How to read odds — and find the value the market mispriced

The conversion that changes everything

Implied probability = 1 / decimal odds. A price of 2.00 implies a 50.0% chance. A price of 1.50 implies 66.7%. A price of 4.00 implies 25.0%. That is the entire formula. Once you internalise it, every coupon turns from a row of numbers into a row of forecasts you can agree or disagree with.

The numbers worth memorising

• 1.50 = 66.7% — the "strong favourite" line. Disagreement here is rare and usually expensive.
• 1.91 = 52.4% — the standard "coin-flip with vig" price both sides of a 50/50 market.
• 2.00 = 50.0% — true coin flip, only seen at the sharpest books.
• 3.00 = 33.3% — your number must clear ~36% to bet it (after 4–5% vig).
• 5.00 = 20.0% — niche / underdog territory; the soft books are softest here.

Why fractional and American odds aren't worth your time

Decimal is the only format that lets you do math at a glance. Fractional (5/2) and American (+250) require an extra mental step before you can ask "is this a fair price?". If your default book quotes in another format, switch the display setting before you do anything else — small friction, large compounding payoff.

How to spot the vig in any market

Add the implied probabilities of every outcome. On a fair market they should sum to exactly 100%. They never do. The excess is the vig. On a two-way market priced 1.91 / 1.91 the implied probabilities are 52.36% + 52.36% = 104.71%, so the vig is 4.71%. That 4.71% is what you pay the book on every bet, win or lose.

Vig by market type — what to expect

• Soft-book pre-match moneyline: 4–6%. The standard tax.
• Sharp-book pre-match moneyline (Pinnacle, sharp Asian books): 2–3%. Half the cost, identical work.
• Player props and exotics: 6–10%. The book is slower here — sometimes the vig is the only thing protecting them, and sometimes it isn't enough (see line shopping).
• Live in-play: 5–8%. Higher because the book is buying time to recompute.
• Three-way markets (1X2 football): typically 5–8% combined. Always strip per-outcome, never assume even.

The math of paying 5% vs 2.5%

A bettor who places 1,000 bets a year at 1u flat, hitting break-even on selection, loses 1,000 × stake × vig per year. At 5% vig that is 50u. At 2.5% it is 25u. Same bettor, same skill, same record — the cost of the venue cuts your loss in half. The first decision in finding value is "where do I bet", and the answer is rarely your most familiar app.

The two-way no-vig formula

For a two-way market: no-vig probability A = implied A / (implied A + implied B). Example: prices 1.83 / 2.00. Implied: 54.64% / 50.00%, sum 104.64%. No-vig A = 54.64 / 104.64 = 52.22%. The book's honest forecast is 52.22% / 47.78%, the fair price for A is 1 / 0.5222 = 1.92. The 1.83 you were quoted is the fair price plus the vig — the bet has zero edge if your number agrees with theirs.

The three-way no-vig formula

Same idea: no-vig P(outcome) = implied(outcome) / sum(all implied). On a 1X2 football market priced 2.10 / 3.40 / 3.50, the implied probabilities are 47.62 + 29.41 + 28.57 = 105.60%, vig 5.60%. No-vig: 45.10% / 27.85% / 27.05%. Fair odds: 2.22 / 3.59 / 3.70. If your model says home win = 50%, you have a real 5pp edge — and a value bet at 2.10.

When the no-vig method breaks (and what to do)

• On heavily lopsided markets (e.g. 1.10 / 9.00) the proportional no-vig method slightly over-estimates the favourite. Use power-method or Shin's method on these if you want precision; for everyday work the proportional method is within 0.5%.
• On Asian handicaps, treat each leg as its own two-way market. Strip the vig leg by leg, not on the spread as a whole.
• On player props with hidden alternates (e.g. over/under shoulders), use the centre line's no-vig — the alts are usually priced from it.

The 2-percentage-point rule

Edges below 2 percentage points (your number vs. no-vig) are noise. You can't reliably distinguish a true 2pp edge from estimation error in your own probability — and once you account for the bid-ask spread when the book moves, you've given the edge back. Hunt edges at +3pp or larger; below that, sit out and save the bankroll for a better disagreement.

The expected value (EV) formula

EV per 1u = (your probability × (decimal odds − 1)) − (1 − your probability). Worked example: odds 2.10, your honest probability 50%. EV = (0.50 × 1.10) − 0.50 = +0.05u per bet. A 5% positive EV repeated 1,000 times is +50u — and the only reason it shows up is because your probability disagreed with the no-vig probability of 47.6% by ~2.4pp.

What looks like value but isn't

• "They're due." Markets don't remember; runs don't mean-revert on demand. The price already includes whatever recent performance is informative.
• "The line moved my way after I bet." That is closing-line value, not value at the time of the bet. Useful diagnostic, not a justification.
• "It pays 5.00, the upside is huge." Upside is irrelevant; the only question is whether 5.00 (= 20% implied) overpays your honest probability. If you say 18%, it's a losing bet, no matter how exciting it is.
• "I have inside info." If the info is accurate and tradable, the line will already have moved before you can act. If it hasn't, the info is either wrong or already priced.

How CLV is measured

You bet at 2.10. The market closes at 2.00. Your CLV is +5% — the no-vig closing probability is higher for your side than the no-vig probability when you bet, meaning the market moved toward your bet. Over hundreds of bets, average CLV is the strongest known predictor of long-term betting profit, more reliable than current ROI, hit rate, or recent results.

Why books care about CLV more than your record

A bettor who is +12u after 100 bets but consistently negative-CLV is on a hot streak; the book leaves them alone, knowing variance will take care of it. A bettor who is −3u after 100 bets but +1.5% CLV is a long-term problem; the book limits their account before they get rich. Public bettor outrage about "limits after winning" almost always misses the real reason: it is CLV, not P&L.

The three rules of CLV-aware betting

• Aim for average CLV of +1% or better. Held even, alone, this beats every soft book over a season.
• Bet early into the market and the closing price has time to move toward you (or away from you). Late bets near tipoff give CLV less room to express itself.
• A single bet's CLV is noise. CLV becomes signal at 50+ bets and decisive at 200+.

The half-percent math

Line shopping that finds an average +0.4% better price on every bet compounds to +40 units per 100 bets at 1u flat. Over a 1,000-bet season, that's +400u for clicking three tabs instead of one. It costs nothing, requires no extra information, and most bettors don't bother.

How to do it without dying to friction

• Maintain accounts at 2–3 books with consistently competitive lines and reliable withdrawals.
• For every bet you intend to place, check the price at all three before you click. Five seconds of work.
• If the best price and your default book's price differ by less than 0.5%, save the friction and stay with default. Above 0.5%, switch books.
• Track which book wins the comparison most often per market — over time you'll learn that book A is sharpest on NBA, B on European football, C on niche markets. Default to the right book per sport.

Trap 1 — Anchoring on the opening line

The first price you see on a market becomes the price you "remember" as fair. When the line moves against you, you treat the new price as worse, even though it may be more accurate. Cure: ignore the open; evaluate every price against your no-vig calculation, not against memory.

Trap 2 — Public bias and the favourite tax

Public money piles onto favourites and household-name teams. The book moves the line to balance their risk, which means favourites are systematically over-priced and underdogs under-priced on popular markets. Real-Madrid-1.50 rarely has value; the underdog opponent at 7.00 sometimes does.

Trap 3 — Recency bias

A team's last three results dominate your estimate even though they are statistical noise. The market knows this and prices the next match against a much larger sample. When you catch yourself saying "they're on fire, must be 1.40", check what they were 1.40 against three weeks ago — and how that ended.

Trap 4 — Chasing losses into worse prices

After a losing run, you accept worse value to "get back to even". The next bets are statistically your worst bets of the week. If your chase bet would have failed your normal value criterion, it is not a bet — it is a coping mechanism with money attached.

Trap 5 — Counting wins, not CLV

You log winners, dismiss losers, build a mental highlight reel, and feel sharper than the market. CLV does not care about your highlight reel. The honest scoreboard is the spreadsheet — and an honest spreadsheet usually finds you flat or down where memory says you're up.