Why I Stopped Betting Multis, Even When I Had an Edge on Every Leg
Multis feel like the best way to turn small bets into big wins. The maths says the opposite. Here's what happened when I ran the numbers on my own.
Tom Rafferty
Columnist
10 min read·Published 29 July 2025
I used to put a multi on every weekend. Four legs, usually - two AFL games, one NRL, one soccer - at roughly $15 odds combined. Sometimes they hit, most of the time they didn't, and I always felt like I was one good weekend from a proper windfall.
Then I actually tracked them for six months. The answer was that I'd lost $3,400 on multis alone over that period, and the bets I'd have made separately on the same games would have turned a small profit. I stopped betting multis that week and haven't placed one since. This is the explanation of why.
The thing about compounding vig
Every bet at an AU bookmaker carries a margin, roughly 4-6% on a typical two-way market. When you bet a single leg, you pay that margin once. When you bet a multi, you pay it on every leg.
A four-leg multi at AU bookmakers with each leg at a 5% vig compounds to: (1 - 0.05)^4 = 0.814, or an effective 18.6% vig on the combined bet. A six-leg multi compounds to 26.5%. A ten-leg multi compounds to 40%.
That's the first problem. Even if every leg is fairly priced, the multi structure itself is a compounding tax on your bankroll. Ten-leg multis at most AU bookmakers are effectively a 40% house-edge product, comparable to keno.
Say you think you have +3% EV on each of four independent legs. Your single-bet expected value per dollar staked is 3%. Easy.
For a four-leg multi, expected value compounds multiplicatively in odds space, not in edge space. If each leg has a true probability of, say, 55% (with a 3% edge against the bookmaker's 52% implied probability), your multi's true probability of winning is 0.55^4 = 9.15%. The bookmaker's implied probability on the multi is 0.52^4 = 7.31%. Your multi EV is 9.15% / 7.31% - 1 = 25.2% expected value.
That sounds amazing. It's also misleading, because we haven't accounted for the multi's compounded vig yet. After the bookmaker's margin, the actual payout odds compound at the margin-inclusive price, not the true-probability price. Once you do that calculation, the 25% EV on the multi collapses to something closer to 4-6% - roughly the same as betting each leg separately.
Except, and this is the real kicker, your variance on the multi is enormous. The multi either hits big or loses the whole stake. Placing the four legs separately produces the same EV with a fraction of the variance. Kelly sizing on a multi would return a bet size roughly 1/10 the size of the equivalent separate-leg sizing. You can't deploy capital efficiently in multi form.
Same-game multis are worse
Regular multis are bad. Same-game multis (SGMs) are substantially worse.
The issue is correlation. When you SGM "Collingwood to win + Collingwood to score over 90 points + Nick Daicos over 28 disposals," those outcomes are heavily correlated. If Collingwood wins by 40, all three are more likely. If Collingwood loses, all three are less likely.
The bookmaker knows this. Their SGM pricing engine bakes in a conservative correlation assumption that protects them from correlated losses. In practice this means SGM legs are priced as if they're slightly more correlated than they actually are, which produces payouts that are 15-25% worse than a naive multiplication of the individual leg prices would suggest.
The total vig on a typical four-leg SGM at an AU bookmaker is 25-35%. Before any consideration of whether you picked the right legs. Before anything else. That margin alone makes SGMs structurally a terrible use of money, even when promo boosts attempt to mitigate it. A 20% promo boost on a 30% vig product still leaves you paying 10% effective vig, which is roughly triple what you'd pay on straight single bets.
The emotional trap
The maths is only half of why people keep betting multis. The other half is emotional, and it's worth being honest about.
A winning four-leg multi at combined odds of $15 returns 15x your stake. On a $20 bet that's $300. On a $100 bet it's $1,500. These numbers feel meaningful in a way that a $3 profit on a single bet does not, even when the expected-value maths says the single bet is the better wager.
The bookmakers understand this exactly. Their UI design pushes you toward multi building at every step. The "Build a Bet" and "Quick Multi" features on Sportsbet, Ladbrokes and TAB are prominent because multis are higher-margin. Every time you add a leg to the multi builder, the potential payout refreshes in a way that makes the next leg look free and valuable. It isn't. It's a compounding vig charge with a carrot on the front.
The six-month numbers
My own data from the period I actually tracked multis (April to September 2024):
Multis placed: 67
Legs: 2-7, average 4.2
Total stake: $3,340
Total return: $1,160 (including partial cash-outs)
Net loss: $2,180
Expected loss if legs had been priced fairly: $0 (since I had a real +EV edge on the legs)
Expected loss from compounded vig alone: approximately $1,400
Additional loss from variance: approximately $780
The variance portion would have averaged out over a larger sample. The compounded vig portion would have been exactly $1,400 no matter how the variance went. I was paying $1,400 a year in structural vig tax to the bookmaker for the privilege of variance exposure I didn't want.
What I do now instead
Single-leg bets, every time. If I want exposure to a theme across a weekend - say, I think underdogs are mispriced in AFL this round - I place three separate single bets, each sized at my regular unit. The aggregate EV is identical to having bet them as a treble. The variance is much lower. The bankroll survives drawdowns. The sizing can be deployed at real Kelly fractions.
Occasionally I still feel the pull toward a multi - a weekend where three games all look like strong picks, a temptation to package them and chase the 8x. I don't do it. The opportunity cost of the multi structure is enough that I'd rather bet three units singly than one unit as a multi, every time.
If you bet multis for entertainment - small stakes, knowing you're buying variance exposure rather than value - that's fine. Call it entertainment spending. For anyone betting for expected value, multis are a direct tax on edge. The bankroll guide walks through why Kelly sizing breaks down on high-variance products, and multis are essentially the highest-variance product AU bookmakers sell.
For the specific case of same-game multis, the Sportsbet promotions piece covers why SGM boost promos don't rescue the underlying product. The whole category is a trap for edge-seeking bettors.
The hardest part of quitting multis is the feeling that you're leaving free payouts on the table. You're not. You're leaving compounded vig on the table. Stay with the boring singles. The bankroll will thank you.
About the author
Tom Rafferty
Columnist
Tom has been punting in Australia long enough to have strong opinions about most of it. Writes the opinion column — multis, Betfair, why your mate is wrong about betting, and the cultural side of being a sharp AU punter.
Story9 min
The Week I Ran Eighth Kelly and It Still Felt Terrifying
Everyone talks about Kelly sizing until they hit a twelve-bet losing streak. Here's what it actually feels like, and why I didn't deviate from the plan.
