How many bets do you need to know if you're profitable?

At a 3% edge on even-money bets, roughly 4,000–5,000 before results alone are convincing. Closing line value gives a useful read in 50–100 bets.

Is a losing streak of 10 normal?

Yes. At about a 51.5% win rate, the median longest losing streak over 1,000 bets is 9, and about three seasons in ten include a run of 10 or more.

What is a normal drawdown for a winning bettor?

In simulations of 1,000 even-money bets at a 3% edge, the median worst drop is about 26 stakes, and one season in ten sees 44 or more.

Does betting favourites reduce variance?

Yes, per bet: shorter odds win more often, so results swing less for the same stake. The edge still takes many bets to show.

+EV

Variance in sports betting: how long until an edge shows

At a 3% edge on even-money bets, you are still behind about one time in six after 1,000 bets. Results take thousands of bets to prove an edge; closing line value takes far fewer.

Updated September 27, 2026 · TrueEdge Academy

Variance is how far your actual results stray from your expected value, and in sports betting it is large. A bettor with a real 3% edge on even-money bets is still behind more than a third of the time after 100 bets, about 17% of the time after 1,000, and only drops below 2% after about 5,000. So results alone cannot tell you whether you have an edge until you have placed thousands of bets. Closing line value can tell you in dozens.

The standard deviation of one bet

A $100 bet at +100 either wins $100 or loses $100. Its standard deviation, the typical distance between a result and the average, is about $100: a full stake. Its expected value at a 3% edge is $3. So one bet's noise is more than thirty times its signal. Over N bets the expected profit grows with N, but the standard deviation grows only with the square root of N, which is the only reason an edge ever shows at all.

A 3% edge at +100, $100 flat stakes

Bets: Expected / one std dev / chance of being behind
100: +$300 / $1,000 / 38%
250: +$750 / $1,580 / 32%
500: +$1,500 / $2,240 / 25%
1,000: +$3,000 / $3,160 / 17%
2,500: +$7,500 / $5,000 / 7%
5,000: +$15,000 / $7,070 / 2%

p = 51.5% (0.515 × 2.00 − 1 = 3%). Std dev = $100 × √N × 2√(p(1 − p)). Chance behind uses the normal approximation. The edge is two standard deviations clear only at about 4,400 bets.

Longer odds, longer wait

The same 3% edge at +300 wins about a quarter of the time, so results are lumpier. Each $100 bet has a standard deviation of about $175 instead of $100, and the number of bets needed for the same confidence grows with the square of that: roughly 13,600 bets to reach two standard deviations, against about 4,400 at +100. After 1,000 bets at +300 you are still behind about 29% of the time. A strategy built on longshots needs a bigger bankroll and more patience for the same edge.

Drawdowns over 1,000 bets (simulation)

Setup: 1,000 bets, +100, 51.5% win rate, $100 flat
Expected profit: +$3,000
Median largest peak-to-trough drop: $2,600 (26 stakes)
1 run in 10 sees a drop of at least: $4,400 (44 stakes)
Median longest losing streak: 9 in a row
Runs with a losing streak of 8+: about 79%

20,000 simulated seasons. The edge is real in every one of them. A $2,600 drawdown on a $3,000 expectation is the normal case, not bad luck.

What that means for your bankroll

Plan for the drawdown before it happens. If a 26-stake drop is typical and a 44-stake drop happens one season in ten, a bankroll of 20 stakes will not survive the method working as intended. That is why staking is a percentage of bankroll, well under Kelly (see the Kelly guide): stakes shrink as the bankroll does, and the risk of ruin falls sharply. And it is why the bankroll must be money you could lose without it hurting.

Why CLV is the faster signal

Closing line value compares the price you took with the market's final price. It does not depend on who won, so most of the noise is gone. As an illustration, suppose your bets beat the close by 2% on average and the figure varies by about 5 points from bet to bet. After 50 bets the standard error of your average is 5 ÷ √50 ≈ 0.7 points, so a +2% average is nearly three standard errors above zero. Results would need thousands of bets for the same confidence. The CLV guide explains why the closing line is the market's best estimate.

Seeing it in your own numbers

The Bet Tracker records each bet with its price and stake and, where it can match the bet to a market TrueEdge captured, its closing price. Its Avg CLV tile averages that across the settled bets that have a close, alongside profit, so you can see whether a losing month came from bad prices or bad luck. The walkthrough explains why the two tiles can disagree.

Frequently asked questions

How many bets do you need to know if you're profitable?: At a 3% edge on even-money bets, roughly 4,000–5,000 before results alone are convincing. Closing line value gives a useful read in 50–100 bets.
Is a losing streak of 10 normal?: Yes. At about a 51.5% win rate, the median longest losing streak over 1,000 bets is 9, and about three seasons in ten include a run of 10 or more.
What is a normal drawdown for a winning bettor?: In simulations of 1,000 even-money bets at a 3% edge, the median worst drop is about 26 stakes, and one season in ten sees 44 or more.
Does betting favourites reduce variance?: Yes, per bet: shorter odds win more often, so results swing less for the same stake. The edge still takes many bets to show.