There's no single "correct" way to manage a betting bankroll. Professional bettors use different models depending on their edge, risk tolerance, and betting style. Here are the main approaches and when each one makes sense.
The flat betting vs Kelly debate has raged in sports betting circles for decades, and the truth is that the "best" model depends entirely on where you are in your betting journey. A beginner with 200 tracked bets needs a completely different system than a professional with 10,000 bets of closing line value data. If you're new to bankroll concepts entirely, start with our complete bankroll management overview before diving into the specifics below.
You can have a legitimate 3% edge on every play you make and still go broke if your sizing strategy is wrong. That's why understanding these models isn't optional. It's the difference between surviving long enough for your edge to play out and blowing up your bankroll in a brutal two-week stretch.
Flat betting means every bet is the same size, usually 1-3% of your total bankroll. You don't adjust based on confidence, odds, or how strong you think the play is. Every bet gets the same amount. For the fundamentals on what percentage to use, see our unit sizing guide.
Simple to implement. No need to estimate your edge on individual bets. Prevents overbetting on plays you're overconfident about. Works well for bettors who aren't good at distinguishing between strong plays and weak plays.
Here's what most people overlook about flat betting: it's incredibly forgiving. You can't blow 15% of your bankroll on a "lock" that loses. That kind of protection is worth more than most bettors realize, especially during the inevitable losing streaks that test every bettor's discipline.
Leaves profit on the table when you have genuinely strong edges. Treats all bets equally even when some clearly have more value than others. If you're consistently identifying +5% edges and +1% edges, flat betting forces you to treat them identically, which mathematically limits your growth rate over time.
You're betting NFL sides and totals, hitting 54% ATS at standard -110 juice over 500 tracked bets. You know you have an edge, but you're honestly not sure which plays are your strongest. Some weeks your most confident pick loses, and your throwaway pick cashes. If your confidence doesn't reliably predict actual edge size, flat betting is your best friend.
Beginners, recreational bettors, and anyone without a proven track record of identifying edge size accurately.
Kelly adjusts bet size based on your estimated edge and the odds. The bigger your edge, the more you bet. The smaller your edge, the less you bet. It maximizes long-term growth if your edge estimates are accurate. For the full mathematical breakdown, see our Kelly Criterion deep dive.
The Kelly formula is straightforward: (bp - q) / b, where b is the decimal odds minus 1, p is your estimated win probability, and q is 1 minus p. A 2% edge at -110 gets a small bet. A 7% edge at +150 gets a much larger one. The math automatically adjusts for both your edge and the odds, which is something no fixed-unit system can do.
Mathematically optimal for long-term growth. Automatically scales bet size to match edge size. Tells you when not to bet (negative Kelly). Over a large sample, full Kelly maximizes bankroll growth faster than any other strategy. It's the theoretical ceiling for how fast your money can compound.
Requires accurate edge estimation, which most bettors are terrible at. Full Kelly creates massive volatility, with drawdowns of 40-60% being completely normal. If your true edge is 2% and you think it's 5%, Kelly will have you betting more than twice what you should, which can turn a winning strategy into a losing one through overbetting alone.
Almost nobody uses full Kelly in practice. The volatility is simply too brutal. Most professionals use fractional Kelly, typically one-quarter to one-half of the full recommendation. Half Kelly preserves most of the compounding potential while cutting variance roughly in half. Quarter Kelly captures about 50% of the growth rate but with dramatically smoother equity curves. We break this tradeoff down in our fractional vs full Kelly comparison.
You've been betting NBA totals for three years with 4,000 tracked bets, and your model consistently beats the closing number by 1-3 cents. This is where Kelly shines: you size into your strongest outputs (3 cents of CLV) more aggressively than your marginal ones (1 cent of CLV). Your bankroll grows faster than flat betting allows because you're not just guessing at which bets deserve more.
Experienced bettors with proven ability to beat closing lines. If you can't articulate exactly why your edge is 3% instead of 2%, you're not ready for Kelly. Use fractional Kelly (half or quarter) to reduce risk regardless of experience level.
Some bettors use a tiered system: 1 unit on standard plays, 2 units on strong plays, 3 units on best bets. This is simpler than Kelly but allows some variation based on perceived edge. It's the most popular model among semi-professional and serious recreational bettors because it feels intuitive and doesn't require a spreadsheet for every bet.
The hybrid approach is essentially simplified Kelly thinking without the precise math. Instead of calculating exact percentages, you're bucketing plays into categories. The key is keeping the tiers tight. A 1-2-3 unit system where your max bet is only three times your minimum is far more sustainable than a 1-3-5 system where you're risking five times as much on "locks" that inevitably lose sometimes.
More flexible than flat betting. Easier to implement than Kelly. Lets you bet more on your best plays without needing to calculate exact edge percentages. Most bettors find this approach psychologically easier to stick with during losing streaks compared to full or even fractional Kelly.
Confidence doesn't always equal edge. Most bettors overestimate their edge on "best bet" plays and end up overbetting their weakest actual plays. Unless you've tracked your results by tier and confirmed that your 3-unit plays actually hit at a higher rate than your 1-unit plays, you could be doing more harm than good with variable sizing.
You've been betting for a year with a solid record, and you've noticed your plays based on specific situational angles (rest advantages, travel spots, divisional revenge) hit at a higher rate than your general model plays. A tiered system lets you put 2 units on those situations and 1 unit on standard plays. Just make sure you're verifying that pattern with data, not just remembering the wins.
Intermediate bettors who have some ability to identify stronger plays but don't want the complexity of Kelly. Also works well for bettors transitioning from flat betting who want to test whether their confidence actually correlates with their edge.
Some bettors set a profit target (e.g., win $100 per game) and adjust their bet size based on the odds to hit that target. Betting a favorite at -200 requires a $200 bet to win $100. Betting an underdog at +150 requires a $67 bet.
This model bets more on favorites and less on underdogs, which is backwards. Favorites are typically more efficiently priced, so you end up with your largest risk amounts on plays where the books are most accurate. Trying to win $100 on a -500 favorite means risking $500. One bad loss wipes out five wins.
Professional bettors avoid this model entirely. If you're currently using it, switch to flat betting immediately. You'll be better off the same day.
Most modern professional bettors use some version of percentage-of-bankroll betting, whether flat or variable. The key principles:
- Recalculate bankroll regularly (weekly or biweekly)
- Never bet more than 5% on any single game
- Use fractional Kelly if varying bet size
- Track closing line value to verify edge
The recalculation piece is critical. If your bankroll grows from $5,000 to $7,000, your unit size should grow proportionally. If it drops to $3,500, your unit size should shrink. This protects you during downswings and lets you capitalize during upswings. Our bankroll management guide covers the mechanics of when and how to recalculate.
One of the biggest mistakes bettors make is jumping straight to Kelly sizing with no track record. Here's a practical roadmap for evolving your approach as you gain data.
When you're starting out, you don't know what you don't know. Flat bet 1-2% of your bankroll on every play. Your only job during this stage is to build a sample size and track everything obsessively. Record the closing line for every bet, because that's going to matter enormously later.
After 1,000+ bets, analyze your data. Did your "confident" plays actually win more than your marginal ones? If yes, introduce a simple 1-2 unit tiered system. If no, stay flat. Most bettors discover their confidence doesn't correlate with results as strongly as they assumed. Better to know that and stay flat than to pretend otherwise.
Once you have a large enough sample with closing line value data, you can start estimating your actual edge on individual plays. If your model consistently beats the close by measurable amounts, implement quarter Kelly or half Kelly sizing. Start conservative. You can always increase your Kelly fraction later if your edge holds, but you can't un-lose money from oversizing too early. For a deeper look at what fraction to use, see our guide on choosing between fractional and full Kelly.
The vast majority of professionals use some form of fractional Kelly, typically between one-quarter and one-half of the full recommendation. Almost none use full Kelly. The ones who tried it early in their careers have stories about 50% drawdowns that permanently converted them to fractional approaches.
The pros who flat bet tend to be high-volume bettors running automated models across thousands of monthly bets, where sheer volume handles the rest. The ones who use Kelly tend to be more selective, making fewer plays with larger individual edges. If you're making 10-20 bets per week with edges ranging from 1% to 8%, fractional Kelly lets you lean into those 8% spots while staying disciplined on the 1% spots.
What you won't find among professionals is gut-feel sizing or anything resembling "I feel good about this one, let me bet the house." Every professional has a systematic approach, and they follow it regardless of how they feel about an individual play. The system protects them from themselves.
The bottom line? Pick the model that matches your current skill level, not the one that sounds most sophisticated. Flat betting done consistently will always outperform Kelly done poorly. And any systematic approach will crush the "bet more when I'm feeling it" strategy that most recreational bettors default to.
Related Reading: