The Uncomfortable Truth About Winning Tips
Here's something that might surprise you: you can follow a tipster who picks winners 30% of the time and still lose money. Conversely, you can back horses that win only 15% of the time and make a consistent profit.
How? It all comes down to price.
If you back $1.80 favourites that win 30% of the time, you're losing 46 cents for every dollar wagered. But if you back $8 shots that win 15% of the time, you're making 20 cents for every dollar. The first tipster gives you winners to brag about at the pub. The second one quietly builds a bankroll.
This is the core principle behind value betting, and it's the foundation of everything we do at Racin. We don't optimise for winners — we optimise for expected value.
What Exactly Is a Value Bet?
A value bet exists when the true probability of a horse winning is higher than the implied probability from the bookmaker's odds. In plain English: the bookmaker is offering better odds than they should.
Let's make this concrete. Say our model analyses a race and determines Horse A has a genuine 25% chance of winning. We look at the best available odds across bookmakers and see $6.00.
At $6.00, the implied probability is 1/6 = 16.7%. But we think the true chance is 25%. That's an overlay of nearly 50%.
The overlay formula is simple:
Overlay = (Model Probability - Implied Probability) / Implied Probability
In this case: (25% - 16.7%) / 16.7% = 49.7% overlay
At Racin, we flag a runner as a value bet when the overlay exceeds 20%. This threshold exists as a buffer — the model isn't perfect, and we want a meaningful edge before committing money.
Why Bookmaker Odds Don't Reflect True Probability
A lot of punters treat bookmaker odds like gospel. "The bookie has it at $3 so it must have a 33% chance." Not exactly.
Bookmaker odds are distorted by several factors:
The overround. In a perfectly fair market, the sum of all implied probabilities would be exactly 100%. In reality, Australian bookmakers build in a 15-25% margin. Every horse is slightly underpriced to guarantee the bookie a profit regardless of the result. In a 12-horse race, the implied probabilities might sum to 120% — that extra 20% is the bookmaker's edge.
Public money bias. Bookmakers shade their odds based on where the public bets. Popular horses — last-start winners, horses with big names, horses that "look the goods" — often get more public money than they deserve. This shortens their price below fair value and, crucially, pushes other runners' odds out beyond fair value. That's where the opportunities live.
Risk management. If a bookmaker has taken a huge bet on one horse, they'll shorten that horse's price and push others out — not because the other horses are less likely to win, but because the bookie needs to balance their book. This creates artificial value.
Information asymmetry. The market is generally efficient for big metro races where thousands of sharp punters are analysing every angle. But in mid-week country races? Wednesday at Dubbo? The market is thinner, less efficient, and the model often finds bigger edges.
Our Backtest Results: What Value Betting Actually Returns
Theory is nice. Let's talk numbers.
We backtested our value bet detection system across 2,798 races from July 2025 to March 2026. Every time the model identified a 20%+ overlay, we tracked what would happen if you bet on it. Here's the month-by-month breakdown:
The good months:
The tough months:
The bottom line: +$429.60 cumulative value bet P&L, or 24.1% ROI.
Here's what I want you to notice: even in losing months, the drawdowns were manageable. October's -$23.60 hurt, but it didn't wipe out the previous three months of gains. And November came roaring back with +$172. That's what a genuine edge looks like — it's lumpy, it's not a straight line, but it trends upward.
The Hardest Part: Accepting That Good Bets Lose
This is where most punters fail with value betting. Let me say it plainly:
A horse with a genuine 25% chance of winning will lose 75% of the time.
That means if you back four value bets in an afternoon, you should expect three of them to lose. After a Saturday with one winner from six value flags, it's tempting to think the system doesn't work. But if that one winner was at $7 and you staked $10 on each of the six, you've wagered $60 and collected $70. A profitable day — despite a 17% strike rate.
Value betting requires a fundamentally different mindset from traditional punting. You're not trying to feel clever by picking winners. You're trying to make mathematically sound decisions consistently over a long timeframe.
I've seen punters try value betting for two weeks, hit a losing streak, and abandon it. Then three months later they check back and see the system has been up 20%. The maths doesn't care about your feelings. It just needs time and sample size to work.
Common Mistakes When Value Betting
Staking Too Much on Individual Bets
Even genuine value bets lose most of the time. If you put 10% of your bankroll on every value flag, a run of 6 losses (which will absolutely happen) puts you down 60%. Game over.
Professional value bettors typically stake 1-3% of their bankroll per bet. Our Elite tier includes Kelly criterion stake sizing that calculates the mathematically optimal bet size based on the edge and your bankroll. It's usually smaller than you'd think. The Kelly formula is aggressive but it prevents ruin.
Cherry-Picking Value Bets
"I'll only take the value bets that also look good on form." This defeats the purpose. If you filter value bets through your own subjective opinion, you're removing half the edge. The model identifies value by processing 79 features — more than any human can hold in their head simultaneously. Trust the system or don't use it, but don't half-use it.
Chasing Losses After a Bad Day
Friday's losses don't make Saturday's value bets any better or worse. Each bet is independent. Increasing your stakes after a loss (known as the Martingale approach) is mathematically proven to end in ruin. Flat stakes or Kelly-adjusted stakes. Every time. No exceptions.
Only Betting at Short Prices
Value exists across the entire odds spectrum. A horse at $2.50 that should be $2.00 is a value bet. A horse at $15 that should be $10 is a value bet. Some punters only back value flags at $5+ because the payoffs feel better. But value at short prices compounds just as effectively — and with lower variance.
How Racin Automates Value Detection
Finding value manually is hard. You'd need to:
1. Form your own probability estimate for every runner in every race
2. Convert current bookmaker odds to implied probabilities
3. Compare across multiple bookmakers to find the best price
4. Calculate the overlay percentage
5. Decide if the edge is large enough to bet
6. Size your stake appropriately
Most punters do step 1 (badly) and maybe step 2. Nobody consistently does all six across 50+ meetings per week.
Our model does all of this automatically. For every runner in every race:
When you open a race card on Racin, value bets are highlighted with the model probability, market implied probability, and overlay percentage. You see exactly why it's flagged and can make your own decision.
Getting Started Without Going Broke
If you're new to value betting, here's my honest advice:
Start with paper betting. Track the value flags for two weeks without risking real money. Log every flag, note the odds, and mark the results. At the end of two weeks, calculate your P&L. This builds confidence in the approach without financial risk.
Use the free tier first. Racin's free plan covers Saturday metro meetings with top 3 picks and value flags. That's typically 3-4 meetings per Saturday — enough to see the model in action across 25-30 races per week.
When you start betting real money, stake small. 1-2% of whatever bankroll you've set aside. If your racing bankroll is $500, that's $5-$10 per bet. It feels small. It's supposed to. Value betting is a marathon, not a sprint.
Track everything. Keep a spreadsheet. Record every bet, the odds, the stake, and the result. After 200+ bets, you'll have a statistically meaningful sample. That's when you can honestly evaluate whether the approach works for you.
The maths is on your side. You just need the patience to let it play out.