Bankroll Management Systems Compared: Flat, Kelly, and Hybrids

• Why bankroll management matters
• Systems at a glance
• Flat betting explained
• Kelly Criterion explained
• Hybrid systems used in real life
• Which system fits you?
• Worked examples and risk
• Tools, calculators, and tracking
• Responsible gambling
• Key takeaways
• FAQs
• Sources and further reading

Why bankroll management matters

Your bankroll is your betting money. A system is the rule for how big each bet is. Good rules help you last through bad runs. They also help you grow when you win. This is key for long-term play.

Three ideas matter a lot:

• Variance: Results jump up and down by chance. Even good picks can lose many in a row. See simple ideas about chance here: Khan Academy: Probability basics.
• Drawdown: How far your bankroll drops from a high point to a low point.
• Risk of ruin: The chance your bankroll hits zero (or a stop level) before you recover. Read more: Gambler’s Ruin (Wikipedia).

Big bet sizes can grow fast, but swings are hard. Small bet sizes are safer, but slower. A good system finds a balance you can stick to.

Systems at a glance

• Flat Betting: You bet the same unit each time (for example 1% of bankroll). Pros: simple, steady, low stress. Cons: slow growth, even if you have an edge.
• Kelly Criterion: You bet more when your edge is big, less when it is small. Pros: best long-run growth in theory. Cons: can be very swingy; needs a good edge estimate. Learn the idea: Kelly Criterion (Wikipedia), and Investopedia: Kelly Criterion.
• Hybrids: Mix of the two. For example, half-Kelly, or a percent-of-bankroll cap, or simple tiers like 0.5u/1u/1.5u. Pros: smoother ride. Cons: more rules to follow.

Flat betting explained

What it is: You pick one unit size and keep it. A unit is often 0.5% to 2% of your bankroll. For example, if your bankroll is $2,000 and you choose 1%, your unit is $20. You bet $20 each time.

When it fits: You bet for fun. You are not sure about your edge. You want to protect your money. You want simple rules you can follow even on a bad day.

Pros:

• Very easy to do. No math each time.
• Stops tilt. You do not “chase” with bigger bets.
• Good for learning and record keeping.

Cons:

• If you do have a real edge, growth may be slower than it could be.
• No compounding per bet if you keep unit fixed in dollars (you can also set unit as a percent to compound slowly).

Quick example: Say odds are -110 (decimal 1.91). If you win 55% of bets, flat 1% per bet will likely grow over time. If you win near 50%, flat bets will lose slowly due to the vig. The key is that the stake size stays steady, so variance is easier to handle.

Kelly Criterion explained

What it is: Kelly tells you what fraction of bankroll to bet based on your edge. The simple formula for one bet is:

f = (b × p − q) / b

• b is decimal odds minus 1 (for -110, decimal is ~1.91, so b = 0.91)
• p is your chance to win (0 to 1)
• q is 1 − p

Why people like it: If your edge and odds are correct, Kelly can maximize long-run growth of your bankroll. This is known from work by Edward O. Thorp and others. See MacLean, Thorp, Ziemba (SSRN) and Edward O. Thorp’s archive. A clear primer: Pinnacle: Kelly Criterion explained.

Real-world problems:

• It is hard to know your true p. If you guess too high, you overbet. Overbetting can blow up a bankroll fast.
• Drawdowns can be big. You must be okay with swings.
• Markets move, books limit stakes, and you may not get your exact price.

Fractional Kelly: Many use half-Kelly or quarter-Kelly to cut swings. You just take the Kelly stake and use 50% or 25% of it.

Worked example:

• Odds: -110 (decimal ~1.91) so b = 0.91
• Your model says p = 0.55, so q = 0.45
• f = (0.91 × 0.55 − 0.45) / 0.91 = (0.5005 − 0.45) / 0.91 ≈ 0.0555
• So full Kelly says bet about 5.6% of bankroll on this edge.
• Half-Kelly would be about 2.8%.

Guardrails to add:

• Set a per-bet cap (for example, never over 2% or 3% of bankroll).
• Recompute stakes only at set times (daily or weekly), not after each win/loss, to reduce noise.
• Round to units you can track (for example, round to the nearest 0.5% or a dollar step).

Hybrid systems used in real life

Many folks mix ideas to get the best of both worlds. Here are common hybrids:

• Fractional Kelly with caps: Use 25%–50% Kelly, but never more than 2% per bet. This cuts drawdowns a lot.
• Percent-of-bankroll flat: Bet a fixed percent like 1% on each pick, no matter the edge. This compounds slowly and is simple. It does not need a model for p.
• Tiered units: Set clear tiers, like 0.5u, 1u, 1.5u. Tie each tier to rules you write down. For example, 1.5u only if your model shows 3%+ edge and price beats the closing line on average. Learn about closing line value (CLV): What is CLV?
• Kelly–Flat blend: Compute fractional Kelly, then stake the minimum of that and your flat cap. Or round to the nearest unit.

Avoid chasing systems: Martingale and “double after a loss” can wipe you out. They fight variance with bigger and bigger bets. That boosts risk of ruin. Read why here: Martingale (Wikipedia).

Write your rules: Put your rules on paper. Include your unit, caps, when to adjust, and what to do after a bad run. This helps you stay calm and avoid tilt.

Head-to-head: Which system fits you?

Use these points to choose:

• Edge certainty: If your edge is unclear or small, use flat or small percent. If you model well and test results, use fractional Kelly with caps.
• Variance tolerance: If swings stress you out, go small and simple. If you can handle swings, you can use a bit more risk.
• Bet frequency and limits: Many small bets can smooth swings. Book limits can also stop big Kelly stakes.
• Complexity: Flat is easy. Kelly needs math and good inputs. Hybrids sit in the middle.
• Bankroll size and reloads: If you cannot reload, protect it. If you can reload, you might accept more variance.

Quick picks:

• New or casual: Flat 0.5%–1% or a simple 1% percent-of-bankroll plan.
• Model-driven: Fractional Kelly (25%–50%) with a 2% cap per bet.
• Mixed confidence: Tiered units with written rules and a hard max stake.

Worked examples and risk

Scenario A: 100 bets at -110, p = 52%

Assume you place 100 bets. Your true win rate is 52%. Odds are -110 (1.91). This is a small edge.

• Flat 1% per bet: Stakes are steady. Bankroll drift is up, but slow. Max drawdown may be modest (for example 5%–15%), but it depends on streaks.
• Half-Kelly: Kelly for 52% at -110 is about 1.9% (rough guide). Half-Kelly is ~1%. So half-Kelly here looks like flat 1% and feels similar. If odds and edge change, stakes change too.
• 1% percent-of-bankroll: Stake stays at 1% but in dollars it moves with the bankroll. This gives small compounding.

Takeaway: With a small edge, flat and fractional Kelly can look close if your caps are tight. Variance stays in check.

Scenario B: Overconfident estimates

Say you think p = 55% at -110, but truth is only 52%.

• Full Kelly bet size was ~5.6% for 55%. That is large. But the real edge is much smaller. You are overbetting. Drawdowns can get very big.
• Half-Kelly cuts the stake to ~2.8%. Risk is lower, but still notable.
• Flat 1% keeps risk low even when you overrate your edge.

Lesson: Kelly works best when you know your edge well. If you are not sure, use fractional Kelly with caps, or use small flat stakes.

Tools, calculators, and tracking

Helpful tools:

• Investopedia on Kelly for the concept and a quick way to check math.
• Odds formats (Wikipedia) to convert American, decimal, and fraction odds.
• Pinnacle Betting Education for deep dives on variance, CLV, and staking.
• Khan Academy: Stats and probability to learn the basics of chance.

What to log for each bet: date, sport/market, odds, stake, your p (if you have one), result, and the closing line. Over time, check if you beat the closing line. That is a good signal of edge.

Where to compare sportsbooks: Prices, limits, and tools can vary a lot. If you want a simple, neutral list of reviews, see https://onlinecasinoitaliani.it/. Look at odds quality, markets, limits, and safe play tools before you choose.

Simple tests: You can run a spreadsheet simulation. Try 1,000 fake bets with your win rate and odds. Compare flat 1% to half-Kelly with a 2% cap. See average growth and worst drawdown. This shows how stakes change the ride. A light intro to simulation ideas: Monte Carlo method (Wikipedia).

Responsible gambling and legal note

This guide is for education. No system can remove risk. Set limits on time and money. Take breaks. If gambling harms your life, get help now.

• US: National Council on Problem Gambling
• UK: GamCare and BeGambleAware
• Global: Gambling Therapy

Gambling laws differ by place. Know your local rules and age limits. Only use legal sites in your area.

Key takeaways

• Flat betting is simple and safe for most new bettors. Try 0.5%–1% per bet.
• Kelly can grow fastest in theory but swings hard and needs a good edge estimate.
• Hybrids like half-Kelly with a 2% cap, or tiered units, often fit real life best.
• Write your rules. Track every bet. Review results. Change slow, not fast.
• Protect your bankroll. No system beats bad odds or bad habits.

FAQs

What is a betting unit?

It is your standard stake size. Many use 0.5%–2% of bankroll as one unit.

Is Kelly always best?

No. Kelly is best in theory if you know your true edge. In real life, edges are noisy. Many use fractional Kelly with caps to cut risk.

How do I find my edge p?

Use data and test. Track your picks for a long time. See if you beat the closing line. Read more: Closing line value guide.

What is a safe percent for flat betting?

For most casual bettors, 0.5%–1% per bet is a good start. Go smaller if swings feel hard.

Does fractional Kelly help much?

Yes. Half-Kelly cuts drawdowns a lot while keeping much of the growth. See research: Fractional Kelly studies (SSRN).

Why is martingale risky?

Because bet size doubles after each loss. A long bad run can hit table or account limits fast. You can lose your bankroll in one streak.

How often should I adjust stake size?

Pick a schedule (for example weekly) and stick to it. Do not change after each bet. That keeps noise low.

Sources and further reading

• Kelly Criterion (Wikipedia)
• Investopedia: Kelly Criterion
• MacLean, Thorp, Ziemba: Kelly and Risk (SSRN)
• Edward O. Thorp’s papers and talks
• Gambler’s Ruin (Wikipedia)
• Pinnacle: Kelly explained
• GamCare • BeGambleAware • NCPG

Editorial note on methods: Examples here use common odds and simple math. Results can vary a lot with price, market, and sample size. For strong tests, run long samples and simulations. Track CLV and compare systems side by side.