Hold Percentage in iGaming: Definition, Formula and Why It Sets the Sportsbook Margin
Hold percentage is the share of total stakes that a sportsbook keeps after paying out winning bets, before bonuses and other costs. It is the foundational margin metric for any betting operation, the closest sportsbook equivalent to gross margin in retail. Strong trading teams obsess…
TL;DR
Hold percentage is the share of total stakes that a sportsbook keeps after paying out winning bets, before bonuses and other costs. It is the foundational margin metric for any betting operation, the closest sportsbook equivalent to gross margin in retail. Strong trading teams obsess over hold because everything downstream, from GGR to NGR to LTV, rises and falls with it.
Formula
In its simplest form:
The numerator is identical to GGR for a settled book. The denominator is the total amount staked. Hold is therefore the GGR-to-stake ratio, expressed as a percentage.
A typical warehouse-level aggregation looks like this:
Warehouse-level aggregationSELECT
SUM(stake) AS total_stake,
SUM(stake – payout) AS gross_win,
SUM(stake – payout)
* 1.0 / NULLIF(SUM(stake), 0) AS hold_pct
FROM bets
WHERE settled_date BETWEEN :start AND :end
AND bet_status = ‘settled’;Hold can be calculated at any granularity that matters: per sport, per league, per market type, per channel, per trader, per cash-out vs non-cash-out, per pre-match vs in-play. The same formula applies. The only design choices are which bets to include and how to handle voids, partial cash-outs and free bets.
Why it matters in iGaming
Sportsbook profitability is a pricing problem. Every market you offer carries an implied margin (overround), and your realised hold tells you whether that pricing actually held up against player behaviour. A book that prices well but holds poorly is leaking to sharp customers. A book that holds high but acquires badly is winning small pots. Hold is the metric that connects pricing, customer mix and trading performance into one number.
Different teams read hold differently:
- Trading uses hold to evaluate pricing accuracy, market positioning and sharpness of the customer base.
- Risk uses hold drift as an early warning that liabilities or specific customer segments are running against the book.
- Finance uses hold as the leading indicator that explains week-on-week swings in GGR before they hit NGR.
- Executives track hold as a measure of structural margin health independent of volume.
Hold is also one of the easiest metrics to misread. A single high-stake winning bet can crash weekly hold for a whole league. A long-tail of small mug-money bets can prop hold up well above the structural margin. Reading hold without distribution context is a recipe for wrong calls.
Common mistakes and how teams get hold wrong
Confusing hold with margin or overround. Margin or overround is the implied book margin baked into prices. Hold is the realised margin after settlement. They are related but not equal, and a healthy trading operation tracks both side by side.
Reporting hold without volume context. A 12 percent hold on a low-volume week tells you almost nothing. A 5 percent hold on a high-volume week with sharp action might be a stronger result. Volume-weighted hold across longer windows is more informative.
Mixing pre-bonus and post-bonus hold. Free bets and stake-back promotions distort hold whether they are netted in or out. The fix is a clear convention (typically: report structural hold pre-bonus, plus a separate effective hold including promo cost) and apply it everywhere.
Aggregating across sports without context. Soccer, basketball, tennis and horseracing have very different structural holds. Mixing them into one weekly number can mask serious problems in a single sport that happen to be offset by another.
Forgetting cash-out and partial settlement. Cash-outs settle at a price that is not the original stake outcome. Whether they raise or depress reported hold depends on how the warehouse models them. Inconsistent treatment leads to numbers that drift over time without anyone noticing.
Treating short-term hold variance as a trend. Hold is high-variance over short windows. Reading week-on-week swings as performance signals leads to overreaction. Trading teams typically look at rolling four-week and twelve-week windows to filter out noise.
Healthy ranges and benchmarks
Structural hold varies materially by sport, market type and customer mix. Directional ranges that are commonly seen in the industry:
- Mainstream pre-match singles on top European football tend to run lower than exotic markets and same-game multis.
- In-play hold is typically higher than pre-match hold across most sports, driven by faster decisions and softer customer behaviour.
- Bet builders and same-game multis structurally hold higher than singles, sometimes meaningfully so.
- Horseracing tends to hold lower than ball-sports for sharp markets but higher on tote and exotic bets.
- Customer-mix shifts the realised hold significantly. Books with a high share of recreational customers hold higher than books that attract a sharp crowd, even on the same prices.
These are directional patterns based on industry observation, not strict benchmarks. The right comparison is your own hold over time, sliced by the dimensions that explain variance in your specific book.
How Gamblitude handles hold
In Gamblitude, hold is a governed Metric defined once at the semantic layer, with explicit rules for which bets are included, how cash-outs are handled and how promotional bets are treated. Trading, Risk and Finance share the same number across dashboards, reports and AI Agent answers. Variants such as Pre-Match Hold, In-Play Hold, Bet Builder Hold and Effective Hold (post-bonus) coexist as separate Metrics, each with its own ownership and audit trail.
Insight Radar continuously monitors hold across sports, leagues and customer segments, surfacing anomalies before they become structural problems. Sharp drops in hold against a specific market or customer cohort trigger alerts directly to trading teams, often within hours rather than the days it takes to spot the same pattern in a manual report.
FAQ
Closely related but not identical. The bookmaker’s margin (overround, vigorish) is the implied margin baked into the prices on a market. Hold percentage is the realised margin after the bets settle. The two converge over very large volumes of recreational bets, but diverge whenever sharp money or unusual results move outcomes.
In-play betting typically holds higher than pre-match for two reasons. Players make faster decisions with less information, and the prices update continuously based on real-time match state. Both effects favour the book. Operators that do not slice hold by pre-match vs in-play miss one of the most material drivers of margin performance.
Sport outcomes are noisy. A few high-stake winning bets, a major upset in a popular fixture or a quiet promotional weekend can all swing weekly hold by several percentage points. This is why trading teams look at rolling four-week and twelve-week hold rather than reacting to single-week swings.
Both, with clear naming. Structural hold (pre-bonus) tells you how good your pricing and risk management are. Effective hold (post-bonus) tells you how much you actually keep. Reporting only one or mixing them creates the same kind of internal disputes that plague NGR reporting.
There is no single right answer. Mainstream football singles tend to sit at the lower end, exotic markets and bet builders at the higher end, and customer mix moves the realised number significantly. Operators that compare their hold to a single industry benchmark without controlling for product mix and customer profile usually reach wrong conclusions.
Further reading
Found a mistake or want a term added to the iGaming Glossary? Let us know.
Browse the complete glossary or see how governed definitions work across dashboards, reports, alerts and AI answers.