Martingale vs Flat Betting in NBA Markets: What the Simulations Show
Available in US
How Martingale Works — and Why It Fails on NBA Spreads
Every NBA bettor encounters Martingale eventually. Someone at a pub, on a forum, or in a YouTube comment section swears by the system: double your bet after every loss, and when you finally win, you recover everything plus a profit equal to your original stake. It sounds bulletproof. It is not. I tried it for exactly one week in 2016 and my seventh consecutive loss required a bet of 1,280 pounds to continue the sequence — on a bankroll that started the week at 2,000. That was the end of my Martingale experiment.
The system’s logic is straightforward. You bet 10 pounds. Lose. Bet 20 pounds. Lose. Bet 40. Lose. Bet 80. Win. Your total outlay is 10 + 20 + 40 + 80 = 150 pounds, and your return on the winning bet is 80 x 1.91 = 152.80 pounds. Net profit: 2.80 pounds. You risked 150 to make 2.80. The risk-to-reward ratio is grotesque, but the system “works” as long as you eventually win before running out of money.
The problem is that losing streaks in NBA spread betting are longer than most people intuitively expect. A bettor with a 52% win rate will experience a run of seven or more consecutive losses roughly once every 200-300 bets. Over a full NBA season of 300 bets, that means at least one Martingale-destroying streak is virtually guaranteed. After a seven-loss streak, the required bet is 128 times your original stake. After ten losses, it is 1,024 times. Bookmaker maximum bet limits typically kick in before that, ending your sequence even if your bankroll could theoretically sustain it.
Between 1% and 2% of the US adult population meet the criteria for gambling disorder, and research consistently identifies loss-chasing as one of the earliest warning signs. Martingale is loss-chasing formalised into a system. It gives the gambler a mathematical excuse for the exact behaviour that clinical psychologists flag as problematic. That alone should give any serious bettor pause.
D’Alembert and Fibonacci Systems Applied to NBA Betting
If Martingale is the aggressive end of progressive staking, D’Alembert and Fibonacci are its more cautious cousins. Both increase stakes after losses but at a slower rate, which reduces the risk of catastrophic blow-up. Whether that slower escalation produces positive results over a full NBA season is a different question.
D’Alembert increases your stake by one unit after a loss and decreases by one unit after a win. Starting at 10 pounds (one unit), a loss takes you to 20, another loss to 30, a win drops you to 20, another win to 10. The progression is linear rather than exponential, which means a seven-loss streak requires an 80-pound bet rather than Martingale’s 1,280. The system is more survivable, but it still increases exposure during losing periods — which is exactly when your bankroll is least equipped to absorb larger bets.
Fibonacci follows the famous sequence: 1, 1, 2, 3, 5, 8, 13, 21, 34. After a loss, you move one step up the sequence. After a win, you move two steps back. The escalation is faster than D’Alembert but slower than Martingale. A seven-loss streak from a 10-pound base takes you to 130 pounds per bet rather than 1,280 (Martingale) or 80 (D’Alembert). The Fibonacci system is the most popular progressive alternative because it feels like a reasonable compromise — but “feels reasonable” and “produces positive expected value” are not the same thing.
A 52% bettor using online betting reported by the Siena College survey found that 52% of online bettors had chased losses at some point. All three progressive systems — Martingale, D’Alembert, and Fibonacci — institutionalise that chase. They assume the next bet is more likely to win simply because the previous one lost, which is not true. NBA spread outcomes are approximately independent events. A loss on Tuesday does not make a win on Wednesday more probable. Progressive systems cannot create edge where none exists; they can only amplify existing edge (if you have one) while simultaneously amplifying risk.
The Case for Flat Staking: Steady Returns, Controlled Risk
Flat staking is boring. You bet the same amount on every game, win or lose, regardless of recent results or emotional state. Two percent of bankroll, every time. No escalation after losses. No reduction after wins. Just the same stake, applied to selections you have researched and believe carry positive expected value.
A 2020 simulation by Dotan tested staking strategies on NBA betting data and found that 1/5 Kelly — which varies stake size based on estimated edge but never increases after losses — delivered ROI exceeding 98% over one season. Full Kelly, which is more aggressive, led to complete bankroll depletion. Flat staking sits between these extremes: it does not optimise for maximum growth rate like Kelly, but it also does not expose you to the ruin risk of progressive systems. For most bettors, that trade-off is the right one.
The psychological advantage of flat staking is at least as important as the mathematical one. When every bet is the same size, a loss stings equally whether it happens on Monday or Friday, whether you are on a winning streak or a losing streak. There is no temptation to bet bigger to recover, no mental accounting that treats “house money” differently from your original bankroll. The emotional flatness of the system is precisely the point — it removes the decision-making about stake size from the equation entirely, freeing you to focus all your analytical energy on selection quality.
I have used flat staking as my default approach for nine of my eleven years betting NBA. The exceptions — two seasons experimenting with Kelly-based sizing — produced slightly better theoretical returns but significantly more emotional volatility. I went back to flat staking not because it was mathematically optimal but because it was psychologically sustainable. A system I can run for 30 years without burning out is more valuable than a system that squeezes an extra 1% of ROI but demands constant recalibration and emotional discipline that, frankly, I do not always have at 02:00 GMT on a Thursday.
For a detailed exploration of how Kelly-based sizing compares to flat staking in a more rigorous framework, my guide on Kelly criterion for NBA betting lays out the simulation data and practical implementation steps.
Articles
Written by the editors at CourtEdge.
