The common belief is that a book wins because it knows more than you. That is sometimes true. But the structural reason is duller and much more reliable.
Take three prices. Sum their implied probabilities. The total is above 100%, and the excess is the book's cut:
2.10 / 3.40 / 3.60 → 47.6% + 29.4% + 27.8% = 104.8%That 4.8% is collected on every market, regardless of who wins. A book that balances its liabilities does not need to predict the match at all.
Money on one side increases the book's liability on that side. The price on that side shortens, the other side lengthens, and the book returns to balance. Movement is mostly a response to weight of money, not a change of opinion — which is why a line can drift for reasons that have nothing to do with football.
A book holding a balanced position earns the margin whichever way the result falls. A bettor holding one position earns or loses the entire outcome. Those are different games with different variance, and comparing your win rate to a book's "win rate" is a category error.
Your long-run result is roughly your edge minus the margin, multiplied by the number of bets. That is why margin shopping is not a detail: on a 5% margin, a 2% edge becomes a 3% loss, and no amount of study inside the same price fixes it.