

Once a month Iain and I go through what won and what lost, and whether any factor is starting to matter more or less as the season changes.

Above all, we check that the model's prices still match what happens on the track.

If horses it rates as one-in-four chances start winning noticeably more or less than one in four, something has shifted, and we dig until we find it.

It could be a course changing how it prepares the ground, or a big yard changing how it campaigns its young horses.

Catching it early in the numbers saves losing money on it later.

Ground changes through the year and yards go in and out of form, and a model that ignored that would slowly rot.

Keeping it right is like servicing a car, unglamorous work you only notice when someone skips it.

It's working better than ever


I put that down to hard work and plain stubbornness...

A lot of late nights spent finding ways to cut out bad bets...

And to find the horses with the best chance of winning.

No short-priced favourites, and no daft long shots either...

Profitable Picks, Day In, Day Out


These days the software sorts through the day's runners in no time at all...

So all I do is look over the selections in the morning...

Place my bets, and carry on with my day.

I've got more free time than I've had in years...

Money to spare, and the debts are gone...

Nothing hanging over me any more.

Then a few weeks ago, I had an idea...


Why not open it up to a small private group...

And share the selections with a few more people...

Two old colleagues from the desk had already been following my bets...

And my neighbour, Owen, had been looking over my shoulder at the selections for weeks...

All three were winning and kept telling me to let more people in...

In the end I gave in


I'd share Bet Detector Pro with a handful more people and see how it went...

On one condition, that nobody new joined until I'd seen it work for people beyond my two old colleagues and Owen.

So 10 test members were picked just over a week ago, and each agreed to report their results back to me...

Their results were better than I'd hoped, so I'm now opening 25 more places.

Three More From The Test Group


“My daughter set the whole thing up on my phone for me because I'm hopeless with all that, and now she rings every evening to ask what the horses did. We're £680 up in our first week and she's more excited than I am, which at seventy-one is really saying something.”

Graham Pickering, Harrogate

“What won me over wasn't the winners, it was watching how the software handled a bad day without throwing good money after it the very next morning. I finished the week £595 in profit off level stakes, and for the first time in years my betting feels like a plan rather than a bad habit.”

Denise Rowbotham, Wigan

“Twenty quid to find out if it was another con, that was my thinking, and I fully expected to be asking for my refund by the weekend. Instead I'm sat here £1,120 in front after nine days, and I've told the lad at work to keep his fiver because this one is the real thing.”

Trevor Hollingsworth, Kettering

So, Which Way Will You Go?


1) Carry on as you are, backing the same horses, following tipsters who vanish when the results turn, topping up your betting account from your wages, and wondering why nothing improves.

2) Spend 5 minutes joining Bet Detector Pro, get your selections before 8am, and start building a balance that grows week by week, with no form study, no guesswork and no chasing losses.

Bet Detector Pro has now delivered 15 profitable months in a row.

Tomorrow's selections will be worked out while you sleep, and the question is whether you'll be one of the 25 people receiving them.

Each day you put it off is a day of winning selections you'll miss.

Where things stand right now


I'm opening 25 more memberships...

I'm keeping numbers low while I build this up gradually...

Partly so the results stay as good for you as they've been for me...

Partly so I've got time to send the bets and answer your questions...

And so anyone who emails me gets a reply from me, not a helpdesk.

If you're reading this, you're in time...

But only just...

I Expect All 25 Places To Go Within 24 Hours


If you leave this page, I can't promise it'll still be here when you come back.

One of my partners sent you here because they thought you'd be a good fit...

And I'd like to prove them right...

Up To £3,387.50 Over Your First 7 Days


You'll be wondering about the price...

So let's get to it.

Money's tight for a lot of people at the moment...

So I've priced this well below what it's worth...

In my old job, firms paid eye-watering sums every year for access to pricing models, and nobody batted an eyelid.

I'd planned to charge £100 for membership...

And once this grows, that's very likely what it will cost...

But as one of the first people through the door...

You Can Join Today For £20


That's £20, paid once, with no monthly subscription...

No extra charges and nothing hidden in the small print...

Enough to cover my costs while I build this up.

That's less than many people spend on a single afternoon's racing.

And in case you're wondering about risk...

Your £20 Is Covered For 30 Days


Follow the selections for 30 days with your joining fee fully protected...

Write the results down, back them with small stakes, or go in big and make as much as you can...

If you're unhappy at any point, say so and I'll refund your fee in full straight away.

Nobody has asked yet, but the option is there.

What Happens After You Join?


Each morning an email arrives with the day's selections...

Sent moments after I've placed my own bets, so you get the same odds I do.

You place yours in a few minutes, and that's your part done.

Watch the winners come in on your phone, or sit down and enjoy the racing if you fancy it...

Or check your betting account in the evening to see how you've done.

It's that straightforward, and you don't need any previous experience.

Don't leave it too long


Places are going quickly...

I expect them to be gone well before the end of today.

When The 25 Places Are Taken, I Close The Doors


I won't open another intake for a good while, and it won't be £20 when I do.

Use the button below to claim your place:



Grab one of the 25 places before they're gone.

Here's what you get straight away...

Lifetime membership of Bet Detector Pro, paid for once.

Daily selections in your inbox before 8am.

2 to 3 picks a day at value odds, never short-priced favourites.

The same model behind 15 profitable months in a row.

No monthly fees, no hidden charges and no repeat payments, ever.

Direct email support whenever you have a question.

30 Days To Make Up Your Mind


30DAY

Follow the bets for up to 30 days, paper-trade them, use small stakes or back them in earnest.

If you're not satisfied at any point, email me and I'll return 100% of your joining fee the same day, with no questions and no delay.

Tap the button below to join today:



Questions People Ask Me


The 10 test members sent me a lot of questions in their first week.

Most of them came down to the same handful of worries.

So here they are, with the answers I gave.

Q: I know nothing about racing, will that matter?


A: Not at all, and most of the test group had never won a bean from betting before.

The selections arrive fully worked out, so you just place them as written.

Q: How long does it take each morning?


A: Under 5 minutes, from opening the email to placing the last bet.

Read the selections, put 2 or 3 bets on from your phone, and get on with your day.

Q: What's the most I could lose in a day?


A: Whatever you choose to stake on that morning's selections, and nothing more.

Each selection is a simple single bet, so nothing carries over or comes back to bite you later.

And your £20 fee is covered by the 30-day guarantee.

Q: How much should I start with?


A: As little as you're comfortable with, because stakes scale to suit any bank.

Many members begin small to find their feet, then raise their stakes once they've seen the winners come in.

Whatever you use, make it money you can afford to lose, kept apart from the money you live on.

Q: Which bookmakers should I use?


A: Any of the big high street or online bookmakers will do.

I always point you to the best price available, so it helps to have accounts with a few of them.

The same horse can be a point or more bigger with one firm than another, and over a year that adds up.

Q: How do I know the results will hold up?


A: That worried me more than anything while we were building it.

Every rule had to work on races the model had never seen, and it could only ever use information available on the morning.

Its prices are also checked against real results each month, and anything that starts slipping gets fixed before it costs members money.

Q: Who takes my payment?


A: ClickBank handles all payments, as one of the biggest online retailers of digital products.

Your card statement will show a charge from CLICKBANK or CLKBANK*COM, so don't be surprised when you see it.

Refunds under the guarantee go back the same way, through ClickBank.

Q: Will I be charged again later?


A: No, there are no further charges, now or later.

The joining fee is the only payment you'll ever make to me.

Q: Is there tax to pay on my winnings?


A: In the UK, betting winnings aren't taxed for ordinary punters, so there's nothing to declare.

I'm not an accountant though, so if your situation is unusual, have a quick word with one.

Q: Can I bet in a shop rather than online?


A: Yes, provided the shop is open in time and you don't mind taking the email in on your phone.

Most members find an app quicker, since the prices are right there and the bet's on in seconds.

Q: Who shouldn't join?


A: Anyone who has trouble stopping once they start betting.

If you find yourself chasing losses, this isn't for you, and I'd urge you to speak to GamCare, whose support is free and confidential.

Bet Detector Pro is for people who treat betting as a pastime with a plan, not as a way out of a hole.

Q: Why only £20 after 30 years in finance?


A: My income comes from my own betting, not from the £20.

I'd like a small group of members who follow the selections well and tell their friends, and a low price is the quickest way to find them.

Q: What if I'm not happy?


A: Email me within 30 days and I'll send back every penny of your fee straight away.

You stay in control of that decision throughout.

P.S. Thanks for reading to the end, it tells me you're serious about turning your betting around.

Tomorrow's selections go out before 8am, and with your fee covered by the guarantee, joining costs you nothing to try.

Claim one of the 25 places while there are still some left.

P.P.S. There are only 25 places in this intake, and they're going fast.

When they're gone, the price goes up


If you're still unsure, ask yourself this, what's the realistic downside of £20 with a full money-back guarantee?

There isn't one, and the upside could be the best decision you make all year.

One click below and you're in: