Sports betting strategies: which work and which don't

No betting system changes the bookmaker's margin. If both outcomes are priced at 1.90, over time you lose about €5 of every €100 you stake on average, whatever order and amounts you bet in. Sports betting strategies fall into two groups: some decide how much to stake, others what to bet on. The first group only changes how sharply your bankroll swings, and only the second can even try to beat the margin. We worked through the best-known systems, from Martingale to value betting, so you can see what each one does with your money.

Why the order of your bets cannot beat the margin

The margin is built into every price, so every bet loses the same share on average, however large it is.

If a match has two equally likely outcomes, the fair price on both would be 2.00. The bookmaker offers, say, 1.90 and 1.90. Add up the implied probabilities of both prices: 1/1.90 + 1/1.90 = 105.26%. Those 5.26% above a hundred are the margin. If the true probability of the outcome is 50%, every euro you stake returns 0.5 × 1.90 = €0.95 on average. Of €1,000 staked as a hundred €10 bets, €950 comes back on average.

A betting system only decides how much and when you stake. It does not change the value of any single bet, so the total of all your bets loses the same 5% on average. How to read odds and implied probability is explained in our sports betting basics.

Martingale in sports betting

Martingale raises the stake after every loss so that the first win recovers everything lost, but in sports betting it hits the wall even faster than at roulette.

We tested Martingale at roulette in our roulette strategies article. In sports betting it has three extra problems.

The first is the odds. At roulette a bet on red pays 1:1, so doubling is enough. At a sportsbook, an outcome with about a 50% chance is typically priced around 1.90, and plain doubling no longer recovers your losses. Start at €10 and double after every loss, and a win on the fifth bet (€160 × 1.90 = €304) does not cover the €310 already staked, leaving you €6 down. For a win to truly recover everything and make €10 profit, the stakes have to grow faster: €11, €23, €50, €105, €221, €466, €984 and €2,077. After eight losses in a row you will have staked €3,937 to win €10.

The second problem is losing streaks. Even if the true probability is exactly 50%, long streaks come more often than it seems:

• Five losses in a row: the chance that these particular five bets all lose is 3.1%, but over 100 bets such a streak happens at least once 81% of the time.
• Seven losses in a row: 0.78% for one streak, 32% over 100 bets.
• Eight losses in a row: 0.39% for one streak, 17% over 100 bets, 63% over 500 bets.
• Ten losses in a row: 0.1% for one streak, 21% over 500 bets.

The third problem is limits. Sportsbooks set a maximum stake for every event: in Optibet's rules (4.1.3) it ranges from €1 to €20,000 depending on the event, and OlyBet (4.7) sets a maximum win for every event. When the next Martingale stake goes over the limit, the run ends with the whole accumulated loss.

As at roulette, Martingale wins often and a little, and loses rarely and a lot. On average it loses exactly the margin on everything staked - only the amount staked is many times larger.

Fibonacci, d'Alembert and chasing losses

Every system that raises the stake after a loss is the same Martingale idea at a slower pace, and none of them changes the margin.

In the Fibonacci system you move one step up the sequence 1, 1, 2, 3, 5, 8, 13 after a loss, and in the d'Alembert system you add one unit to the stake after a loss. Both raise stakes more slowly, so the wall comes later, but a single win no longer recovers everything. For roulette we worked out both in the same article; in sports betting the result is the same, only the margin is usually higher than European roulette's 2.70%. According to University College Dublin economist Karl Whelan, it averages 5.3% on the English Premier League 1X2 market and up to 9% in lower leagues.

Chasing losses is the same system without rules: a bigger stake after a bad day, to "win it back". It does not raise the chance of recovering what you lost; it only makes the loss bigger if the next bet loses too. If you notice that you are staking more in order to win it back, that is the signal to stop, and responsible gambling tools exist for exactly these moments.

Flat staking and percentage staking

Flat staking and percentage staking do not change the margin, but they are the only systems that protect your bankroll from a single bad run.

With flat staking you stake the same amount every time, for example €10. With percentage staking you stake the same share of your current bankroll, for example 2%. With a €500 bankroll, 2% is €10; after losses the stake shrinks, after wins it grows.

For comparison: eight losses in a row at a flat €10 cost €80, while Martingale costs €3,937. The average loss on every euro staked is the same in both cases; what differs is the amount staked and how painful the bad run is. Why 1-2% of your bankroll per bet is a sensible limit is covered in our sports betting basics.

The Kelly criterion

The Kelly criterion works out how much to stake if you have a real edge over the bookmaker, and without such an edge it tells you to stake nothing.

The formula is f = (b × p - q) / b, where b is the odds minus 1, p is your estimate of the outcome's probability and q = 1 - p. Bell Labs researcher John Kelly published it in 1956 in the Bell System Technical Journal. Example: the odds are 2.20 (b = 1.2) and you believe the outcome has a 50% chance. Then f = (1.2 × 0.5 - 0.5) / 1.2 = 8.3% of your bankroll.

The weak spot is p. If the odds are 2.10 and the true probability is exactly what the odds imply (47.6%), the Kelly criterion gives zero: the bet is not worth making. The formula is only as good as your probability estimate, and an inflated estimate makes you stake too much. That is why people who use the criterion usually stake only part of what it suggests, such as half or a quarter (Pinnacle). Mathematician Edward Thorp calculated that at full Kelly the chance that your bankroll halves at some point is 50% (Thorp, 2006).

Value betting: the only approach with an edge in theory

A value bet is a bet whose odds are higher than the outcome's true probability justifies, and only this approach can beat the margin over time.

If the true probability of an outcome is 50% and the bookmaker offers 2.20, every euro staked returns €1.10 on average - 10% ahead. The hard part is knowing the true probability better than the bookmaker, whose prices are shaped by models and the money of the whole market. One honest yardstick is closing line value: if your odds are regularly higher than the price just before kick-off, you are probably finding value (Pinnacle).

There is a flip side. Sportsbooks reserve the right to cut limits for individual customers (OlyBet rules, 4.9; Optibet rules, 2.4.5.3). A group of researchers who bet real money on football using a statistical model made 8.5% in five months, and then the bookmakers restricted their accounts (Kaunitz et al., 2017). So even someone who finds value cannot use it without limits.

The accumulator as a "strategy"

An accumulator multiplies not only the odds but also the margin.

Four selections at 1.90 give combined odds of 13.03. If each outcome has a true 50% chance, the fair combined price would be 16, so on average 81.5% of the stake comes back, not 95% as with a single bet. Every extra selection on the slip adds another margin. How accumulators and system bets are calculated is explained step by step in a separate lesson.

Arbitrage and "sure" tips

Arbitrage works on paper but rarely in practice, and a "sure" tip for money is a sign of fraud.

Arbitrage means betting on every outcome in different places when the gap between prices guarantees a profit whatever the result. In theory the chances exist: a study of Europe's five biggest football leagues, combining a bookmaker bet with a position on a betting exchange, found them in 19.2% of matches (Franck et al., 2013). In practice the profit is small and short-lived, and sportsbook rules threaten it from several sides: bets at obviously wrong odds can be settled at 1.00 (Optibet 4.2.3, OlyBet 8.2), and individual customers' limits can be cut without explanation.

If someone sells a "100% sure" tip or a "fixed match", it is fraud. If a tip really were sure, its author would bet on it instead of selling it to you. Manipulating sports competitions is a criminal offence in Latvia (Criminal Law, Section 212.¹), so a seller of "fixed matches" is most likely just taking your money.

What actually helps

There is no system against the margin, but a few simple things reduce what you pay for your bets.

• Is the same price higher somewhere else? Compare before you bet. If the true probability is 50%, odds of 1.95 cut the average loss from 5% to 2.5% of the stake. Sportsbooks licensed in Latvia include Optibet, Klondaika, LVBET, Betsafe, Tonybet and Laimz; we compare them all on our sportsbooks page.
• Do you really need the accumulator? A single bet costs one margin; an accumulator costs as many as it has selections.
• How big is one stake? A fixed amount or 1-2% of your bankroll, not however "sure" the match feels.
• Do you know how much you have won and lost? Record every bet with its odds and result. After a hundred bets you will see in black and white whether your approach beats the margin, and in most cases the answer will be "no".

Play within your means

Sports betting is entertainment, and you pay for entertainment. Set a budget before you start and stop when it is spent. Our responsible gambling page gathers the tools that help. If you want to ban yourself from gambling, the State Revenue Service (VID) keeps the self-exclusion register, and how to join it is explained in a separate article. If you want to talk it through, the addiction support line of the Centre for Disease Prevention and Control, 67037333, is anonymous (in Latvian). Betting is for people aged 18 and over only.