Draw No Bet: Paying for the Draw to Be Refunded

Draw no bet is the simplest of the derived markets, and the easiest to price yourself. It is 1X2 with one of the three outcomes deleted, and the whole market is just arithmetic on the other two.

How it settles#

Backing the home side, there are three cases and only two of them involve money changing hands:

Settlement is 90 minutes plus stoppage time, as with almost every football market. A cup tie that finishes 1-1 and is won on penalties is a push on draw no bet — your stake comes back even though your team went through. The refund case is often called a push or a void leg.

One consequence that matters more than it looks: a pushed bet is not a won bet. If you keep records by strike rate, a run of draws will make your numbers look strange in both directions. Track pushes separately.

It is the same bet as Asian handicap 0.0#

Asian handicap quotes a line of 0.0 — the "level ball", sometimes literally labelled "draw no bet" — and it settles identically. Winner paid, draw refunded, loser loses. There is no mechanical difference at all.

There is often a price difference. Draw no bet is sold on the main coupon to a recreational audience; Asian handicap 0.0 sits in the Asian handicap section aimed at people who compare prices. The same bet in the sharper section frequently carries less margin. Before you take a DNB price, check what the level ball is paying — it is free money if it is better, and it costs nothing to look.

Deriving the fair price#

Start from fair 1X2 probabilities — margin already stripped (how and why), because margin distorts derived markets more than it distorts the main line. Say the fair line is 49% home, 28% draw, 23% away.

Delete the draw. You are now betting on which of the two remaining outcomes happens, so you renormalise over their combined 72%:

Fair DNB probability (home) = P(home) ÷ ( P(home) + P(away) ) = 49 ÷ (49 + 23) = 49 ÷ 72 = 68.1% Fair DNB odds = 1 ÷ 0.681 = 1.47

And the away side, for completeness: 23 ÷ 72 = 31.9%, fair odds 1 ÷ 0.319 = 3.13.

Check it with a settlement walk-through, because a formula you have not tested is a guess. Stake 100 units at 1.47:

• Home win, 49% of the time: you get 147 back → 0.49 × 147 = 72.03
• Draw, 28% of the time: you get 100 back → 0.28 × 100 = 28.00
• Away win, 23% of the time: you get 0 back → 0

Total expected return: 72.03 + 28.00 = 100.03 units on a 100-unit stake. Break-even to rounding, which is what "fair odds" means. The formula holds.

Note that P(draw) never appears in the formula. That is not a trick — the draw is genuinely irrelevant to the ratio of home to away, which is all DNB is. It becomes relevant the moment you compare DNB to something that pays out on it.

Paste your own prices into the fair odds calculator to strip the margin off a real 1X2 line first — the numbers on the coupon include the bookmaker's cut, and running this derivation on raw prices will give you a fair DNB number that is quietly too short:

Draw no bet versus double chance 1X#

Both protect you against the draw. They do it differently, and the difference is exactly the price.

Both fair prices come from the same three probabilities: 1X is 49 + 28 = 77%, giving 1 ÷ 0.77 = 1.30; DNB is 68.1%, giving 1.47. Neither has an edge over the other — check 1X the same way and you get 0.77 × 130 = 100.1 units on a 100-unit stake, break-even again. Double chance covers this from the other side.

So how do you choose?

Compare the actual offered prices, not the shapes. Suppose a book offers 1X at 1.25 and DNB at 1.40 on the line above. Expected return per unit staked:

• 1X: 0.77 × 1.25 = 0.9625, so you lose 3.75% per unit
• DNB: (0.49 × 1.40) + (0.28 × 1.00) = 0.686 + 0.28 = 0.966, so you lose 3.40% per unit

Here DNB is the marginally cheaper way to express the same opinion, purely because that book has loaded more margin onto the recreational double chance price. On a different fixture it will be the other way round. There is no general rule — do the two lines of arithmetic.

If you disagree with the market about the draw, take the shape that rewards it. Say the market has 49/28/23 and you think it is 47/33/20 — draw more likely, away win less. Then 1X moves from 77% to 80%, a relative gain of 3.9%. DNB moves from 68.1% to 47 ÷ 67 = 70.1%, a relative gain of 3.1%. Both improve, but the market that pays out on draws improves more. If your disagreement is about how bad the away side is rather than about the draw, DNB is the purer expression.

And if you have no view on the draw at all, DNB is the more honest instrument, because it does not force you to have one.

When the shape earns its place#

Draw no bet is at its most useful when you are confident about the ranking of two teams and genuinely uncertain about whether the match resolves at all — a heavy favourite away from home in a competition where the home side is happy with a point, a match with a defensive coach on one side, a fixture in a league whose draw rate runs high.

It is at its least useful on a short favourite, where the draw is already unlikely and the refund is worth almost nothing but still costs you price. If the fair home win is 78%, the draw 15% and the away 7%, fair DNB is 78 ÷ 85 = 91.8%, or odds of 1.09. You have surrendered price to insure against something that happens once in seven matches. At that end of the board, look at Asian handicap lines like −1.0 or −1.5 instead: you get a real price back for taking real risk.