Correlated Markets in Betting

Correlated markets are selections whose results move together, so multiplying their prices double-counts the shared part. Over 2.5 goals and both-teams-to-score yes are the football example: with independent Poisson λ 1.50 and 1.10, the joint is 39.6% (fair 2.53), not the 25.0% (fair 4.01) you get by multiplying the two singles. A same-game parlay priced as independent legs is therefore too long on paper and too short in the book.

This page is the joint-probability problem. Compounded margin on independent legs is are parlays positive EV. The scoreline engine underneath is Poisson in football betting.

Definition. Two markets are correlated when P(A and B) ≠ P(A) × P(B). Positive correlation: they win together more often than independence says (Over and BTTS; a moneyline and that team's −1.5). Negative correlation: they win together less often (Over and a 0-0 correct score).

Key Takeaways

• Independence is an assumption, not a default. If two legs share a match, check the joint before you multiply.
• Same λ 1.50 / 1.10 as our Poisson article: Over 2.5 is 48.16% (fair 2.08), BTTS yes 51.83% (fair 1.93), Over and BTTS 39.57% (fair 2.53). The independent product is 24.96% (fair 4.01). P(BTTS | Over) = 82.2%.
• Dan Abrams' MEG table on VBF (7 April 2026, re-read 7 November 2026): Tigers ML 45% at +130 is 4.7 bps MEG solo; +1.5 at −140 is 5.7 bps; −1.5 at +195 is 2.7 bps. Full Kelly on all three at once is −5.0 bps. Optimal mix (1% ML + 3% +1.5) is 6.1 bps.
• Books often price a bet builder as a package, not as the product of the singles. Compare the ticket with a joint model, not with 1.91 × 1.91.
• Pinnacle Terminal does not scan same-game parlays. It shows each line's fair price so you can check the legs before you combine them.

What are correlated markets?

Any two prices that share information. Classic pairs:

• Football Over 2.5 and BTTS yes. Most 3-goal games have both teams on the scoresheet.
• A team's moneyline and its −1.5 handicap. The handicap is a stricter version of the same result.
• Home win and home −0.5 Asian handicap. Same winner, different draw rule.
• A player to score and Over 2.5 in the same match.

Uncorrelated: Saturday 15:00 in one league and Saturday 17:30 in another, with no shared players or weather. Those legs can use the independent parlay table.

Football dropping odds already move in packs for this reason: a 1X2 drop drags the Asian handicap and often the totals. The second move is the same information. See football dropping odds.

Independence vs the joint: original Poisson table

Take the same illustrative match as the Poisson article: home λ = 1.50, away λ = 1.10, independent Poisson teams, no Dixon-Coles correction. Scorelines up to 10–10.

The independent parlay of Over 2.5 at fair 2.08 with BTTS at fair 1.93 would be 2.08 × 1.93 = 4.01. The true joint fair price is 2.53. Anyone who multiplies the two singles is paying 4.01 for something worth 2.53.

That gap is the correlation, not extra vig. The vig on a posted parlay is a second problem, on top.

These probabilities are original arithmetic, 7 November 2026, from the same λ as Poisson distribution in football betting. They are not a real fixture.

Same-game moneyline and handicap

The same trap appears on a moneyline plus a run line or Asian handicap. The −1.5 is a subset of the win. Betting both at full size is two bites of one result.

Dan Abrams' MEG article on ValueBetFactory (7 April 2026, re-read 7 November 2026) puts numbers on a Tigers example. VBF's sharp line implied a 45% chance the Tigers win. Three related offers:

MEG is expected bankroll growth in basis points (1 bp = 0.01%), not a profit forecast. The +1.5 has a slightly smaller edge than the moneyline and a higher MEG, because the shorter price lets you stake more for the same Kelly logic.

Stake all three at full Kelly at once: MEG −5.0 bps. The correlation turns three pluses into a minus. Abrams' optimal mix on that table is 1% of bankroll on the moneyline and 3% on +1.5, for 6.1 bps, with 0% on −1.5.

We do not have a Tigers capture of our own, so this page repeats VBF's table as a worked staking example, not as a pick and not as Pinnacle Terminal data.

The same idea in football: home win + home −1.5 is two claims on one attack. Price the joint (home wins by 2+ vs home wins by 1 vs the rest) before you size both.

How do books price correlated tickets?

Two common methods:

1. Product of the singles, then a haircut. The book multiplies the decimal prices and then shortens the ticket because it knows the legs overlap. The independent table in are parlays positive EV is the starting point; the haircut is the correlation charge.
2. A packaged price. Bet builders and scorecasts are often built as one custom market. BetUnfair's niche-markets explainer (16 July 2026, re-read 6 November 2026 on the parlay page) treats a combined bet as a product of listed prices in its 1.50 × 1.80 × 2.00 = 5.40 example. That product is valid only if the legs are independent. For a same-game builder, ask whether the book is selling the joint or the product.

If you cannot write P(A and B) from a model or from the scoreline matrix, you cannot tell whether the ticket is long or short. Do not use the independent 1.91ⁿ table on a same-game slip.

How Pinnacle Terminal handles this

Pinnacle Terminal does not build parlays, bet builders or same-game scans. It streams Pinnacle's prices and limits, shows the fair price (NVP, power devig by default) on each line, and grades every pre-match move 0–10 (Big move, Sharp drop, Limit-backed, Downtrend, Rebound, Standard). Live in-play football drops are shown and filtered, without a grade. Use the fair price on each leg, then compute the joint yourself if you combine them. The terminal does not accept bets and is not affiliated with Pinnacle. Product: Pinnacle dropping odds. Plans: pricing.

FAQ

What are correlated markets in betting?

Selections that win or lose together more (or less) often than if they were independent. Over 2.5 and BTTS yes in the same football match is the usual pair.

Why can't I multiply Over 2.5 and BTTS?

Because most games with three or more goals have both teams scoring. With λ 1.50 / 1.10 the joint is 39.6%, not the 25.0% you get from 48.2% × 51.8%.

Are same-game parlays +EV?

Only if the joint is mispriced in your favour after the book's correlation charge. Independent +EV legs in different matches are a different calculation; see are parlays positive EV.

Can I bet a moneyline and a −1.5 on the same team?

You can. They are correlated. Sizing both at full Kelly overbets the shared result. Abrams' Tigers table shows all-three full Kelly at −5.0 bps MEG.

Does Pinnacle Terminal find correlated parlays?

No. It shows each Pinnacle line and its fair price. Combining them is your arithmetic.

Sources

• Joint Over-and-BTTS table: original Poisson arithmetic, 7 November 2026, same λ 1.50 / 1.10 as Poisson distribution in football betting.
• Dan Abrams, "Maximum Expected Growth explained", ValueBetFactory, 7 April 2026 (re-read 7 November 2026): Tigers MEG table (4.7 / 5.7 / 2.7 bps solo; −5.0 bps all-three full Kelly; 6.1 bps optimal 1% + 3%).
• Compounded-margin independent table: are parlays positive EV, 6 November 2026.
• BetUnfair, "Niche Markets: Specials, Scorecasts and Combined Bets Explained", 16 July 2026, cited 6 November 2026 for the 1.50 × 1.80 × 2.00 = 5.40 product of listed prices.

Try it

Check each Pinnacle line's fair price before you combine anything: start the Pinnacle Terminal 7-day free trial. Nothing is charged before day 8. Plans: pricing.

18+ · Bet responsibly. Only bet money you can afford to lose.