You have some money you're prepared to put on player props, and you don't yet have a rule for how much goes on each one. By the end of this page you'll have the money set aside, a unit size in dollars, three hard limits and a log with your first bets in it. It takes about an hour. Most of the hard part has nothing to do with math.
One idea holds up everything else
How much you stake on each bet decides whether you are still betting at the end of the season. Picks matter too, but nothing else in bankroll management matters as much.
Three terms first. A bankroll is money set aside only for betting, taken from disposable income and kept apart from rent, bills and savings. A unit is your standard bet, set as a fixed share of that bankroll, usually somewhere between 1% and 5%. Risk of ruin is the chance of losing the whole bankroll before your edge, if you have one, has time to show up.
Here is why the unit matters more than it looks. DeucesCracked ran the numbers for a bettor who wins 54% of bets at -110. Ten losses in a row happen to winning bettors too. At 10% of the bankroll per bet, that streak wipes out roughly 65% of the money. At 1% per bet, the same streak costs under 10%. It's the same bettor with the same picks, and only one of them is still in business afterward.
An edge doesn't protect you on its own. Wizard of Odds worked through a bettor with a real edge: 55% winners at -110, $100 flat bets, 100 bets. The expected profit is $500, but the 95% range of outcomes runs from about -$1,362 to +$2,362. Even a fairly modest plan carries real risk. On the same source's math, a $1,000 bankroll betting $50 a time at that 55% win rate has a risk of ruin of about 1.96%.
One more thing to understand before you start. Bankroll management can't make a losing bettor win. Michael Shackleford's verdict on betting systems is that "all they can do is affect volatility," and he's right. What sizing does is keep you in the game long enough for your picks to be judged on a few hundred bets instead of a bad weekend.
What you need before you start
You need less than the sportsbook ads suggest. Setting up costs nothing beyond the bankroll itself, and the setup takes about an hour.
-
A sum you can lose without it changing anything else in your life
TopEndSports suggests $100 to $500 as a beginner bankroll. DeucesCracked puts a casual bettor's yearly entertainment budget at $100 to $1,000, and a full-time professional's bankroll at 3 to 6 months of living expenses. A quick test from the same source: if losing $100 would hurt, your biggest single bet should be about $5.
-
Enough depth for 50 to 100 units
DeucesCracked recommends a bankroll that can absorb at least 50 to 100 units without strain. At a 2% unit you get 50, and at 1% you get 100.
-
A funding method that uses money you already have
Debit card or bank transfer. Most of the market has stopped taking credit cards anyway (more on that below).
-
Deposit and loss limits switched on in each sportsbook app
Regulated books offer deposit limits, bet amount limits, loss limits, time limits and cool down periods. All of them are opt-in, so you have to set them yourself.
-
A log
A free spreadsheet works, or the prop tracker on this site. You'll set up the columns in step 6 below.
Setting up a flat-unit bankroll in eight steps
Use flat betting: one unit on every bet, whatever your confidence. DeucesCracked calls flat staking "ideal for beginners," and it's the right call. It's simple, it keeps your risk steady, and it takes the emotion out of sizing at exactly the stage when your confidence ratings mean nothing yet.
From cash to first logged bet
-
Write down one number
Decide your bankroll and write it down: say $1,000. Worked when: you can name the figure and none of it is earmarked for rent, bills or savings.
-
Move it somewhere separate
Fund your sportsbook accounts with debit or bank transfer, and keep any undeposited part of the bankroll apart from your spending money. Worked when: your sportsbook balances plus any reserve add up to exactly the number from step 1.
-
Lock in deposit and loss limits
In each app's responsible gaming settings, set a deposit limit no higher than your bankroll and a loss limit to match. Worked when: the app confirms the limits are active.
-
Pick a unit of 1% to 2%
Multiply the bankroll by your percentage. $1,000 at 2% is a $20 unit. TopEndSports calls 1% to 2% the safest starting point. TheSpread.com goes narrower, suggesting 0.5% to 2% with 1% as a popular starting point. Worked when: you have a dollar figure written next to the bankroll figure.
-
Set three caps
No more than about 2 units on any one game. No more than 5% to 7% of the bankroll staked in a day. Stop for the day at -4 units. All three figures are from TheSpread.com's risk limits. On a $1,000 bankroll with a $20 unit, that's $40 per game, $50 to $70 per day, and stop at -$80. Worked when: all three are written down where you'll see them before betting.
-
Build the log
Columns: date, sport, market, sportsbook, odds, stake in units, closing price, result, profit or loss, and a one-line reason. Add a running bankroll column. Worked when: the header row exists and the first row shows your starting balance.
-
Place and log your first bets
Bet one unit, say a points over at -110 for $20. Log it before the game starts, then add the closing price once the line closes. Worked when: every bet you've placed has a row, losers included.
-
Review monthly, resize only on a trigger
Look at the log once a month, not after every game. Change the unit only when the bankroll has moved by 25% to 33% (TheSpread's recalibration trigger). Consider withdrawing once you're about 10 units above where you started (TopEndSports). Worked when: your unit changes on a date in the calendar, not the morning after a bad night.
Step 4 is where people overreach, so here is the whole grid. The 1% to 2%, 3% and 4% to 5% tiers are the conservative, moderate and aggressive bands from TopEndSports' bankroll guide. The dollar figures are just bankroll times percentage. Note that TopEndSports itself says 4% to 5% is for experienced bettors only, and that no unit should ever exceed 5%.
| Bankroll | Conservative 1% | Conservative 2% | Moderate 3% | Aggressive 4% | Aggressive 5% |
|---|---|---|---|---|---|
| $100 | $1 | $2 | $3 | $4 | $5 |
| $500 | $5 | $10 | $15 | $20 | $25 |
| $1,000 | $10 | $20 | $30 | $40 | $50 |
| $2,000 | $20 | $40 | $60 | $80 | $100 |
| $5,000 | $50 | $100 | $150 | $200 | $250 |
A $1,000 bankroll through its first 100 bets
A hypothetical bettor starts with $1,000 in November 2026 and flat bets a 2% unit of $20 on props priced at -110. A winner returns $18.18 profit and a loser costs $20. The record comes in four stretches, including a cold run early on.
| Start | Bankroll $1,000, unit $20 |
|---|---|
| Bets 1 to 10 (4 won, 6 lost) | -$47.27, bankroll $952.73 |
| Bets 11 to 20 (3 won, 7 lost) | -$85.45, bankroll $867.27, the low point |
| Bets 21 to 40 (12 won, 8 lost) | +$58.18, bankroll $925.45 |
| Bets 41 to 100 (34 won, 26 lost) | +$98.18, bankroll $1,023.64 |
| Total wagered | 100 bets x $20 = $2,000 |
| ROI | $23.64 / $2,000 x 100 = 1.18% |
Look at the example's 53% win rate. At -110 you need about 52.4% just to break even (110 divided by 210). That's why the result is barely positive, and why 100 bets tell you very little. DeucesCracked calls 100 bets a small sample, several hundred a rough picture, and a thousand or more the point where the picture gets clear. Your first month's job is to get the mechanics right. It won't tell you whether you can beat the books.
Four ways the first month goes wrong
1. Chasing losses. This is the one every source names. NCPG's guidance ahead of Super Bowl LX was to set a budget before betting, set a time limit, and stop when you reach either one. It also warned specifically against chasing. Doubling systems like the Martingale are a formal version of the same mistake, and DeucesCracked calls loss-chasing the fastest way to bust a bankroll. How to tell it's happened: your log has a stake bigger than one unit placed the same day as a loss. What to do: adopt TheSpread's "no chase rule" in so many words. A loss never increases the unit. Close the app for the day.
2. The oversized "lock." A tipster, a group chat or your own gut flags a bet that can't lose, and you put five units on it. TopEndSports puts the ceiling at 3 units on any bet and notes that most bettors stay within 1 to 2. On a flat plan the answer is simpler still: it gets one unit, like everything else. How to tell: any row in the log above 2 units.
3. Stacking props on one slate past your daily cap. This one is specific to prop bettors, and it's the easiest to miss because each bet looks small. Wizard of Odds shows that ten simultaneous props, each sized at 5% Kelly, add up to risk equivalent to about a 15.8% single bet. That's well beyond what any of the tiers above allow. How to tell: add up today's open stakes. If they're above 5% to 7% of the bankroll, you're over the cap. What to do: stop adding bets until some of the open ones settle.
The best argument against stop-losses deserves a hearing here. Shackleford says "there is no magic quitting point," and mathematically he's correct. Quitting at -4 units doesn't change the expected value of the bets you've already chosen. But that isn't what a stop-loss is for. It's there to stop the next bet, the one placed angry. The next.io guide and NCPG both point to emotion-driven betting as the route into chasing. A rule set in advance does the deciding that you won't do well after a bad hour.
4. Reading the log too early, or reading the wrong number. DeucesCracked's warning is blunt: "avoid making large changes based on 20 or 30 bets." Win rate misleads on its own, too. TopEndSports notes that 45% on +200 underdogs can be profitable while 60% on -200 favorites can still lose money. Other tracking errors come up repeatedly: logging only the wins, and mixing bonus bets in with cash bets, which distorts ROI because free bet stakes usually aren't returned. How to tell: you've changed your unit or your markets since the start of the month. What to do: change them back, and decide at the monthly review instead. If a cold run is what's tempting you to change things, there's a page on getting through a losing streak.
Once flat betting is working
The next thing isn't a new staking system. It's a better log. Two numbers matter. ROI is profit divided by total wagered, times 100: $50 profit on $1,000 wagered is 5%. TopEndSports describes 5% to 7% sustained over the long run as excellent, and DeucesCracked puts professionals hitting 52% to 55% at roughly 4% to 5%. Expect thin margins. Closing line value (CLV) compares the price you took with the final price before the game, so taking +150 on a prop that closes at +130 is positive CLV. DeucesCracked treats beating the close more than 55% of the time as a strong sign, and an average CLV of 1% to 2% as meaningful over a large sample. The prop tracker logs both. Be careful with thin markets, though: closing lines on small props can be set by a single book with little action, so they say less.
After a few hundred logged bets, you can consider letting your stakes move. Percentage staking recalculates the unit from the current bankroll on every bet. At 2% of $1,000 the first bet is $20, and after a loss the next is $19.60. Your stakes shrink automatically in a downswing and grow as profits build. The cost is more arithmetic.
The Kelly Criterion sizes each bet by your edge: f = (bp - q) / b, where b is the decimal odds minus 1, p is your win probability and q is 1 - p. It comes from John Kelly's 1956 work at Bell Labs. At -110 with a 55% win probability, full Kelly says 5.5% of the bankroll, or $55 on $1,000. If the result is zero or negative, it says don't bet. Full Kelly is too aggressive for almost everyone, because it trusts your probability estimate completely. MetricGate's 10-bet sample averaged a full Kelly stake of 22.46% of the bankroll. The standard fix is a fraction of it. Per Wizard of Odds, half Kelly ($27.50 in the example) keeps about 75% of the growth with 50% of the variance, and quarter Kelly ($13.75) keeps about 50% of the growth with 25% of the variance. To work a stake out from your own projection and the posted price, use the prop edge calculator. Honestly, though, if you can't yet say with evidence what your win probability is, Kelly is just a guess with a formula attached. Stick with flat units until the log can answer that.
Ready to size bets by edge instead of flat units?
The Kelly page covers when fractional Kelly makes sense and how to get from a projection to a stake.
Read the Kelly Criterion guide
Comments
No comments yet. Be the first to comment!
Leave a Comment