Line shopping at speed — the 3-screen routine

Line shopping is the only edge in betting that requires no model, no angle, and no opinion. It is pure execution, it is available to everyone, and most people do it badly because they do it slowly. This walkthrough builds a three-screen routine that takes a flagged opportunity to a confirmed, logged position in well under a minute — and shows exactly what that speed is worth in EV.

What you'll need

Funded accounts at a minimum of four sportsbooks, ideally six to eight, spread across pricing philosophies (see [#3 — Setting up your sportsbook stack — which books for which plays])
Real-Time Odds, filtered to only the books you actually hold accounts at
A bet log open and ready, so the price you took and the price at close both get recorded without a hunt later
A second monitor, or a phone propped next to a laptop — the routine assumes two visible surfaces, not tab-switching
Roughly 20 minutes to set the layout up once

Why the price gap is bigger than it looks

Start with the arithmetic, because the number surprises people.

The break-even win rate at a given price is the implied probability of that price. At -115, that is 115 / (115 + 100) = 53.49%. At -103, it is 103 / (103 + 100) = 50.74%.

So the same bet at -103 instead of -115 drops your required win rate by 2.75 percentage points. That is not a rounding difference. It is roughly half the size of the entire vig on a standard -110/-110 market.

In EV terms it is starker. On a genuine coin flip, -115 returns 0.5 × 0.8696 − 0.5 = −6.52% per unit. The same coin flip at -103 returns 0.5 × 0.9709 − 0.5 = −1.46%. The gap is 5.07 percentage points of every unit staked. Over 500 one-unit bets, price alone is worth 25.3 units. That is a full season of edge, produced by nothing but where you clicked.

Even the modest version pays. Moving from -110 to -105 is worth 2.16 points per bet, or about 10.8 units over 500 bets.

Step 1 — Screen one: the comparison view

Screen one is Real-Time Odds, and it stays open. Set it up once:

Filter to your books only. A best price at a book you cannot bet at is noise, and noise costs you seconds you do not have.
Sort by the market you actually work — spreads, totals, or a prop category — not by game time.
Turn on the best-price highlight and learn to read it peripherally. You want to see the outlier without reading the row.

Real-Time Odds refreshes at roughly 50 milliseconds across 100+ books. Your eyes are the slow component, which is the whole reason the layout matters.

Step 2 — Check the number before the price

This is the step people skip, and skipping it inverts the entire routine.

In NFL, key numbers dominate price. Roughly 14-15% of games are decided by exactly 3 points. Consider a game truly worth 3: say the favorite wins by exactly 3 nine percent of the time, with the remaining 91% split evenly.

-3 at -115: 0.455 × 0.8696 − 0.455 = −5.93%
-3.5 at -105: 0.455 × 0.9524 − 0.545 = −11.17%

The worse price on the better number wins by 5.23 percentage points. The half point across 3 is worth about 21 cents of price; a book offering you 10 cents to cross it is not doing you a favor.

So the order of operations on screen one is: find the best available number first, then find the best price at that number. Never the other way around.

Step 3 — Screen two: the book, already open

Screen two is the betting surface, and the key word is already.

Log in before the slate, not when you find something. Sessions expire; that expiry is where members lose the price.
Have the sport and market page loaded at the book you use most, so a click lands you on the game rather than the lobby.
Know your own limits at each book by heart. Discovering a $300 maximum at the moment of entry costs you the number.

If you use One-Click Betting at an OAuth-connected book, this step compresses: Sharp pre-fills the betslip with the market, side, and stake. You still confirm it yourself. That confirmation is deliberate — it is not auto-betting, it does not bypass a book's account limits, and the caps you set (daily bet count, weekly bankroll, per-bet maximum, cool-down timer) all still apply.

Step 4 — Screen three: confirmation and logging

Screen three is where the bet becomes data.

Log the bet the moment it confirms: the book, the market, the number, and the price you got, with a column left for the price at close. Do it in the same session — write it down before the next opportunity, or it will not get written down at all.

What you are building here is a CLV record. Closing line value is the difference between the price you took and the market's final price, and it is the metric that proves an execution habit is working. Line shopping produces CLV mechanically: if you consistently take the best of eight numbers, you are consistently ahead of the number those eight books converge on.

Step 5 — Run it against the clock

Time yourself for a week. The target is under 45 seconds from flagged opportunity to logged position. Most members start at three or four minutes, and the entire gap is layout and login friction rather than decision-making.

Two habits that produce most of the improvement:

Pre-decide the stake. Half-Kelly on the measured edge, capped at 5% of bankroll. If you are computing stake size at the moment of entry, you are computing it while the price moves.
Pre-decide the walk-away. Write down the worst number you will accept before you open screen two. If the price is gone, it is gone. Chasing a moved line is how a disciplined shopper turns into a bad bettor.

If it doesn't work

The best price is always at a book you cannot use. Your filter is wrong or your stack is too narrow. Four books is a working minimum, and they should not all be the same kind of book — a stack of three retail apps prices almost identically.

The number moves between screen one and screen two. That is normal at 50ms refresh and it means your screen-two friction is the bottleneck. Log in earlier and pre-load the market page. If it keeps happening on the same market, you are arriving after the move rather than before it.

Your CLV is flat despite good shopping. Shopping cannot rescue a bad read. If you are consistently getting the best number available and still not beating the close, the problem is upstream in selection, not execution. See [#6 — CLV — the only metric that actually matters] for how to read the scorecard.

Your bets stop getting accepted at full size. Fast, precise, best-price-only betting is exactly the profile books limit. Rotate books, mix in ordinary-looking positions, and treat account longevity as part of the routine.

You keep taking the worse number to get the better price. Re-read Step 2 and put the key-number table where you can see it. This is the most expensive mistake in the entire routine.