The 'risk-free' or 'bet credit' promotion has been the marquee customer-acquisition tool of US sportsbook expansion. The math behind it is straightforward; the math behind why it is now being reined in is more interesting. A walkthrough of token expected value, the math literacy and match-betting populations that broke the original model, and what is replacing risk-free in 2026.
Every US sportsbook acquisition campaign since 2018 has been built around a version of the same offer: bet $100, get $200 if you lose, or its cousin, deposit $50, get a $1,000 bet credit. The numbers are huge by design. The reality, for the typical recreational customer, is that the offer is worth a fraction of its face value, and the books have been priced on the assumption that customers will not extract the full economic value. This article walks through the math, the populations that broke the original model, and what is replacing the 'risk-free' construct in 2026.
Surface vs substance: how the promo actually pays
A 'risk-free $100' promo, in the most common US construction, works as follows: the customer places a $100 wager. If the wager wins, they get the winnings as cash, as normal. If the wager loses, they receive a $100 'bet credit' (sometimes called a 'bonus bet' or 'free bet') that can be wagered but not withdrawn. When the bet credit is then wagered, the customer can withdraw any winnings — but the original $100 of credit is consumed by the bet whether it wins or loses.
The key mechanic is that the bet credit pays winnings on its face value, not on the stake-plus-winnings model of a normal wager. If a customer bets a $100 credit at +200 and wins, they receive $200 (the winnings), not $300 (winnings plus stake return). The stake portion of the credit is consumed in placing the bet, regardless of outcome. This is the structural fact that drives the credit's expected value to a fraction of its face.
Token expected value, three customer profiles
Profile A — the math-literate customer
This customer treats the credit as a derivative asset and works out its EV explicitly. The optimal play depends on the conversion mechanics, but a common strategy is to wager the credit on a high-odds bet (e.g., +400 to +800) at a price that has been verified against fair-value calculations. The fair-value EV of a $100 credit wagered at +400 on a true 20% probability is: 0.20 × $400 = $80 expected winnings. Net of the consumed credit stake, the credit is worth approximately $80 in expectation, against a $100 face value. Across a population of math-literate users, the empirical conversion rate has run 60-80% of face.
Profile B — the median recreational customer
This customer uses the credit on a popular short-odds market (e.g., a favorite at −150 or even money), often without explicit EV calculation. A $100 credit wagered at −150 on a coin-flip wager has expected winnings of: 0.50 × $67 = $33 expected return. The credit is worth approximately $33 in expectation, against a $100 face value. Empirical data on median recreational behavior puts conversion at 25-40% of face. The difference between Profile A's 70% and Profile B's 35% is essentially pricing literacy, applied once.
Profile C — the matched-bettor
This customer bets both sides of a market, one side at the sportsbook with the credit and the other side at an exchange or competing book. The mechanic locks in a near-guaranteed extraction of a fraction of the credit's face value. Empirical conversion in well-executed match-betting on US-licensed promos has run 70-90%. The match-betting community emerged from the UK in the 2010s and arrived in US markets in volume around 2019-2020, expanding rapidly through YouTube and Reddit.
| Risk-free $500 first bet | Math: ~$300 | Median: ~$150 | Match: ~$400 |
|---|---|
| Deposit $50, get $1,000 in bet credits | Math: ~$550 | Median: ~$200 | Match: ~$700 |
| Bet $5, get $200 bonus on first deposit | Math: ~$120 | Median: ~$60 | Match: ~$150 |
| 100% deposit match up to $1,000 (with 5x playthrough) | Math: ~$450 | Median: ~$150 | Match: ~$600 |
| 10 second-chance bets, $50 each | Math: ~$200 | Median: ~$100 | Match: ~$320 |
Why books are deboost-ing in 2025-2026
The 2018-2023 promo budgets at major US books were sized on actuarial assumptions about conversion. The assumptions were calibrated against US market data from the early years of legalization (2018-2020), when the match-betting community was small, the YouTube tutorial ecosystem was nascent, and most customers extracted closer to Profile B's 35% than Profile A's 70%. Acquisition-cost models built on those numbers projected breakeven-positive returns within 12-18 months of customer onboarding.
The actual conversion population shifted faster than the models predicted. By 2022-2023, the math-literate and match-betting segments were collectively extracting closer to 50% of face on aggregate promo spend — well above the 30-35% the actuarial models had assumed. The implication: every dollar of promo spend was returning less customer LTV than projected, and the negative gap was widening as the educated population grew. The deboost cycle started in 2023 and accelerated through 2025.
Three mechanics being used to deboost
1. Tighter conversion mechanics
The promos still appear at the same headline face value ('Bet $5, get $200'), but the fine print has shifted. Common changes include: longer playthrough requirements (was 1x, now 5x), maximum stake per bet credit ticket (was unlimited, now $25 max so the credit must be wagered across 8 tickets), and minimum odds requirements (was +100, now −200 to suppress the high-EV long-shot strategy). Each change reduces the math-literate and match-betting customer's conversion rate without changing the headline.
2. Shift toward odds boosts
Odds boosts are increasingly the primary marketing surface. A boost takes a market the book offers at, say, +250 and offers it at +400 for a maximum stake of $25. The customer perceives generous value; the book retains pricing control. The catch is that boosts are usually offered on selections the book wants action on for risk-management reasons (one-sided heavily-public lines that the book needs balance against). On boosts where the book is offering a true price improvement, max stakes are typically limited and only available to customers without a sharp profile.
3. Profile-targeted promo eligibility
Mass-market customers see the public promo headlines and get full eligibility. Customers whose accounts profile as sharp (positive closing-line value, match-betting patterns, geographic clustering with known sharp accounts) see promo offers degraded or removed entirely. Books are not subtle about this — a customer with a documented sharp pattern is often offered the same promos with maximum stakes capped at $5 or excluded entirely. The mechanic is not new but has become more sophisticated.
What is replacing risk-free
The marquee offer is shifting from 'risk-free' to two replacements: 'bet credit' (a credit that pays only winnings on the credit's face, exactly the mechanic the original 'risk-free' obscured) and 'odds boost' (a one-time price improvement on a single bet). Both are honestly priced — the bet credit is what the original was always doing, and the odds boost is structurally a temporary price cut. Both reduce the gap between perceived and actual value, which is what regulators and the math-literate customer base were demanding.
The shift is also legal. State regulators in New York, Massachusetts, and Ohio have pressed sportsbooks to remove the 'risk-free' language because the offer is not, in fact, risk-free — the customer can lose their original stake even on the offer itself in some constructions. The replacement language ('bet credit,' 'bonus bet,' 'second chance') reflects regulatory pressure as much as customer-acquisition economics.
What this means for the recreational bettor in 2026
Promo value per customer is materially lower than it was in 2020-2022. A recreational bettor opening a new US sportsbook account today is likely receiving a promo with 20-40% of the headline face value in actual expected value, down from 40-60% three years ago. The shift toward odds boosts means the value is more transparent — the customer can see the implied price improvement and decide whether the underlying market is one they would have bet anyway.
The simplest recommendation: do not let promo offers drive your bet selection. The expected value of a bet you would not have made without the promo, even with a 50% boost, is usually negative on the underlying odds. The expected value of a bet you would have made anyway, with a true price improvement from a boost, is positive. The distinction matters more than the headline number.
Is match-betting still profitable in 2026 on US-licensed books?
Less profitable than in 2020-2022, but still positive expected value for customers who put in the work and have access to a US-licensed sportsbook plus a legal exchange or competitor book. Per-promo profits have shrunk roughly 30-40% as books have deboosted, and the customer-targeting algorithms have gotten better at identifying match-bettors and limiting their accounts. The community has shifted toward more selective promo participation, focused on the few books that still offer high-EV promos.
Why do books still offer promos if the math is unfavorable?
Because customer acquisition cost in a competitive market is structurally above zero. Even a promo with negative direct return can be net positive on customer LTV if a meaningful fraction of customers continue to bet after promo conversion. The promo's job is not to break even on its face; it is to recruit a customer who will bet for years afterward. The deboost is calibration, not abandonment.
What is the math-literate customer's optimal strategy for a 'bet $5 get $200' promo today?
Generally: place the qualifying $5 wager on a near-coin-flip event so the qualifier itself is approximately neutral EV. Then wager the $200 credit at +400 to +600 on a market with verifiable fair value. Expected extraction is roughly 60-75% of face, or $120-$150 on a $200 credit, depending on the specific conversion mechanics and odds availability. The math is straightforward; the discipline to do it consistently and avoid the temptation of high-action low-EV markets is the harder part.
Is there any reason to take a promo with a 7-day expiry over a same-face promo with 30-day expiry?
Generally no — the shorter expiry forces conversion under time pressure, which usually means worse market selection and lower EV per credit. A 7-day expiry is roughly equivalent to a 15-25% deboost on the same headline face value, depending on the bettor's market familiarity. Books often impose short expiries on their highest-face-value offers for exactly this reason.
The promo era of US sportsbook expansion is winding down — not because books have stopped offering promos, but because the gap between perceived and actual value has compressed materially as customers got more literate and regulators got more skeptical of the original language. The recreational bettor who pays attention to mechanics over headlines, and who treats promos as occasional value rather than as a strategy, will continue to extract real economic value. The bettor who chases every offer at face will see returns that look like the median Profile B math — roughly a third of headline, often less after the discipline costs of forced conversion.