Are 2.2 Odds Good Value at a 50 Percent Chance?

Decimal odds of 2.2 return 2.2 units for every unit staked when the bet wins, a net profit of 1.2 units. If the true probability of winning is 50 percent, the expected value per unit staked is:

0.5 times 1.2, minus 0.5 times 1, which is 0.6 minus 0.5, or plus 0.10.

Equivalently, probability times decimal odds minus one: 0.5 times 2.2 is 1.10, minus one is 0.10. Either way, on the stated assumption the position returns 10 percent of stake on average. In other formats, 2.2 is American +120 or fractional 6/5.

The break-even probability is the reciprocal of the price: 1 divided by 2.2 is about 45.5 percent. Any true probability above that makes the price positive in expectation; any below makes it negative. At 50 percent the gap is about 4.5 percentage points, which is large by the standards of liquid markets.

The calculation is exact. The input is not. "50 percent" is an estimate, and the honest question is how far off it could be.

Where does it come from? If it is the de-margined price of a sharp market, it has a known and testable source. If it comes from a model, the model's calibration decides how much it can be trusted. If it is a coin-flip intuition about two evenly matched teams, it has no source at all, and the 10 percent figure inherits that.

How precise can it be? Suppose the estimate comes from historical frequency: in 100 comparable past situations, the outcome occurred 50 times. The standard error of that proportion is the square root of 0.5 times 0.5 divided by 100, which is 5 percentage points. A rough 95 percent interval runs from about 40 to 60 percent, and the break-even of 45.5 sits comfortably inside it. On that evidence alone, the bet's expected value could be anywhere from clearly negative to strongly positive.

Is the market telling you something? A two-way market where one side is offered at 2.2 might price the other side at 1.70. Those imply about 45.5 and 58.8 percent, summing to about 104.3 percent; removing the margin proportionally leaves roughly 43.6 percent for the 2.2 side. If a liquid market thinks the true figure is near 44 percent and your estimate says 50, the burden of proof is on your estimate. Six points of disagreement with a sharp market is far more often an estimation error than a market error.

Recompute the expected value at decimal 2.2 across nearby probabilities (illustrative):

• At 52 percent: plus 14.4 percent.
• At 50 percent: plus 10 percent.
• At 48 percent: plus 5.6 percent.
• At 47 percent: plus 3.4 percent.
• At 45.5 percent: about zero.
• At 44 percent: minus 3.2 percent.

Every percentage point of probability is worth 2.2 percentage points of expected value at this price. That leverage is why a confident point estimate is less informative than the range: if the sign of the result flips within the range you consider plausible, nothing has been established yet.

Even if 50 percent is exactly right, results are noisy. Each unit bet returns either plus 1.2 or minus 1, so the standard deviation per bet is about 1.1 units against a mean of 0.1. The noise is eleven times the signal on any single bet.

Using the binomial distribution with a true 50 percent and a price of 2.2 (illustrative):

• After 50 bets, the chance of being behind is about 24 percent.
• After 100 bets, about 18 percent.
• After 500 bets, about 2 percent.
• After 1,000 bets, about 0.2 percent.

So a genuine plus 10 percent position loses money over a hundred repetitions roughly one time in five or six. And because the estimate itself is uncertain, a short losing run cannot tell you whether you are in that unlucky fraction or whether the true probability was 44 percent all along.

The Kelly criterion gives the growth-optimal stake for a known edge: f equals b times p minus q, divided by b, where b is 1.2, p is 0.5 and q is 0.5. That is 0.1 divided by 1.2, about 8.3 percent of bankroll at full Kelly. Because the probability is uncertain and full Kelly assumes it is exact, fractional Kelly (a quarter or a half) is the common practice, giving roughly 2 to 4 percent. If the true probability were 45.5 percent, the correct Kelly stake would be zero.

Are 2.2 odds good value at 50 percent? Arithmetically, yes: plus 10 percent per unit. Practically, the answer is only as good as the 50 percent. Before trusting it, write down its source, compute the range of plausible probabilities, compare it with a de-margined sharp price, and track whether prices like it beat the closing line over many observations.

This page works through a calculation and is not betting advice. Set limits before betting and keep to them.