Roulette's neighbour bets hedge your losses — at 9% vig

Let’s skip the preamble and get to the number that matters: neighbour bets on a single-zero wheel carry a 9% house edge, roughly double the 2.7% you’re giving up on a straight-up number. That’s the price you pay for buying a slice of the wheel around a chosen number, and it’s why these bets are the quietest way to bleed a bankroll dry while feeling like you’re playing smart. The pitch is that you’re hedging — covering five numbers in a row so a nearby win softens the sting of a miss — but the maths says you’re paying a premium for that comfort, and most punters never stop to check the vig on the call bets section of the layout.

What a neighbour bet actually buys you

On a European wheel, a neighbour bet means you pick a number, say 17, and you cover 17 plus the two numbers on either side of it — 16 and 19 on one side, 18 and 21 on the other (depending on wheel order, not table order). You’re staking five units, one per number, and if the ball lands on any of those five, you get paid at straight-up odds of 35:1 on that single unit. The other four units are gone.

Here’s where the 9% comes from. Five straight-up bets at 35:1 each, on a wheel with 37 pockets, gives you an expected return of 5 × (1/37) × 35 = 4.73 units for every 5 units wagered. That’s a 94.6% return, or a 5.4% house edge — not 9%. So why does every serious roulette reference quote 9% for neighbour bets?

Because you’re not actually placing five straight-up bets. You’re placing a single call bet that pays a bundled rate. On most tables, a neighbour bet on five numbers pays 31:1 on your total stake, not 35:1 per individual number. The casino takes the extra 4 units of theoretical payout and pockets it. A 31:1 payout on a 5-unit stake means you win 31 units plus your 5 back — total return of 36 units on a bet that should mathematically return 47.3 units if it were five independent straight-ups. That gap is the vig, and it works out to a 9.0% house edge after you account for the 1/37 hit probability.

Why the hedge feels good but pays worse

The psychological trick is that neighbour bets smooth out variance. If you’re playing 17 straight and the ball lands on 19, you lose everything. With neighbour bets, you’ve got a 13.5% chance of hitting something on each spin (5 out of 37 pockets), versus 2.7% for a single number. That’s a massive difference in how often you walk away from the table with a win in your pocket — but the size of that win is smaller relative to your total stake.

Let’s run a concrete comparison. Over 100 spins, betting $5 per spin on a straight-up number costs you $500 total. Expected loss at 2.7% edge: $13.50. You’ll hit roughly 2.7 times, and each hit pays $175 profit (35:1 on $5), so you’re looking at about $472.50 returned on average. Now do the same with $25 per spin on a neighbour bet covering five numbers — same $500 total exposure across 100 spins. Expected hits: 13.5. Each hit pays $155 profit (31:1 on $25 stake, minus the $25 you get back plus the $5 units you didn’t win — the actual payout structure varies by casino, but the standard is 31:1 on the total bet). Your expected return is $418.50. You’re losing $54 more over the same number of spins, but you’ll have a winning session far more often — about 13 wins instead of 3.

That trade-off is the entire appeal. You’re buying frequency of reinforcement at a price of roughly 6.3 percentage points of extra house edge. For a casual punter who gets a kick out of multiple small wins across an evening, that might be worth it. For anyone trying to grind a bonus or stretch a bankroll, it’s a slow leak that makes the 2.7% edge look like a gift.

The split between European and American wheels

One number that matters: on a double-zero wheel, neighbour bets get even uglier. The 38th pocket doesn’t just add a house slot — it changes the payout maths. A five-number neighbour bet on an American wheel pays the same 31:1, but your hit probability drops to 5/38 (13.2%), and the house edge jumps to 12.6%. That’s the difference between losing $63 per $500 wagered versus $54 on a single-zero table. If you’re in a venue that only runs American wheels, neighbour bets are close to the worst wager on the floor, second only to the five-number bet on the top line (which carries a 7.89% edge — ironically better than a neighbour bet on the same wheel).

Australian players are lucky on this front. Most land-based venues and regulated online casinos here run single-zero wheels as standard, and some even offer French rules with la partage on even-money bets. But here’s the catch: neighbour bets are exempt from la partage. That rule only applies to red/black, odd/even, high/low. The casino isn’t giving you half your stake back when the ball lands on zero if you’ve got a neighbour bet covering 23 and its neighbours. So even at a French table with 1.35% effective edge on even-money bets, your neighbour bet is still paying the full 9%.

When neighbour bets make sense (and when they don’t)

There’s a narrow case where the 9% vig is a reasonable trade: when you’re playing a short session with a strict win target. Say you’ve got $200 and you want to walk away if you hit $300. A neighbour bet on a single number cluster gives you a 13.5% chance per spin of hitting a $155 profit on a $25 stake. You need roughly two hits to hit your target, and the probability of getting two hits within 20 spins is about 44% — versus a 23% chance of hitting a single straight-up number twice in the same span. The extra frequency genuinely improves your odds of hitting a session target, even though the expected value is worse. That’s the one legitimate argument for these bets: they’re a volatility management tool, not an edge play.

The flip side is bonus wagering. If you’re working through a 40x wagering requirement on a $100 deposit bonus, you need to turn over $4,000. At a 2.7% edge, your expected cost of clearing is $108. At 9% for neighbour bets, it’s $360. That’s the difference between walking away with a small profit and eating your entire bonus plus a chunk of your deposit. No casino is going to tell you that, but the math is right there on the layout.

The real question nobody asks

The next time you’re at a table and the dealer offers “neighbours of 17” or you see the call bets section on an online client, ask yourself what you’re actually buying. The casino has already priced the comfort of a wider net into the payout, and they’ve done it at a rate that makes the games with the worst reputation — like keno or the big six wheel — look almost fair. A 9% vig isn’t a rounding error; it’s a deliberate pricing structure that exploits the gap between how often you win and how much you win.

So the open question is whether that extra frequency is worth the extra 6.3 points of edge. For a casual night out, maybe. For anyone who’s ever calculated an expected value in their head, it’s hard to see the appeal. But if you do play them, at least know the number you’re fighting — and remember that on a double-zero wheel, you’re fighting 12.6%, which isn’t a hedge at all. It’s a donation with extra steps.