A +EV bet still loses because value improves the price, not the outcome: +110 (2.10) on an outcome worth +100 (2.00) still wins only about one time in two. The 5% edge is real, but it only shows over many bets. After 100 bets like that there is still a 31% chance of being down; after 1,000 it falls to 7%. That swing is called variance.

What is variance in sports betting?

It is the gap between what you expect on average and what actually happens over a given number of bets. A value bet has a positive expected return, for example +5 cents per dollar staked, but each bet ends only one of two ways: you win (decimal odds − 1) or you lose the whole stake. The +5% is an average that only emerges once the bets add up.

In the short run variance rules: you can bet well and lose for a month, or bet badly and win for a month. It is also why a lucky tipster can look skilled, as How do you verify a tipster or handicapper? explains.

How many bets does it take for the edge to show?

Take bets of 1 unit at +110 (2.10) on outcomes with fair odds of +100 (2.00): a 50% chance to win, a +5% edge. Exact calculation with the binomial distribution:

Bets placed	Expected profit	Chance of being down
100	+5 units	30.9%
250	+12.5 units	24.3%
500	+25 units	15.2%
1,000	+50 units	6.9%
2,000	+100 units	1.7%

Losing streaks are normal too: over 500 bets at 50%, more than half the time there is at least one run of 8 losses in a row. It is not a sign the method has stopped working.

What does variance look like in real results?

NinjaBet members log the value bets they place in the Bet Tracker, and the ValueBet results page publishes them all. Here is the return on turnover month by month since September 2025 (as of October 16, 2026, settled straight bets, all NinjaBet sites):

Month	Settled bets	Return on turnover
September 2025	481	−1.28%
October 2025	821	+11.83%
November 2025	308	+33.53%
December 2025	263	+16.69%
January 2026	648	−2.73%
February 2026	464	+13.84%
March 2026	488	+47.81%
April 2026	601	+11.70%
May 2026	652	+9.30%
June 2026	803	+22.30%
July 2026	452	+28.12%
August 2026	373	+16.19%
September 2026	2,616	+16.81%

The same kind of bet returned anywhere from −2.73% to +47.81% in a month. In the ten days from October 6 to 16, 2026, the 490 settled value bets saved from the tool lost 6.84%, despite a positive average CLV (+2.40%): What is CLV? tells that story. The average return over the whole period is +9.59% on 21,793 bets: a figure members entered themselves, not a promise.

How do you handle variance?

1. Stake a small, flat amount: 1-2% of your bankroll per bet. With big stakes, a run of 8-10 losses empties the account before the edge shows. The Bankroll and Kelly guide explains how to size it.
2. Don't chase losses by raising stakes after a bad run: it raises the risk without raising the edge.
3. Judge the method by CLV, not by the week's balance: if you keep beating the fair closing odds, the results come with volume.
4. Build volume: the more +EV bets you place, the closer the result gets to the expectation. ValueBet is there to find enough of them.
5. Log everything in the Bet Tracker and review results by quarter, not by day.
6. Be careful with long odds: at +500 (6.00) you win one time in six and the variance is far bigger than at +100. With a small bankroll, stay with mid-range prices.

Want to see it in practice? +EV picks shows members' verified picks and their public record, Check a pick prices any tip you are given, and the Telegram channels get them first.

Frequently asked questions

I lost for a month betting only value bets. Does the method not work?

A month is a small sample: members have had losing months too. Check your CLV and the number of bets before you change anything.

How big a bankroll do you need to ride out variance?

Big enough to stake 1-2% per bet: with $1,000 that is $10-$20 a bet. Then even a run of 10 losses costs 10-20% of the bankroll, not all of it.

Can you remove variance?

Not with value bets: it only shrinks with volume and small stakes. If you want a margin locked in on every play, that is the territory of surebets, which carry different risks.

Short or long odds to keep variance down?

At the same edge, shorter odds swing less. Longer odds are not worse, but they need more bets and a bigger bankroll to reach the expected result.

Does the Kelly criterion help?

Kelly works out the optimal stake from the edge and the odds, but full Kelly is very aggressive: many bettors use a fraction (a quarter or a half). The Variance and long-term results guide covers it.

Sources

• NinjaBet, ValueBet results: monthly return of the value bets members logged, figures as of October 16, 2026.
• NinjaBet calculation: binomial distribution for bets at 2.10 with a 50% win probability; losing streaks simulated over 20,000 sequences of 500 bets.

You must be 21 or older (18 in a few states) to bet in the United States, and bet only with operators licensed in your state. Gambling can be addictive: if you or someone you know has a gambling problem, call or text 1-800-GAMBLER, or the National Problem Gambling Helpline at 1-800-MY-RESET (ncpgambling.org), free and confidential, 24/7. Results vary and require time and capital; this article is information, not financial or legal advice. Last updated: October 16, 2026.