I worked the betting shops in London from 1972 to 2002. Took bets on everything. Football, horses, snooker. Totals were the most interesting because they are the easiest to get wrong.
A total is a bet on whether the combined score of both teams will be over or under a certain number. The sportsbook sets the line at 45 points. You bet over or under.
It sounds simple. It is not.
The Difficulty
An NFL game score depends on multiple variables. How many passing yards each team gets. How many touchdowns. How many field goals. How many safeties. The interaction between offensive and defensive quality on both sides.
The sportsbook has to estimate all of these and produce a number that is right in the middle.
If the sportsbook overestimates, over bettors win. If they underestimate, under bettors win. The sportsbook makes money off the vig, not off being right.
Where Smart Bettors Win
Smart bettors look for totals that are mispriced. They have their own models. They estimate what the real total should be. If the sportsbook line is far from their estimate, they have an edge.
Example: your model says the total should be 45 points. The sportsbook set it at 47. Over bettors are paying for a line that is two points too high. Under is the bet.
This requires research. You need to know the quarterbacks' passing statistics. You need to know the defenses' yards-per-play average. You need weather forecasts. You need injury reports.
The Trend Trap
Amateurs look at a team's recent game totals and assume that pattern continues. The Dallas Cowboys just played 55 and 49 and 52 total-point games. So this week will also be high.
This is nonsense. Opponents matter. If Dallas played bad defenses two weeks in a row and now plays a good defense, the total drops.
Trends are coincidence. The only thing that matters is the matchup.
How to Be Profitable
Build a model. Use yards-per-play data. Use historical offensive and defensive efficiency. Use game-specific variables (weather, injuries, travel).
Estimate the total yourself. Then compare to the sportsbook line. Bet when there is a meaningful gap (at least half a point).
Do not bet every week. Bet only when you have genuine edge. This means betting maybe 5 to 10 percent of available games.
Manage your bankroll. You will have losing weeks. You need capital to survive variance.
The House Edge
The sportsbook builds in a 4.5 percent vig. Your model needs to be better than 4.5 percent for you to profit long-term.
This is a high bar. Most casual bettors are worse than 4.5 percent. Smart bettors are better.
Why I Stopped
I closed my shop when online betting took off. But back when I was a bookie, the totals were where I saw the smartest money move. The point-spread guys were okay. The total guys were dangerous.
They knew the game. They had data. They understood variance. They did not chase losses.
They bet less often than casual bettors but they won more overall. That is the whole game.
