What is a totals middle worth
How do you price an over/under middle without guessing?
A totals middle is over 44.5 at one book and under 47.5 at another. Both bets win if the combined score lands strictly inside — 3 numbers here. One always wins, so the position costs the hold on one leg and pays the window.
The two numbers
Break-even is arithmetic. At -110 both ways the pair needs the window to hit 4.76% of the time to be worth taking. That figure comes from the two prices and nothing else — no distribution, no model, no assumption you can get wrong.
The window probability is not. A normal curve with a 13-point spread puts this window at 9.19%, which clears the bar comfortably. That number is an approximation, and the spread behind it is a stated prior rather than a measurement of any sport.
Why the approximation is the weak part
Real scores are lumpy. They cluster on numbers the sport's scoring produces often, and no smooth curve reproduces a lump. A window sitting on a cluster is worth more than the curve says; one sitting in a gap is worth less. The curve cannot tell you which you have.
Counting recorded finals can. It needs 400 games for totals against 200 for margins, and the reason is structural: a margin is folded — the absolute difference — so it piles onto a few small integers, while combined scores spread across a far wider range. The same game count buys much thinner evidence per number, so a second rule applies: at least 8 games on each distinct total actually observed, measured rather than assumed.
Below either bar the answer is the approximation, labelled as one. A counted probability backed by three games in the window is worse than an admitted estimate, because it does not announce itself.