Learn

The Linked Legs Test: same game parlay correlation and the math to price any two legs

Same game parlay correlation is why multiplying the legs gives the wrong price: in a hypothetical where two 50% outcomes happen together 40% of the time, multiplying says +300 but the fair price is about +150. The Linked Legs Test is five steps to check any pair: do the legs move together, which way, what is the joint chance, what fair price does that give, and how does the book's price compare. 21+, nothing is guaranteed.

Updated October 21, 2026 · 6 min read

The Linked Legs Test in one example

The video behind this guide makes one point with one hypothetical, and this page opens with the same example. Two outcomes in the same game are each a 50% chance, but because they tend to happen together, they both happen 40% of the time. It is a hypothetical built on the same idea as OddsPapi's correlation examples.

  • Multiplied: 50% x 50% = 25%, which is a fair price of about +300.
  • Linked: the real joint chance is 40%, which is a fair price of about +150, half the profit (fair price = 1 / chance, the Wizard of Odds rule).

Multiplying the legs only works when they don't affect each other. OddsPapi's walkthrough of same-game parlay correlation puts it plainly: legs on the same match are not independent events, and multiplying their prices as if they were two separate coin flips misprices the ticket. That is why books price same-game parlays on a separate engine that models the link, then add their margin on top.

Why the multiplied price is wrong

A normal parlay multiplies leg probabilities because the legs are independent, the rule Wizard of Odds uses for regular parlays. In one game, results are tangled. If a soccer match has lots of goals, both teams scoring becomes more likely. If a football favorite covers a big spread, it probably scored a lot of points, which pushes the total toward the over.

OddsPapi fitted a goals model to one real fixture and compared the multiplied price with a correlation-aware fair price. For Over 2.5 goals plus both teams to score, the multiplied decimal price was 3.24 and the fair price was 2.42, so the multiplied number was about a third too long. For Under 2.5 plus both teams to score, the opposite happened: 5.77 multiplied vs 14.31 fair, because those two outcomes push against each other. Those figures are from the same OddsPapi analysis, for one match.

Positive links make the multiplied price too generous. Negative links make it too stingy. Either way, multiplying gives you the wrong fair price.

The Linked Legs Test, step by step

  1. Ask whether the legs move together. Do they depend on the same thing, like total scoring, one team's pace, or one player's role? If yes, the legs are linked and multiplying is off the table.
  2. Name the direction. Legs that rise together (lots of goals and both teams scoring) are positively linked. Legs that work against each other (few goals and both teams scoring) are negatively linked.
  3. Estimate the joint chance. This is the hard part. You need the chance that both legs happen together, from a model or from how often the pair happened in comparable games. OpticOdds warns that history can't capture injuries or coaching changes and that correlated prices need updating in real time, so treat any estimate as rough.
  4. Turn it into a fair price. Fair American price for a joint chance p = (1 / p - 1) x 100, shown as a plus price. Our fair odds calculator does the conversion.
  5. Compare with the book's price. If the book's same-game price is shorter than your fair price, the gap is the margin you'd pay. The expected value calculator shows it per $100.

Joint chance to fair price, two 50% legs

For two legs that are each a 50% chance, the joint chance can sit anywhere from 0% to 50% depending on the link. Here is the fair price at each level (our arithmetic with the formula in step 4, the one-over-the-chance rule from Wizard of Odds).

Joint chance both legs win Link Fair price
15% negative +567
20% negative +400
25% none (independent) +300
30% positive +233
35% positive +186
40% positive +150
45% positive +122
50% perfectly linked +100

A worked comparison, hypothetical: you estimate the joint chance at 40%, so the fair price is +150. A book offers +135 on the pair. Your expected return is 0.40 x 2.35 = 0.94 per dollar, about 6% below fair, which is the margin on that ticket. The fair-price formula is the same one-over-the-chance rule Wizard of Odds applies to parlays.

What the research says about linked parlays

Linked parlays are not new. A study in The Journal of Prediction Markets, Correlated parlay betting by Davis, Dawson and Krieger, looked at college football from 2005 to 2015 and found that games where the favorite covered were more likely to go over. The authors noted that many books prohibit some or all same-game parlays and argued books had generally been too conservative in refusing them. Wikipedia's parlay entry also notes traditional books often refused correlated parlays, and that newer apps price the correlation instead.

The modern answer is the separate pricing engine. The book models the link, as OddsPapi describes, and adds margin, so the linked price you see is usually below a correlation-aware fair price, not above it.

What to watch out for

The book is usually right about the link. You'd think a short same-game price means the book is shortchanging you, but most of the gap is the link itself, priced correctly, with the margin on top. Finding real value means your joint-chance estimate has to beat the book's model.

Three or more legs compound the problem. Every extra linked leg adds another relationship to estimate. Keep the test to pairs you understand.

Estimates drift. Lineups, weather and in-game news change the link. A pair that moved together last season may not today, which is the warning OpticOdds makes about historical data.

How to tell it's working

You'll know the test is working when you stop being surprised by same-game prices. Before you add a leg, you can say whether it moves with the others and roughly what the fair price of the pair is. Track your same-game tickets in a bet tracker with your fair estimate next to the price you took. If the price you take is regularly shorter than your fair estimate, those slips are costing you margin, and fewer, simpler legs is the honest fix.

21+. For informational purposes only. Nothing is guaranteed. If you or someone you know has a gambling problem, call 1-800-GAMBLER or visit the National Council on Problem Gambling.

Questions people ask

Why do same game parlays pay less than regular parlays?

Legs in one game are usually linked, so the chance they all win is higher than multiplying suggests. Books price that link on a separate engine and then add margin.

Can I just multiply the odds of a same game parlay?

No. Multiplying assumes the legs are independent. In one analysis, Over 2.5 goals plus both teams to score multiplied to 3.24 while the fair price was 2.42.

Are negatively linked legs a good deal?

They can look short when multiplied, because the true joint chance is lower than the product. The fair price is longer, but the book's model prices that too.

How do I estimate the joint chance?

From a model or from how often the pair happened in comparable games. Treat it as rough, because injuries, lineups and news change the link.

Sources

  1. Same-Game Parlay Correlation: Why Multiplying Odds Fails, OddsPapi
  2. Correlated parlay betting, The Journal of Prediction Markets (UWF IRCommons)
  3. Correlation in Same Game Parlays, OpticOdds
  4. Same-Game Parlays: The Mathematics of Correlation, Wizard of Odds
  5. Parlay, Wikipedia
  6. National Council on Problem Gambling, NCPG