Your sportsbook stack — which books for which plays

One account means you take whatever price that book posts, you can never arb, and you convert one welcome offer instead of six. This lesson covers why the sportsbook stack is infrastructure rather than preference, what each tier of book is for, and how to build a working 5-8 account stack without tying up your bankroll or getting flagged in month two.

One book is a ceiling on your edge

Start with the arithmetic. Say your read on a side wins 53% of the time. Book A has it at -110; Book B has the same side at -105.

  • At -110 you need 110 / (110 + 100) = 52.38% to break even. Your EV is 0.53 × 1.909 − 1 = +1.18%.
  • At -105 you need 105 / 205 = 51.22% to break even. Your EV is 0.53 × 1.952 − 1 = +3.48%.

Same read, same risk, nearly three times the expected value — purely because you had a second account open. That gap is line shopping, the highest-return five seconds in betting, and over a season it dwarfs almost everything else a beginner worries about.

Two more things a single account makes impossible:

Arbitrage needs two sides at two books, simultaneously. Say Book A posts Team X at +125 (implied 100/225 = 44.44%) and Book B posts Team Y at -115 (implied 115/215 = 53.49%). Those sum to 97.93% — under 100, meaning the two books together pay more than the outcome is worth. Split $1,000 proportionally: $453.82 on X, $546.18 on Y. X wins, collect 453.82 × 2.25 = $1,021.10. Y wins, collect 546.18 × 1.8696 = $1,021.10. Either way, $21.10 of covered profit on $1,000 — a 2.11% return — available only because you held both accounts.

Promo conversion scales with account count. Every book runs its own welcome offer and reload cycle. Six accounts means six offer streams to convert (Lesson #14). One account means one, then nothing.

The three tiers of a stack

Sort books into three tiers and the "which book for which play" question answers itself.

Tier 1 — sharp / reference books. Low hold, high limits, sharp-tolerant. Pinnacle is the market's default reference price; its two-sided market devigs to the cleanest estimate of true probability available for free (Lesson #4). Circa operates in a small number of US states and posts early numbers with real limits. Pinnacle is not licensed in US regulated markets, so for most American bettors it is a price you read, not a book you hold — the reference value is in the number, not the account.

Tier 2 — large public books. DraftKings, FanDuel, BetMGM, Caesars, ESPN BET, Fanatics. Broadest state licensing, deepest prop menus, heaviest promotional volume, and the OAuth connections One-Click Betting uses to pre-fill a betslip. Fast on marquee markets, slow on second-tier props, and quicker than anyone to limit a winning account.

Tier 3 — regional and soft books. BetRivers, SuperBook, Hard Rock Bet, Bally Bet and the state-specific operators. Fewer eyes on the lines, slower repricing after news, smaller limits. Your best arb partners and richest source of stale numbers (Lesson #16).

Prediction markets sit alongside all three as a fourth venue with a different regulatory footprint (Lesson #18).

Which book for which play

Tier Reference pricing Prop depth Promo volume Arb partner Live speed
Sharp / low-hold Primary — this is the fair price Thin, but very clean None Rarely (prices are already right) Fast, tight
Large public Secondary consensus Deepest menus Highest Occasionally, on the popular side Fast on mainlines, slow on props
Regional / soft Not usable Moderate Moderate Best — this is where the gap lives Slow, which is the point
Prediction market Clean, no vig baked in Event contracts, not player props None Cross-venue hedges Varies by contract

Read the table as a routing rule. Reference price comes from Tier 1. Prop volume comes from Tier 2. The second leg of an arb almost always comes from Tier 3, which is also where live edges cluster, because the slowest book to reprice is the one paying you.

A starter stack of six to eight accounts

Build in this order. Each account has one job.

  1. A reference read — Pinnacle's no-vig number, or Circa where it operates. Sets the fair price everything else is measured against.
  2. Three large public books — prop depth, promo volume, One-Click support, and three independent quotes on every mainline. Three is where line shopping starts paying real returns.
  3. Two regional or soft books — arb partners and stale-line sources. These do the heaviest lifting per account in the stack.
  4. One book you rate as fast in-game — live markets. Test this yourself; live speed varies by sport and by book, and it changes.
  5. A prediction-market account — event contracts and cross-venue hedging (Lessons #18 and #20).

Six to eight accounts is the practical sweet spot. Below five you leave line-shopping value on the table. Above ten, funding overhead and KYC paperwork cost more time than the marginal book returns.

Funding and withdrawals

The unglamorous part that decides whether the stack works.

  • Name match is mandatory. The funding account has to be in your legal name. A joint account or a partner's card will fail verification and can freeze a balance mid-season.
  • ACH and online banking are the cheapest rails. PayPal is usually fastest where supported. Debit cards frequently decline on the merchant category code; credit cards are blocked at most US books.
  • Withdrawals are closed-loop. Money generally returns by the method it arrived, and your first withdrawal triggers a document upload. Do that early, on a small amount, so it is finished before you need cash out.
  • Budget for float. An arb needs funded balances at both books at the same moment, and money in transit cannot take a leg. Expect a meaningful share of your bankroll to sit distributed across books.
  • Your bankroll is the sum, not the pile at one book. The 5% cap on any single bet applies to the total across every account. A $5,000 bankroll spread over six books still caps a single bet at $250, not $250 per book.

Where Sharp links to a sportsbook offer, an affiliate disclosure applies. We surface the price; we do not take the other side of it.

State availability is the real constraint

Legal betting in the US is licensed state by state. The roster of books, the market menus, and even the prop types differ across state lines, and geolocation must confirm you are physically in-state at the moment you place. Your best arb partner in one state may not exist in the next one over.

So build the stack you can actually have where you live, not the one you read about on a forum — then filter your tools to match. Real-Time Odds is filterable to the books you actually hold accounts at, which turns a wall of prices into a shortlist. Prediction markets run under a different framework, so their availability map does not match the sportsbook one.

Keeping the accounts healthy

Accounts are a depreciating asset — every winning bet spends a little of one. The stack is your insurance, but only if you manage it (Lesson #13).

  • Round your stakes. $50, not $47.23. Perfectly optimized stakes are the loudest signal a bettor sends.
  • Rotate which book takes the arb leg. Hitting the same soft book every time gets that book gone in weeks.
  • Do not always take the maximum on a stale number. Grabbing the full limit seconds after a line error is the fastest route to a restriction.
  • Leave a working float. Do not sweep the whole balance after every win.
  • Spread the mainline action. A book that only ever sees your sharpest bets works that out.

Limits come anyway; that is the cost of doing this well. A six-book stack absorbs one restriction. A one-book stack does not.