Value Betting & Odds Picking

Expected value is the average result of a bet if you could repeat it endlessly. It combines the chance of winning with what you win, and what you lose otherwise.

The formula is: EV = (probability of winning x profit if you win) - (probability of losing x stake). Using decimal odds, a simpler version is EV per unit staked = (true probability x decimal odds) - 1.

Here is an example. Suppose you believe a team has a 50% chance of winning and a bookmaker offers 2.10. EV per unit = (0.50 x 2.10) - 1 = 0.05, so +5%. On a 100 unit bet, you would expect to make 5 units on average. But you would not make 5 units on that bet. You would either win 110 or lose 100. The 5% only shows up as an average over many repetitions.

If the true chance were 45% instead, the same bet would be (0.45 x 2.10) - 1 = -5.5%, a losing proposition in the long run. The odds did not change; your estimate of the probability did. Everything in value betting depends on how good that probability estimate is.

A bet has value when the bookmaker's implied probability, which is 1 divided by the decimal odds, is lower than the true probability. At 2.10, the implied probability is 47.6%. If the real chance is 50%, there is value. The difficulty is that nobody knows the true probability. You can only estimate it.

There are two main approaches. The first is to build your own model from data such as team form, injuries, schedule and ratings. That is hard, time-consuming and needs testing against results over a large sample. The second, much more accessible approach is to use the market itself as the estimate, which is covered in the next section.

In either case, be sceptical of your own edge. If you think you have found a 15% advantage on a major market, it is far more likely that you are missing information, such as a late injury, than that the market is wrong. Genuine edges tend to be small, a few percent, and they are most likely in less efficient places such as lower leagues, niche markets and soft bookmakers' slow-to-update lines.

Not all bookmakers are equal. Sharp books, with Pinnacle the best-known example, accept large bets from skilled players, run low margins and move their prices quickly in response to informed money. Their closing odds are widely regarded as among the most accurate public estimates of true probability. Soft books, which cater mostly to recreational players, are slower to adjust and carry bigger margins, so their prices sometimes sit above the sharp market.

The method is to take the sharp book's odds, remove its margin to get a no-vig price, and compare that with soft-book odds. For example, if a sharp book prices a two-way market at 1.95 and 1.95, the implied probabilities are 51.3% each, adding up to 102.6%. Removing the margin gives 50% each, so the fair price is 2.00. If a different bookmaker offers 2.10 on one side, the value is 2.10 / 2.00 - 1 = 5%.

There are several ways to remove the margin, and results differ slightly in lopsided markets. Treat the no-vig figure as a good estimate, not as truth. Odds comparison sites and screeners can automate this, but check that the price is still available and that the markets are defined identically before you bet. Remember that a sharp reference can be wrong too, and that a rapidly moving sharp line often means the soft price is simply stale for a good reason.

Your short-term results are mostly luck. Closing-line value, or CLV, is a far better indicator of whether you are actually finding good prices. It compares the odds you took with the odds available just before the event starts, which are usually the most efficient prices of all, because they absorb all the information and money that arrived in the meantime.

If you bet at 2.10 and the sharp market closes at 2.00, you beat the closing line. Consistently getting better prices than the close is strong evidence of a real edge, even if your actual profit and loss looks poor over the same period. Equally, a winning streak achieved while regularly taking worse odds than the close is likely to be luck.

To use CLV, record the odds you took and the closing odds from a sharp book for each bet, convert both to no-vig probabilities, and track the average difference. A few hundred bets is the minimum for a rough signal. CLV is not perfect, since markets can be less efficient in obscure leagues, but it is the best feedback loop most bettors have.

Even with a genuine edge, staking too much will ruin you. Bankroll management is about choosing bet sizes that survive the swings. The simplest method is flat staking, risking a fixed small percentage, often 1% to 2%, of your bankroll on each bet.

The Kelly criterion is a formula for the stake that maximises long-run growth. For a bet with decimal odds, Kelly fraction = (b x p - q) / b, where b is the decimal odds minus 1, p is your estimated win probability and q is 1 - p. In our example, b = 1.10, p = 0.50 and q = 0.50, so the Kelly fraction = (0.55 - 0.50) / 1.10 = 0.045, or about 4.5% of bankroll.

Full Kelly is aggressive and assumes your probability estimate is exactly right, which it never is. If you overestimate your edge, full Kelly can lead to heavy losses. Most practitioners use a fraction, such as one quarter or one half Kelly, which here would be roughly 1.1% to 2.3% of bankroll. That sacrifices some growth for much smoother results. Never stake money you need for rent, bills or anything else essential, and set a bankroll limit that you treat as entertainment spending.

Arbitrage covers all outcomes, so the result is mostly fixed once both legs are placed. Value betting covers nothing. Each bet is a one-sided wager that you expect, but cannot be sure, to win on average.

Here is what that means in numbers. Take bets at even-money-like odds with a 5% edge. After 100 bets, the expected profit is about 5 units of stake, but the standard deviation of the outcome is around 10 units. That puts the chance of being behind after 100 bets at roughly one in three, even though every bet was positive expected value. Losing runs of 20 or more bets are completely normal for bets at longer odds, and drawdowns lasting months happen to genuinely skilled bettors.

This variance has practical consequences. It is psychologically hard to keep following a process through a losing streak, and it is tempting to raise stakes to chase losses, which is the fastest route to ruin. It also makes it hard to know whether you actually have an edge, which is why CLV and a large sample size matter. Bookmakers will also limit successful value bettors, just as they limit arbers, so even a real edge can have a shelf life.

Crypto sportsbooks can suit value bettors thanks to fast deposits and withdrawals and, at some operators, higher limits and slower restriction of winning accounts. But policies differ, and many of these books use sharp-derived pricing, which leaves less value on offer. Check each operator's licence, maximum payouts and rules on voided bets, and use stablecoins if you want to avoid mixing coin price swings with your betting results.

Keep a detailed record of every bet, including odds, stake, closing odds and reasoning. Set deposit and loss limits in advance, take regular breaks, and never treat betting as a source of income you rely on. If gambling is causing you stress or financial harm, stop and contact a support service in your country or use self-exclusion tools. Value betting is a long-term, probability-based discipline, not a path to guaranteed winnings.