You can have a genuine edge and still go broke. Bankroll management is the discipline that ensures your edge survives long enough to produce returns. This guide covers the principles that professional bettors actually use.
Read the Guide →
Most bettors who fail long-term don't fail because they have no edge. They fail because they don't have the capital structure or the staking discipline to survive the variance that comes with any betting edge. A 5% positive expected value approach, run on an under-capitalised bankroll with poor staking, will lose money through ruin before the edge has time to manifest.
This is the ruin problem : mathematically certain failure for undercapitalised operations regardless of edge. Bankroll management is the set of disciplines that prevents it.
This guide covers the principles: what a betting bankroll is, how to size it, how to structure stakes, how to handle drawdown, and how to organise capital across multiple betting operations and platforms.
A betting bankroll is a dedicated, ring-fenced pot of money set aside exclusively for betting. The key characteristics:
Many bettors confuse "having money in a bookmaker account" with having a bankroll. A bankroll is a structured, deliberately sized allocation with defined staking rules applied to it. Without the structure, it's just money in an account.
Bankroll size should be determined by your staking plan, not the other way around. The question to answer is: how large must the bankroll be for my staking approach to survive expected drawdown without requiring a rebuild?
For a flat staking approach at 1% per bet, a 100-unit drawdown (a plausible run for a bettor with a real edge in a tough period) represents a 50% drawdown from an initial 100-unit bankroll, painful but survivable. At 2% per bet, the same drawdown wipes the bankroll.
| Staking Level | Bankroll Needed for 100-Unit Drawdown Survival | Recovery Time Estimate | Variance Level |
|---|---|---|---|
| 0.5% per bet | 200 units | Slow but very stable | Very low |
| 1% per bet (standard) | 100 units | Moderate | Low–medium |
| 2% per bet | 50 units | Fast but volatile | High |
| Full Kelly | Varies by edge | Theoretically optimal | Extreme : not recommended |
In practical terms, most professional bettors operating at volume in sharp markets use 1–2% flat staking and a starting bankroll equivalent to 100–200 units. A bettor using €50 per unit would need a €5,000–€10,000 bankroll to operate safely at this level.
The two most widely used staking approaches are flat staking and Kelly Criterion. Both have genuine advantages. The right choice depends on your confidence in your edge estimates and your operational discipline.
Flat staking means betting the same amount (or the same percentage of your current bankroll) on every bet regardless of your estimated edge. This approach is simple, transparent, and robust to errors in edge estimation. If your model overestimates the edge on a given bet, you're not over-exposed.
The limitation is that flat staking doesn't differentiate between high-confidence and low-confidence bets. A bet you believe is 5% EV gets the same size as one you believe is 15% EV. Some edge is left on the table.
Kelly calculates the theoretically optimal stake size for each bet based on your estimated edge and the odds. The formula:
Example: You estimate 55% probability, odds are 2.00. b = 1, p = 0.55, q = 0.45. Kelly = (1 × 0.55 − 0.45) ÷ 1 = 10% of bankroll.
Full Kelly is theoretically optimal for bankroll growth but produces extreme variance. A single mis-estimated edge can lead to severe over-exposure. Most professionals use fractional Kelly (typically 25–33% of the full Kelly amount) which reduces variance substantially while retaining most of the growth advantage.
Drawdown is the reduction in your bankroll from its peak to a subsequent trough. Every bettor experiences drawdown, even bettors with strong, validated edges. The question is whether your operation is structured to survive it.
| Drawdown Level | What It Means | Appropriate Response |
|---|---|---|
| 0–10 units | Normal variance : expected for any bettor | Continue as normal. Do not change approach. |
| 10–25 units | Significant but within statistical range for most edges | Review recent bets for process errors, not results. Continue if process is clean. |
| 25–50 units | Serious drawdown : possible edge degradation | Reduce stakes by 25–50%. Investigate whether market conditions have changed. |
| 50+ units | Potential structural problem | Pause. Conduct systematic review of all bets from the drawdown period. Seek external review if possible. |
The most damaging response to drawdown is to increase stakes to recover faster. This is "loss chasing" at a structural level : it turns manageable drawdowns into terminal ones. The correct response is almost always to reduce stakes, not increase them.
It's also important to distinguish between a drawdown caused by bad variance and one caused by an actual problem. If your closing line value remains positive during a losing run, the run is variance. If CLV has also turned negative, something structural has changed.
Most serious bettors operate across multiple platforms, at minimum a broker account for sharp book access and an exchange for trading and lay betting. Some run two broker accounts to maximise market access.
Capital allocation across platforms should reflect how actively each is used:
Keep a small liquidity reserve (10–15% of total capital), not deployed in any account. This ensures you can fund an opportunity without waiting for a withdrawal to process.
The platform structure you operate through has direct implications for bankroll management. Soft bookmakers create a unique bankroll management problem: the account can be restricted or closed at any point, effectively removing capital from the operation without warning.
Professional bettors avoid concentrating capital in platforms that can arbitrarily restrict accounts. The broker model solves this : broker accounts don't experience the restriction cycle, and capital held in a broker account is stable. You're not at risk of waking up to find a €5,000 balance frozen in a gubbed account.
This makes the broker model structurally better for bankroll management, independent of the odds quality argument. A stable capital base that you control is a prerequisite for executing any bankroll management plan over the long term.
For bettors currently operating through soft bookmakers who are experiencing restrictions, transitioning to a broker account solves both the access problem and the capital stability problem simultaneously. See our best betting brokers guide for a comparison of the major options.
For flat staking, 1–3% of your total bankroll per bet is the standard range for most betting operations. For Kelly-based staking, most professionals use fractional Kelly at 20–33% of the full Kelly stake. The exact figure depends on your confidence in your edge estimate and your tolerance for drawdown. Erring on the smaller side is nearly always correct : losing a betting operation to poor bankroll management is more common than leaving edge on the table due to conservative staking.
Kelly Criterion is a mathematical formula for calculating optimal bet size based on your estimated edge and the odds available. Full Kelly maximises long-run growth rate but produces large, volatile swings. Most professionals use fractional Kelly (20–33% of the full Kelly amount) which reduces variance while preserving most of the growth advantage. The key limitation: Kelly requires an accurate edge estimate, which is difficult to produce reliably. If your edge estimate is too high, Kelly leads to overbetting.
A properly sized bankroll should be able to survive at least 50–100 units of drawdown without requiring a rebuild. For a 1% flat staker, this means the bankroll can withstand 50–100 losing bets before dropping below a functioning level. In practice, drawdowns of 20–30 units are common for bettors with genuine edges in any given season. The key is sizing the starting bankroll so these expected drawdowns don't force you to stop betting.
Yes. Ring-fencing bankrolls by activity (value betting, exchange trading, each-way racing, Asian Handicap) allows you to track performance separately and size stakes appropriately for each context. Mixing betting activities into a single pool obscures which activity is producing returns and which is losing. Most professionals treat each betting approach as a separate operation with its own capital allocation.
A broker account typically holds your primary working capital for sharp market betting. Because broker accounts aren't subject to restriction or closure cycles, they provide a stable capital base. You don't need to manage the risk of losing a funded account to a bookmaker decision. The working balance should reflect 2–4 weeks of normal betting activity, enough to operate without constant top-ups, but not excess capital sitting idle.
The standard professional response to a losing run is to reduce stakes by 25–50% until performance returns to within expected variance bands, not to increase stakes to recover losses faster. Review whether the losing run is within normal statistical variance for your edge (most are) or whether something has changed in your approach or market conditions. Only restructure the approach if you have concrete evidence of a systematic problem, not because of a short-term result.