Staking Plans & Bankroll Management for Greyhound Punters

Betting Strategy · 7 min read · Updated October 2026

Two punters can pick the same winners for a year and one ends up broke while the other profits — the difference is almost always staking. Here are the mechanics of betting the right amount, from simplest to sharpest.

Start with a real bankroll

Your bankroll is money set aside exclusively for betting — an amount you could lose entirely without affecting rent, bills or mood. If that separation doesn't exist, nothing else in this guide applies. Most serious recreational punters run a bankroll of 50–100 times their average bet.

Flat staking: the baseline

Bet the same amount on every selection — say $10, or $5 on exotics. Flat staking is boring, which is exactly why it works: it makes your results purely a function of your selections, survives losing streaks without mathematical drama, and produces clean records for review. If you have no measured edge yet, flat staking is where you start.

Percentage staking: growing with the bankroll

Bet a fixed percentage of your current bankroll — commonly 1–2%. After winning runs your stakes naturally increase; losing runs shrink them, which is mathematically the correct behaviour (bet bigger when your bankroll — and confidence — supports it). Percentage staking never goes mathematically bankrupt, though a brutal run can reduce stakes to trivial levels.

The Kelly criterion: staking your edge

Kelly sizes bets by edge: stake = (edge × bankroll) ÷ (odds − 1), usually applied at a fractional level (half or quarter Kelly) to tame variance. Example: bankroll $1,000, you assess a dog at a true $3.00 chance ($4.00 fair) betting at $4.60 fixed — edge 15%. Full Kelly suggests roughly 4% of bankroll; quarter Kelly suggests 1%.

Kelly's catch: it requires honest probability estimates, and humans overestimate their edge. Used with discipline and fractional sizing, it is the mathematically optimal growth strategy; used with inflated confidence, it accelerates ruin. Only graduate to Kelly after months of recorded results prove your estimates are calibrated.

What to avoid

• Martingale (doubling after losses): the classic account-killer. A six-leg losing streak at even money requires 64 times your base stake to continue — and greyhound losing streaks of six are routine, not unlucky.
• Chasing within a session: raising stakes after a bad quaddie leg to 'recover'. This converts an ordinary losing night into a catastrophic one.
• Bet sizing by confidence feelings: your feeling of confidence is not calibrated. If it were, you wouldn't need the records.
• Staking more on exotics 'because the payout is bigger': the opposite is true — higher variance bets deserve *smaller* stakes for the same expected value.

Records make it real

Track date, track, race, dog, bet type, odds, stake, result and your reasoning. After 200 bets you will know your actual strike rate, average winning price, and where your edge lives — and you can then choose a staking plan that fits measured reality rather than hope. See [responsible greyhound punting](/guides/responsible-greyhound-punting) for the wider discipline framework, including when to stop entirely.

The hierarchy is simple: bankroll first, records second, staking plan third. Get those three right and even a modest edge compounds; get them wrong and even a good one doesn't matter.