Bet slip edits lock at leg 3 — the stake you typed doesn't

Most Australian sportsbooks let you keep editing a bet slip right up until the moment you hit place — change a leg, swap the market, drop the stake from $50 to $5, all fine. What almost none of them tell you is that the legs stop being yours much earlier than the stake does. Once a multi has three or more legs on it, the standard behaviour across the major apps is that the combination gets priced and locked at leg three. The stake box stays live. The selections behind it don't.

That distinction matters more than it sounds, because it means the number you're typing at the end of a build-a-bet isn't a fresh bet at all. It's a quantity applied to a slip that was finalised two or three taps ago.

Where the lock actually sits

Open any of the big Australian operators and build a same-game multi. Add leg one. Add leg two. The odds update as you go, and the price on the slip is still provisional — pull a leg out and the whole thing reprices cleanly. Add leg three, and something changes in the plumbing. The combined price is now being calculated against a specific set of correlated outcomes, and the app treats that set as committed. From that point, removing a leg doesn't just subtract its odds. It can invalidate the price entirely and force a rebuild.

You won't see a padlock icon. What you'll see, if you're watching, is that the odds on a four-leg same-game multi stop moving in the way you'd expect when you tinker with leg two. The price you're shown is anchored to the combination as it stood when leg three landed.

The practical tell is the "same game multi" toggle itself. Turn it on, add three legs from one match, then try to swap the third. On most platforms the price collapses back to a single or gets wiped and repriced at a worse number. That's not a glitch. That's the lock doing its job.

Why three is the trigger

Three legs is the threshold where correlation modelling gets expensive and where the book's exposure stops being trivial. A two-leg multi across unrelated markets is close to a straight multiplication of independent probabilities — cheap to price, easy to void. A three-leg same-game multi involves overlapping outcomes: a player to score, their team to win, and total goals over 2.5 are not independent events, and the book has to solve for the joint probability rather than multiply marginals.

Once that model is running, the operator wants the inputs frozen. Letting you keep editing the legs after pricing would mean re-solving the correlation for every tap, and it opens a small but real arbitrage window: build a combination the model misprices, watch the live odds shift, then adjust a leg to capture the difference before placement.

The stake field is a different animal

Here's the part that catches people. The stake input has no such lock. You can leave a slip sitting for twenty minutes, come back, change $20 to $200, and the app will accept it without a murmur — right up until the market suspends or the odds drift beyond the tolerance the book sets.

That asymmetry is deliberate. Stake is a scalar. Applying a new scalar to an already-priced combination costs the book nothing computationally and doesn't change the underlying correlation. Editing a leg changes the shape of the bet. So one is gated and the other isn't.

The consequence is that a lot of punters believe they're "still building" a bet when the bet has already been built for them. They see a live stake box and read it as a live slip. It isn't.

What the odds drift tolerance does

Most Australian books apply an acceptance tolerance somewhere between 5% and 15% on multi odds. If the combined price moves more than that between the moment you last touched a leg and the moment you confirm, the bet gets referred or rejected rather than placed at the shown number. On a three-leg multi priced at $6.40, a 10% tolerance means anything from about $5.76 to $7.04 goes through silently. Outside that band, you'll get the "odds have changed" prompt.

That tolerance is measured against the locked combination, not against whatever you've been fiddling with in the stake field. Changing your stake doesn't reset the clock on the lock or the tolerance. Only changing a leg does — and changing a leg after leg three is often what triggers the repricing you didn't want.

The same-game multi is where this bites hardest

Single-game multis are the product where leg-locking is most visible, because the correlation is tightest. A three-leg SGM on an NRL match — say, a team to win, a player to score first, and total points over 42.5 — has genuinely interdependent legs. The book can't price that as three separate bets and hope for the best.

Compare that to a four-leg multi spread across four different codes and four different days. The legs are close to independent, the correlation model barely engages, and you'll often find you can edit legs more freely because repricing is just arithmetic. The lock is still nominally there at leg three, but it's less likely to bite because the price doesn't swing much when you change something.

So the rule of thumb: the more correlated your legs, the earlier and harder the lock. Same-game multis are the extreme case. Cross-code accumulators are the mild one.

A worked example

Say you're building a three-leg SGM on a Friday night AFL game. Leg one: home team to win at $1.65. Leg two: a key forward to kick 2+ goals at $1.90. Leg three: total score over 165.5 at $1.85.

Straight multiplication gives you $5.80. The actual SGM price will be lower — often meaningfully lower — because those outcomes correlate. If the forward kicks multiple goals, the team is more likely to win and the total is more likely to go over. The book prices that overlap and you might see $4.30 instead of $5.80. That gap is the correlation adjustment, and it's calculated the moment leg three lands.

Now try to swap leg two for a different player. On most apps, the $4.30 evaporates. You get repriced from scratch, possibly at a worse number if the market has moved, and possibly with the new leg rejected outright if it's already suspended. The stake box, meanwhile, sat there the whole time looking perfectly editable.

Why this is worth knowing before you build

The lock isn't a trap in the sense of being hidden on purpose — it's disclosed in the terms, usually in a paragraph most people scroll past. But the user interface tells a different story. A live stake field implies a live bet. When the legs are frozen and the stake isn't, the interface is showing you the one part of the slip you can still change and hiding the part you can't.

For anyone who builds multis methodically — adding legs, checking the price, thinking about it, adjusting — the implication is that the thinking window is shorter than it looks. The moment your third correlated leg lands, the number you're looking at is the number you're getting, give or take the tolerance band. Everything after that is just deciding how much to put on it.

Which raises the obvious question: if the stake is the only live variable after leg three, why do so many apps keep the leg-edit buttons active and responsive, letting you tap them and watch the price repaint, rather than graying them out the moment the lock engages? Is that a design oversight, or is a slip that feels editable right up to the end simply better for turnover?

And if you're the sort of punter who builds a multi across an afternoon, checking prices as team news drops, the answer probably changes how you sequence the build. Legs first, stake last — and don't count on the last leg being yours to change once the third one's in.