Greyhound Betting Systems Explained: Dutch, Martingale and More

---

Author : Finlay Dennis Last updated: April 2026 Reading time : 21 min

Beyond Luck: How Betting Systems Actually Function

A betting system is discipline encoded—nothing more, nothing less. Strip away the marketing language, the promises of guaranteed returns, and the breathless testimonials, and what remains is a structured approach to two decisions every punter faces: what to back, and how much to stake. That structure is what separates the methodical bettor from the one tossing darts at a racecard.

Greyhound racing occupies a peculiar sweet spot for systematic betting. Six dogs, four bends, roughly thirty seconds of action. The field is small enough that meaningful analysis is practical, but large enough that the market regularly misprices runners. Compare that to horse racing, where fields of sixteen require exponentially more homework, or football, where a single deflection can render ninety minutes of tactical analysis irrelevant. Greyhounds offer frequency too—BAGS meetings alone produce hundreds of races daily across the UK, which means a system can be tested against live data faster than almost any other sport.

None of which means greyhound betting systems are easy money. They are not. What they provide is a framework for making repeatable decisions under conditions of uncertainty. A staking system tells you how to manage your money across a sequence of bets. A selection system tells you which dogs deserve your attention based on specific, measurable criteria. A hybrid system does both. The value of any system lies not in its complexity but in its consistency—whether you follow it when the results are running against you, not just when the winners flow freely.

This guide examines the systems most commonly applied to UK greyhound racing. Some have genuine merit. Some have elegant mathematics and terrible outcomes. Understanding the difference is worth more than any single bet you will ever place.

Categories of Greyhound Betting Systems

Know which lever you’re pulling before you pull it. Every greyhound betting system falls into one of three categories, and confusing them is the fastest route to frustration.

Staking systems concern themselves exclusively with how much you bet. They accept any selection—your gut feeling, a tipster’s pick, a pin stuck in the racecard—and impose a mathematical structure on the stake amounts. Martingale, Fibonacci, and level staking all live here. The appeal is obvious: you don’t need to know anything about greyhounds to follow a staking plan. The danger is equally obvious. No staking plan transforms bad selections into profitable ones. If you consistently back dogs at odds shorter than their true probability, no amount of clever stake manipulation will rescue your bank.

Selection systems occupy the opposite end. They define criteria for choosing which dogs to bet on—trap bias filters, sectional time thresholds, grade-drop indicators, trainer form metrics—and leave staking decisions to the bettor. These systems demand more work. You need data access, analytical tools (even a basic spreadsheet will do), and enough patience to resist betting on races that don’t meet your criteria. The reward is a set of selections that, over time, should carry a positive expected value. Should. The qualifier matters.

Hybrid systems combine both elements: defined selection rules paired with a structured staking approach. For example, a system might say: back any dog drawn in trap one at Romford when it has the fastest sectional time in the field, staking 2% of your current bank. This is where most successful long-term punters eventually arrive—not because hybrids are inherently superior, but because they force you to be explicit about every decision you make. Ambiguity is the enemy of profitable betting. A hybrid system leaves no room for it.

The Dutch Betting System In Depth

Dutching traces its name to Dutch Schultz, the Prohibition-era gangster whose accountant Otto Berman devised a method for spreading bets across multiple runners to guarantee a fixed return regardless of which one won. The maths has survived rather better than Schultz did.

The core principle is straightforward: instead of backing a single dog to win, you back two or more, adjusting the stake on each so that the total payout is identical whichever of your selections crosses the line first. You are not predicting a single winner. You are predicting that the winner will come from a defined group—and pricing that prediction accurately.

In greyhound racing, Dutching works best in competitive races where no single dog dominates the market. A six-dog race with an odds-on favourite is poor Dutching territory; the favourite consumes most of the probability, leaving scraps. But a race where three dogs sit between 3/1 and 5/1 is different. Here, the market is telling you the race is genuinely open, and a Dutch across two or three of those runners can produce a positive expected return if the combined implied probability of your selections exceeds the true probability that one of them wins.

The formula for calculating individual stakes in a Dutch is: divide the target profit by the decimal odds of each selection, then adjust proportionally so the total outlay matches your budget. Suppose you want to Dutch two dogs at decimal odds of 4.0 and 5.0 with a total stake of £20. The proportion for Dog A is 1/4.0 = 0.25; for Dog B, 1/5.0 = 0.20. The total is 0.45. Dog A receives (0.25/0.45) × £20 = £11.11, and Dog B receives (0.20/0.45) × £20 = £8.89. If Dog A wins, the return is £11.11 × 4.0 = £44.44; if Dog B wins, £8.89 × 5.0 = £44.45. Equal return, either way.

The critical question is whether the combined implied probability of your selections leaves room for profit after the bookmaker’s margin. If the overround is steep—common in greyhound markets—Dutching across too many runners becomes unprofitable by definition. Two or three carefully chosen dogs in a competitive race is the sweet spot. Trying to Dutch four or five runners almost always means paying the bookmaker for the privilege of spreading your risk.

Dutching does not pick winners. It hedges against picking losers from a defined shortlist. That distinction matters more than most punters appreciate.

Calculating Dutch Stakes

Arithmetic before ambition. The maths behind Dutching is not complicated, but getting it wrong defeats the entire purpose. Here is the process stripped to essentials.

Step one: convert all odds to decimal format. A dog at 7/2 becomes 4.50; a dog at 3/1 becomes 4.00; a dog at 9/2 becomes 5.50. Step two: calculate the inverse of each selection’s decimal odds. For our three dogs, that gives us 0.222, 0.250, and 0.182. Step three: sum those inverses—0.654 in this case. If that sum exceeds 1.0, the combined implied probability of your selections is over 100% and you cannot profit from Dutching them at those prices. Step four: divide each inverse by the total and multiply by your total stake. With a £30 budget, Dog A gets (0.222/0.654) × £30 = £10.18; Dog B gets £11.47; Dog C gets £8.35. The return if any one wins sits around £45.87.

Several online tools automate this entirely. The Timeform Dutch calculator and OddsMonkey’s version are both reliable and free. For punters who prefer working in spreadsheets, a simple formula column handles the maths in seconds. The manual method is worth learning regardless—not because calculators fail, but because understanding the maths reveals when a Dutch is marginal and when it is strong. A combined implied probability of 60% on three dogs in a six-runner race looks appealing. The same probability spread across four dogs at shorter prices looks like paying for insurance you don’t need.

One refinement worth noting: reduced-stake Dutching. Instead of equalising the return across all selections, you weight more heavily toward the dog you consider likeliest. The payout is no longer flat, but the total outlay drops. This hybrid approach trades some of Dutching’s mathematical purity for a more opinion-driven allocation. It suits punters who have a view on the race but want partial protection if they are wrong.

The Martingale System: Why It Fails

Martingale is seductive—and mathematically doomed. The pitch sounds irresistible: double your stake after every loss, so that the first win recovers everything plus one unit of profit. It works on paper. It works in practice too, right up until the moment it catastrophically doesn’t.

The mechanics are simple. You stake £5 on a dog at even money. It loses. You stake £10 on the next selection. Loses again. £20. £40. £80. £160. After six consecutive losses—which happens more often than most people intuit—you have invested £315 to chase a £5 profit. If the seventh bet also loses, the cumulative damage reaches £635. Greyhound racing produces losing runs of six, seven, eight or more regularly, even for selections with a 40-50% strike rate. The Martingale system does not account for reality; it assumes infinite capital and no betting limits, neither of which exist.

The psychological trap is what makes Martingale dangerous rather than merely inefficient. Each escalating stake feels justified because the logic of eventual recovery seems sound. Punters deep in a losing sequence are not thinking about the exponential growth curve of their liability. They are thinking about the relief of the next winner. This is exactly the emotional state in which catastrophic financial decisions get made.

Bookmakers, incidentally, love Martingale players. The escalating stakes generate turnover, and the system’s inevitable collapse events—where a long losing run wipes out weeks of accumulated small profits—ensure the house always wins over a sufficient sample. Some betting firms have even been known to restrict accounts less aggressively when they identify Martingale patterns, because the long-term mathematics so reliably favour the operator.

The numbers do not care how reasonable the logic sounds. A system that requires exponentially increasing risk to produce linear returns is a system designed to fail. The only question is when, not whether.

Progressive Staking Alternatives

Slower failure is still failure—but the pace matters, because slower systems at least give you time to recognise the problem before your bank evaporates.

The Fibonacci system increases stakes following the famous sequence: 1, 1, 2, 3, 5, 8, 13, 21. After a loss, you move one step forward in the sequence. After a win, you step back two places. The escalation is gentler than Martingale’s doubling, which means a losing run of eight produces total exposure of roughly £54 rather than £1,275 on a £5 base. The relief is relative, though. Fibonacci still requires progressively larger stakes during losing periods, and a long enough losing run will overwhelm any realistic bankroll. It simply takes longer to get there.

The Labouchere system (sometimes called the cancellation method) lets you define a target profit and work toward it through a sequence of calculated stakes. You write a number sequence—say, 1-2-3-4—and your stake is always the sum of the first and last numbers. Win, and you cancel those two numbers. Lose, and you add the lost amount to the end. It offers more flexibility than Fibonacci and feels more controlled, but the tail risk is identical: a persistent losing sequence extends the number list and inflates the stakes.

D’Alembert takes the most conservative approach of the progressive family, increasing stakes by a single unit after a loss and decreasing by one after a win. It is the least aggressive escalation method and, consequently, the least likely to produce a dramatic blowout. It is also the least likely to recover losses quickly, which means it requires patience and a genuinely positive selection strike rate to generate profit. If your selections are running at 30% when you need 35%, D’Alembert will not paper over the gap. It will just make the decline more gradual.

All progressive systems share a fundamental flaw: they assume the next bet exists in isolation from the sequence, when in reality the sequence itself creates pressure that distorts decision-making. The best thing any progressive system can teach you is that if your selections are profitable at level stakes, you do not need progression at all.

Statistical Selection Systems

Data doesn’t guarantee wins—but ignorance guarantees losses. Statistical selection systems flip the emphasis from how you stake to what you back, using measurable criteria to identify dogs that the market has undervalued.

The simplest statistical approach exploits trap bias. Every UK greyhound track has structural characteristics—track geometry, distance from trap to first bend, hare rail position—that favour certain starting positions over others. In a perfectly fair six-dog race, each trap would win approximately 16.7% of the time. In practice, some traps at some tracks exceed 25% over large samples. That gap between expected and actual win rate is your edge, and it exists because most casual punters ignore it entirely. Tracking trap performance across a few hundred races at your chosen track costs nothing except time and a spreadsheet.

Sectional time analysis operates at a finer resolution. Most UK tracks record split times—typically at the first bend and sometimes at intermediate points—that reveal a dog’s pace profile. A dog with fast early splits but fading finish times is a front-runner who needs a clean break. A dog with moderate early pace but strong closing splits is a closer who benefits from interference ahead. By comparing sectional profiles against the trap draw and the likely running styles of opponents, you can identify dogs whose pace profile gives them a structural advantage in a specific race. This requires access to form data from sources like the Racing Post or the Greyhound Board of Great Britain’s records, and it requires patience to build a meaningful dataset.

Grade-drop systems focus on dogs that have recently been moved down a grade. A dog dropping from A3 to A4 is competing against theoretically weaker opposition, and while grade drops sometimes signal declining form, they can also indicate a dog returning from rest, recovering from a poor draw, or simply being managed by a trainer who knows the grading cycle. Filtering grade drops by additional criteria—recent sectional times, track familiarity, trainer strike rate—sharpens the selection considerably.

Building your own statistical model does not require advanced mathematics. A spreadsheet with columns for trap, grade, sectional time, running style, and result, accumulated over a few weeks of racing at one or two tracks, will start revealing patterns that casual observation misses. The GBGB publishes race results with full form data, and specialist sites aggregate this into searchable databases. The investment is time, not money. The return is a selection process grounded in evidence rather than instinct—which is exactly what separates systematic bettors from recreational ones.

Following Smart Money

Follow the money, not the hype. Market movements in greyhound racing carry information, and learning to read them gives you a window into what sharper bettors are doing.

A “steam move” occurs when a dog’s price shortens rapidly across multiple bookmakers in a short period. Oddschecker’s price history feature makes these movements visible in near real-time. When a dog drifts from 5/1 to 7/2 to 3/1 across four or five firms within minutes, someone with confidence—and usually money—is backing it. Professional bettors, kennel connections, or those with access to trial information can move a greyhound market more dramatically than equivalent money moves a horse racing market, simply because greyhound betting pools are smaller.

Distinguishing professional money from public money is part art, part timing. Public money tends to arrive late, close to race time, driven by tipsters and social media. Professional money typically appears earlier, often in the morning markets, and moves prices before the general public even notices the race. A dog that opens at 6/1 and drifts to 8/1 by early afternoon before suddenly contracting to 4/1 in the final thirty minutes is showing a pattern consistent with professional interest arriving after early casual money pushed the price out.

A word of caution: following market movements is not a standalone system. Steam moves occur in races with thin liquidity, and sometimes the money behind them is wrong. It works best as a confirming signal layered on top of your own analysis. If your statistical model highlights a dog and the market then moves in its favour, that convergence of independent signals strengthens the case. If the market moves against your pick, it doesn’t automatically invalidate your analysis—but it should prompt a second look before you commit your stake.

Level Stakes: The Boring Winner

Boring protects the bank. Exciting empties it. Level staking—placing the same amount on every qualifying bet regardless of confidence, odds, or recent results—is the least glamorous approach to bankroll management and, for most punters, the most effective.

The case for level stakes rests on one brutal truth: human beings are terrible at calibrating confidence. We overbet when we feel certain and underbet when doubt creeps in, and our feelings bear almost no statistical relationship to actual outcomes. Level staking removes this variable entirely. Every bet carries equal weight, which means your results over time reflect the quality of your selections rather than the quality of your emotional state. It is, in effect, the control experiment of betting—isolating the one variable that matters.

A standard approach is to stake 1-2% of your starting bank per bet. On a £500 bank, that means £5-£10 per selection. The amount feels small, and that is precisely the point. Small stakes absorb losing runs without threatening the bank. A fifteen-bet losing streak at £10 a time costs £150—painful, but survivable. The same losing streak under a Martingale system starting at £10 would cost over £300,000 in theoretical escalation. Level staking does not eliminate losing streaks; it prevents them from being terminal.

Where level staking truly proves its value is in performance measurement. If you are flat staking and your records show a positive return on investment after 200+ bets, you know your selection method works. If the ROI is negative, you know it does not. No ambiguity, no noise from variable staking obscuring the signal. This clarity is worth more than any progressive system’s promise of accelerated profits, because it tells you the truth about your betting ability—and the truth is the only foundation on which a sustainable approach can be built.

Building Your Own Hybrid System

The best system is the one you understand and execute consistently. That sounds like motivational poster material, but it contains a genuine insight: a mediocre system followed with discipline will outperform a brilliant system followed erratically. Building your own hybrid approach forces you to define every element explicitly, which makes consistent execution far more likely.

Start with selection criteria. Pick two or three measurable factors from the statistical methods discussed earlier—trap bias at your chosen track, sectional time advantage, grade drop within the last two races—and define precise thresholds. Not “the dog looks quick” but “first-bend split time within 0.05 seconds of the fastest in the field.” Not “favourable trap” but “trap with a win rate above 20% over the last 200 races at this track.” The more specific your criteria, the less room there is for subjective interpretation to creep in on a bad day.

Layer in a staking rule. Level staking at 1-2% of your current bank is the simplest and most robust option. If you want to add a single variable, consider a two-tier system: standard stake when one criterion is met, double stake when two or more criteria align. Keep it simple. Every additional rule adds a decision point, and decision points introduce opportunities for error.

Paper-test before committing real money. Track your system’s selections for at least fifty races without placing actual bets. Record every qualifying selection, the price available, and the result. This phase is not optional—it is where you discover whether your criteria produce enough qualifying bets to be practical, whether the strike rate supports profitability, and whether the system generates selections at odds that offer genuine value. A system that qualifies one bet per week is not usable. A system that qualifies thirty bets a day is probably not selective enough.

After the paper trial, review and adjust. If the strike rate is close to profitable but not quite there, tightening one criterion may be enough. If the system is wildly unprofitable, revisit your assumptions rather than patching with staking adjustments. Good selection criteria occasionally need refinement. Bad selection criteria need replacing. The difference is usually visible in the data within fifty to a hundred bets—which, given the volume of UK greyhound racing, can represent as little as two or three weeks of results.

Systems Don’t Bet—You Do

A system on paper is worthless. A system in practice is everything. The gap between the two is filled entirely by the person holding the pen—or, more accurately, clicking the mouse.

Every system discussed here has theoretical merit in some form. Dutching spreads risk intelligently. Statistical selection grounds decisions in evidence. Level staking protects capital. But none of these approaches functions on autopilot. A Dutching system requires you to assess race competitiveness honestly, not hopefully. A statistical model requires you to trust the data when it contradicts your instinct. Level staking requires you to resist the urge to increase stakes after a good run or chase losses after a bad one. The system provides the framework. You provide the discipline.

Set your expectations at ground level. A long-term return on investment of 5-10% at level stakes is an excellent result—achievable by skilled, disciplined bettors and essentially unreachable by those who jump between systems, abandon methods after short losing runs, or treat every race as a must-bet event. The edge in greyhound betting is real but narrow. It rewards patience, record-keeping, and a willingness to sit out races that do not meet your criteria rather than forcing a bet because the next race is only twelve minutes away.

Greyhound racing will still produce upsets, interference, and results that no system could have predicted. That is the nature of the sport—six live animals in a confined space moving at forty miles per hour. The role of a system is not to eliminate uncertainty. It is to ensure that when uncertainty resolves in your favour, you are positioned to benefit, and when it resolves against you, the damage is contained. Over hundreds and thousands of bets, that asymmetry compounds. And compounding, not any single winner, is where the real edge lies.