Line shopping and exact market equivalence
A better price only counts if it is for exactly the same contract. Learn to check period, settlement rule, line and voids before comparing prices — and why Value 7's no-push rule changes the answer.
Learning objectives
- List the contract features (event, period, settlement rule, line, player identity, void rules) that must match before two prices are comparable.
- Show that Asian handicap −0.5 and the 1X2 home win are the same bet, and that draw-no-bet equals Asian handicap 0 under general rules.
- Compute fair odds and EV for a whole-number line with and without pushes.
- Explain how Value 7's win-or-lose settlement changes the fair price of whole lines and handicap 0.
Line shopping compares contracts, not numbers
Line shopping means taking the best available price for a bet you already want. It is the least controversial way to improve results: the same probability at a higher price is worth more. But it only works when the two prices refer to exactly the same contract. A higher number attached to a different contract is not a better price; it is a different bet.
- Event: same fixture, same competition and same date (watch for re-arranged fixtures).
- Period: 90 minutes plus stoppage time, or including extra time and penalties, or a single half.
- Settlement rule: what happens on a draw, on a whole-number line or when the result lands exactly on the handicap.
- Line: −0.5, −1 and −1.5 are three different contracts; quarter lines split the stake.
- Player identity: the right player (namesakes exist) and the right team.
- Voids: postponement, abandonment and non-starter rules differ between operators.
Period: 90 minutes versus extra time
Match-result markets normally settle on regular time. Cup markets such as 'to qualify' include extra time and penalties, so they are much more likely to land for the stronger side. Using a simple independent Poisson model with hypothetical rates λ = 1.5 (home) and μ = 1.1 (away), the home side wins in 90 minutes with probability 46.42%. If we assume extra time runs at a third of those rates and penalties are a coin flip, the home side qualifies with probability 60.51% (fair odds 1.653 versus 2.154).
Those are different contracts. A 'to qualify' quote of 1.70 is +2.87% EV on its own probability, but if you mistook it for the 90-minute home win it would be −21.08%. The extra-time assumptions above are deliberately simple; the point is the size of the gap, not the precise figure.
Settlement rules: when two markets are the same bet
| Market A | Market B | Same contract? |
|---|---|---|
| Asian handicap home −0.5 | 1X2 home win | Yes: both need a home win |
| Asian handicap home 0 (general rules) | Draw no bet, home | Yes: a draw refunds the stake |
| Asian handicap home +0.5 | Double chance home or draw | Yes |
| Asian handicap home −1 (general rules) | 1X2 home win | No: a one-goal win is a push, so you need a two-goal win to profit |
| Total goals Over 2 (general rules) | Over 2.5 | No: exactly two goals refunds the Over 2 stake |
| Quarter line −0.75 | Half on −0.5, half on −1 | Yes, by definition |
With a push (refund), the fair price depends only on the win and lose probabilities, because the push returns your stake and earns nothing:
EV = P(win) × (o − 1) − P(lose) fair o = 1 + P(lose) / P(win)Value 7 settles differently: no pushes
The difference is large. With λ = 1.5 and μ = 1.1, Asian handicap 0 for the home side wins 46.42%, pushes 25.77% and loses 27.81%. Under general rules its fair price is 1.599; under Value 7 rules a draw loses, so handicap 0 becomes a plain home win with fair price 2.154. A hypothetical quote of 1.62 is +0.97% EV with refunds and −24.79% without them. Likewise Over 2 at 1.75 is +9.38% with pushes and −15.73% when exactly two goals loses.
Quarter lines, players and a canonical market key
Under general bookmaker conventions a quarter line splits the stake: home −0.75 is half on −0.5 and half on −1, so a one-goal home win pays half the stake at the quoted odds and refunds the other half. Its fair price must be built from both halves, never compared directly with either neighbouring line. Value 7 picks never split or refund, so treat any quarter-line quote from elsewhere as a different product.
Player markets add identity and participation risk: two players can share a surname, a player can move clubs mid-season, and operators differ on whether a player who never comes on is void. Value 7 settles such a pick as a loss.
In data, the safest habit is a canonical market key for every quote: event identifier, period, market type, line, side and settlement ruleset. Two prices are comparable only when their keys match exactly; anything else must be converted to a probability first.
Worked example
Worked example · The 2.30 that was not a home win
Hypothetical Harbour City v Northvale. For illustration, goals follow independent Poisson rates λ = 1.5 (home) and μ = 1.1 (away). Book X quotes 1X2 Harbour City 2.10. Book Y quotes Asian handicap Harbour City −0.5 at 2.12. Book Z's handicap tab lists 'Harbour City −1' at 2.30, which a hurried reader takes for a better home-win price.
- Score-model probabilities
a Poisson score grid (λ = 1.5, μ = 1.1) → home / draw / away probabilitiesHome 46.42%, draw 25.77%, away 27.81% (fair home odds 2.154) - Is AH −0.5 the same contract as 1X2 home?
Home −0.5 on the score grid: win 46.42%, push 0%, lose 53.58%Yes. Same probability, so 2.12 is simply 0.95% more than 2.10 - EV of the equivalent prices
0.4642 × 2.10 − 1 and 0.4642 × 2.12 − 1−2.51% at 2.10 and −1.58% at 2.12: the shopping gain is real, but both remain negative under this model - What AH −1 actually is
Home −1 on the score gridWin (by 2+) 23.19%, push (by exactly 1) 23.24%, lose 53.58% - Fair price of AH −1
General: 1 + 0.5358 / 0.2319. Value 7: 1 / 0.23193.311 with pushes; 4.313 under Value 7 rules (a one-goal win loses) - EV of the 2.30 quote
General: 0.2319 × 1.30 − 0.5358. Value 7: 0.2319 × 2.30 − 1−23.43% with pushes; −46.67% under Value 7 rules - Bonus check: draw no bet at 1.62
AH 0: win 46.42%, push 25.77%, lose 27.81%+0.97% with refunds; −24.79% under Value 7 rules, where AH 0 is just a home win
The 'better' 2.30 belongs to a contract that wins less than a quarter of the time; it is far worse than the 2.10 home price. The honest line-shopping gain is 2.10 → 2.12 on equivalent markets, and under this model even that is still negative, so the respectable conclusion is no action. Always map each quote to a probability from one score model before comparing numbers.
Practice · Price the contract, not the label
In the Fair Odds and EV lab, enter odds 2.30 with probability 0.2319 (Harbour City −1 under Value 7 rules) and a range of 0.21–0.25. Then change to odds 2.10 with probability 0.4642 and a range of 0.44–0.49 (the 1X2 home win). Compare the two verdicts.
Lab 1 · Fair odds and expected value
Synthetic · deterministicNegative across your range. Even the high end of your range is below break-even. There is no advantage under any probability you consider plausible: no action.
Show data table
| True probability (%) | Expected ROI |
|---|---|
| 11% | -75% |
| 12% | -72% |
| 13% | -70% |
| 14% | -68% |
| 15% | -66% |
| 16% | -63% |
| 17% | -61% |
| 18% | -59% |
| 19% | -56% |
| 20% | -54% |
| 21% | -52% |
| 22% | -49% |
| 23% | -47% |
| 24% | -45% |
| 25% | -43% |
| 26% | -40% |
| 27% | -38% |
| 28% | -36% |
| 29% | -33% |
| 30% | -31% |
| 31% | -29% |
| 32% | -26% |
| 33% | -24% |
| 34% | -22% |
| 35% | -20% |
Educational calculator. Uses only the numbers you enter or seeded synthetic data — no live prices, no staking, no real money.
Show what you should find
At 2.30 the break-even probability is 43.48%; with p = 23.19% the EV is −46.66%, ranging from −51.70% to −42.50%: negative across the whole range. At 2.10 the break-even is 47.62%; EV is −2.52% at the estimate, −7.60% at 0.44 and +2.90% at 0.49, so the verdict depends on the estimate. The higher number is the far worse bet because it is a different contract.
Common mistakes
- ✗ Comparing a 'to qualify' price with a 90-minute match-result price.Qualification includes extra time and penalties. In the example it is 60.51% versus 46.42% for the 90-minute home win.
- ✗ Assuming Asian handicap −1 is a home win at a better price.AH −1 needs a two-goal win to profit; a one-goal win is a push under general rules and a loss in Value 7.
- ✗ Using general push rules when pricing Value 7 picks.Value 7 has no pushes: use fair o = 1 / P(win). AH 0 fair odds jump from 1.599 to 2.154 in the example.
- ✗ Ignoring void and non-starter rules on player markets.Many bookmakers void a bet on a player who takes no part; in Value 7 that pick loses. The two contracts are priced differently.
When it fails
Supporting research
- Maher (1982): Modelling association football scoresStatistica Neerlandica · Peer-reviewed empiricalTeam-strength Poisson models give a coherent score distribution, which lets us price different contracts (1X2, handicaps, totals) from one set of probabilities.
- Dixon & Coles (1997): Modelling Association Football Scores and Inefficiencies in the Football Betting MarketJournal of the Royal Statistical Society: Series C (Applied Statistics) · Peer-reviewed empiricalScore-matrix models as a basis for comparing related markets consistently; lesson 10 adds their low-score correction.
Check your understanding
Practice only: saved on this device, no XP and no account record.