GLI-31 v1.1 Electronic Raffle Certification: What It Covers and What It Misses

GLI-31 v1.1 certifies that an electronic raffle system sells tickets, accounts for sales and draws winners in a way that can be audited. It does not certify an online raffle business. The standard was released on 24 July 2015 and is written around Raffle Sales Units and bearer tickets, so a lottery operator running an electronic charity raffle should treat a GLI-31 certificate as proof of draw integrity and ticket accounting only.

The bottom line for licensees is this. Use GLI-31 to scope and evidence the draw engine, the ticket ledger and the prize reports. Then close the gaps the standard leaves open, including player accounts, payments, age and identity checks, and safer-gambling controls, through the regulator’s own rules and a broader certification path. Obtain jurisdiction-specific legal advice before relying on the certificate for licensing.

What GLI-31 v1.1 Actually Is

GLI Standard 31 describes itself as a document produced “for the purpose of providing independent test reports and/or certifications to regulators indicating the state of compliance of various suppliers’ devices and systems.” It was created in response to regulators who asked for technical guidance on electronically conducted raffles. As of September 2026, Version 1.1 is the version reviewed for this article.

“This standard is not intended to represent a set of rigid requirements that every raffle system must comply with, as the operation of raffles varies widely due to jurisdictional laws, varying levels of acceptable security measures and differing implementations of raffle technologies.”

That disclaimer comes from the “About This Standard” preface of Version 1.1. GLI-31 is therefore a technical baseline, and the regulatory body that adopts it decides which provisions bind, which are waived and which are supplemented. The phrases “as allowed by the regulatory body” and “as required by the regulatory body” appear repeatedly in the text wherever the choice is operationally significant.

Section 1.3.1 defines an Electronic Raffle System as “computer software and related equipment used by raffle licensees to sell tickets, account for sales, and facilitate the drawing of tickets to determine the winners.” A Draw Number is the uniquely identifiable number given to the purchaser for each chance, and the standard also defines the Bearer Ticket issued to the purchaser. Those definitions fix the scope at sales, accounting and the draw.

Section 1.4.1 sets out two phases of certification. The first is initial laboratory testing of the system together with the Raffle Sales Units, assembled in the laboratory. The second is on-site testing after installation, “as required by the regulatory body,” which can include event simulations with and without challenges to system operations, stability testing at maximum anticipated loads and verification of the venue’s internal controls and IT infrastructure. The second phase is therefore regulator-dependent, and licensees should confirm whether it applies in their jurisdiction before building the project plan.

Draw Integrity: Chapter 4 and the Winning Number

Section 4.1.1 states the rule in one sentence.

“The selection process for the winning number shall be random.”

The standard allows this to be met through an approved random number generator. It leaves the method open, which fits section 1.2’s stated intent to allow a wide range of methods to conform to the standard while encouraging new ones.

The detail sits in the surrounding sections. Section 4.2 requires the independent test laboratory to apply recognised statistical tests, and the standard lists chi-square, equi-distribution, gap, overlaps, poker, coupon collector’s, permutation, Kolmogorov-Smirnov, runs, serial correlation and Poisson distribution tests, among others. The laboratory chooses the appropriate tests case by case, depending on the RNG under review.

Section 4.2.3 requires the RNG period, together with the method of implementing outcomes, to be “sufficiently large to ensure that all valid, sold numbers are available for random selection.” Section 4.2.4 requires the raw range to be large enough for adequate precision when scaling and mapping. Section 4.3.1 requires scaling methods that are linear and introduce no bias, pattern or predictability, and requires scaled outcomes to be proven against recognised statistical tests. The standard also keeps seeding and re-seeding to an absolute minimum unless the manufacturer proves no adverse effect on randomness.

Section 4.4.2 addresses interference. An electronic raffle system “shall use appropriate protocols to protect the random number generator and random selection process from influence by associated equipment, which may be communicating with the electronic raffle system.” For a charity raffle, this is the provision that governs how a front-end, a reporting tool or a broadcast overlay connects to the draw engine.

Ticket Accounting, Voids and Prize Verification

Chapter 2 builds the accounting model around the bearer ticket. Under section 2.3.1, a ticket must carry, at minimum, the name of the organisation conducting the raffle, an event identifier or location, one or more unique draw numbers, the issue date and time in 24-hour format, the RSU identifier, the ticket value, a unique validation number or barcode, and the licence number where applicable. Where one operator runs a series of raffles, tickets for each raffle must be differentiated.

Section 2.3.2 requires the validation-number method to be “unpredictable” and to prevent duplicate validation numbers within the raffle in progress. Section 2.4 deals with counterfoils. Where a manual draw is used, counterfoils must be identical in size, shape and weight. Where an electronic RNG selects the winner, a printed counterfoil is not required. Reprinting of counterfoils requires supervised access controls, and every reprinted draw number must be flagged in the system as a reprint (section 2.4.2).

The closing sequence is where ticket accounting meets draw integrity. Section 2.6.2 requires the system to close the sale of bearer tickets at an operator-determined time, with no sales afterwards, and to show the operator that all sales from each device have been uploaded. Once the transfer is verified, the RSU must be capable of being reset or closed and must not be enabled for further sales. Section 2.6.3 then provides:

“Voided tickets shall not be qualified toward any prize. The system must be capable of reconciling voided sales for the raffle purchase to identify all voided tickets which may be committed to the draw.”

The system must also record an acknowledgement from the event manager that voided tickets have been reconciled before it permits a winning number to be entered. A draw cannot proceed in a certified system until that acknowledgement is on record.

Prize verification follows. The system must be able to verify winning draw numbers and allow validation either manually or by barcode scanner or equivalent. A note in the standard states that amounts won above any jurisdictional limit require the appropriate documentation to be completed before the winning participant is paid.

Section 2.8.1 sets the reporting minimum. The Raffle Drawing Report must show the date and time of the event, the organisation running it, sales totals and refunds, prize value awarded, prize distribution (total raffle sales against prize value awarded), refund totals, the count of draw numbers in play, and the winning numbers drawn including draw order, call time and claim status. An Exception Report covers system exceptions. For a charity raffle, the prize distribution line is the figure a charity regulator or lottery board will want reconciled against the declared proceeds.

Supporting controls in chapter 5 protect these records. Section 5.2.4 bars alteration of accounting, reporting or significant event data without supervised access controls, and requires the system to log the element altered, its value before and after, the time and date, and the user login. Section 5.2.5 bars any means for an operator to program the server, with the example that an operator should not be able to run SQL statements to modify the database. Section 5.3 requires an internal clock in 24-hour format that time-stamps significant events, sales and draw events. Section 5.4.1 requires a master list of every authorised RSU, and section 5.7.2 requires recovery from the last viable backup to restore significant events, accounting and reporting information.

Raffle, Bingo and Pull-Tab: Where GLI-31 Sits

GLI-31 governs a pooled draw where each paid chance carries a unique draw number and the winner is selected from the numbers sold. Other GLI standards cover other lottery-style products, and they differ in what they treat as the outcome event. The versions and dates below are as published in the standards themselves, as of September 2026.

Standard	Version and date	Product	What the standard certifies
GLI-31	1.1, 24 July 2015	Electronic raffle	Ticket sales, accounting and the drawing of tickets to determine winners
GLI-15	1.3, 6 September 2011	Electronic bingo and keno	Displaying selected balls or numbers, verifying and awarding winnings, accounting and game history reporting
GLI-14	2.2, 6 September 2011	Finite scratch ticket and pull-tab	Central System selecting an Electronic Ticket from a Game Set, with the outcome displayed on a Player Terminal

The structural difference is in where randomness is applied. In GLI-31 the RNG selects a winning number from the numbers sold, and the system must close the purchase period and reconcile voids before a winning number can be entered. In GLI-14 the Central System selects a predetermined ticket from a finite game set at the point of purchase, and the player is paid according to the paytable. In GLI-15 the system draws balls or numbers in a session.

In practice, a lottery operator whose electronic charity raffle is sold alongside instant or bingo-style games should expect each product to be tested against its own standard, because the certified outcome event differs. For the wider question of how GLI selects between standards for online platforms, see our comparison of GLI-19 and GLI-33.

Does a GLI-31 Certificate Cover an Online Charity Raffle?

No, not on its own. GLI-31 certifies the raffle management, sales unit, RNG, server and communications layers. It leaves online-specific obligations to the regulator, and a licensee selling raffle tickets over the internet will usually need controls that sit outside the standard.

The standard is not silent on technology. Section 3.1.1 allows chances to be sold from attendant-operated or player-operated RSUs, with “any other methods” reviewed “on a case-by-case basis, as allowed by the regulatory body.” Section 5.1.1 allows servers to be remote, over a wide area network, where the regulatory body permits. Chapter 6 requires mutual authentication between components, encryption of all communications critical to the raffle, and authorised devices only (sections 6.1.2 and 6.1.3). It also requires communications to pass through an application-level firewall with audit logs (section 6.2.2) and permits remote access only where authorised (section 6.3.1). Wireless RSUs may communicate only through authorised access points (section 3.2.6).

What the chapters reviewed do not contain is any player-account model, identity or age verification, payment processing, deposit or spend limits, self-exclusion or customer-interaction requirements. These are the controls that a jurisdiction such as Ontario, where AGCO and iGaming Ontario run a conduct-and-manage model with Registrar’s Standards for Internet Gaming, applies to online products (see our AGCO and iGaming Ontario profile). Whether a charity raffle falls under those regimes, or under a separate charitable gaming or lottery statute, depends on the jurisdiction’s definition of the activity, and legal counsel should confirm that classification before the product is scoped.

The counter-argument deserves weight. GLI-31 states that it is not intended to mandate specific “operational practices” and invites regulators to layer their own rules over it. A regulator that runs charitable raffles under a lottery act may reasonably treat the standard as sufficient for the draw and rely on licence conditions for everything else.

The standard’s age also matters. Version 1.1 dates from 2015, and section 1.1.2 commits GLI to update documents “as often as possible to reflect changes in technology, testing methods, or cheating methods.” The chapters reviewed contain no clause on cloud hosting, API integrations or third-party payment services. Where a supplier’s architecture relies on these, the laboratory and regulator will need to agree how the existing clauses apply. Information-security assurance beyond the standard’s own server and communications requirements is commonly handled through a separate framework, discussed in our review of ISO/IEC 27001 in iGaming.

One drafting point affects test planning. The firewall clause in chapter 6 refers to “section 5.5, Significant Events,” while the significant events provisions in the version reviewed appear at section 5.6. Suppliers and laboratories should agree the intended cross-reference in the test plan so that a significant event generated by a full firewall audit log is tested against the right event list.

What Compliance Teams Should Change

The following steps apply to licensees and suppliers preparing a GLI-31 engagement or relying on an existing certificate. They are analysis and recommended practice, not requirements of the standard.

Start by obtaining the regulator’s adoption instrument. Confirm whether the regulatory body has adopted GLI-31 v1.1 in full, adopted it with modifications, or only cited it as guidance, and whether it requires the section 1.4.1 on-site phase. Record the answer in the certification file.

Build the void reconciliation and event manager acknowledgement into the draw runbook as a named, logged step. Section 2.6.3 makes it a system gate, so the operating procedure must identify who holds the event manager role and who covers absence. Align the charity’s proceeds reconciliation to the section 2.8.1 prize distribution line so that declared proceeds, prizes and refunds trace to one report.

Document the RNG integration boundary. Section 4.4.2 protects the selection process from associated equipment, so any livestream overlay, results page or reporting export must be listed in the architecture and shown to be one-way or otherwise isolated from the draw engine.

Map the online gaps in writing. For each control the regulator expects of an online licensee, such as identity and age checks, payment controls, limits and self-exclusion, state where it is delivered and which standard or licence condition governs it. Where the raffle platform shares infrastructure with other products, plan separate certification scope for the shared components.

Finally, set a change control rule. The standard requires software components to be verifiable by Program ID and version with an independent integrity check (section 5.8.1) and the RSU to support the same check (section 3.6.1). Any release that touches the RNG, scaling, ticket validation or reporting should be treated as a candidate for laboratory re-evaluation, with the regulator’s change policy deciding the outcome.

This analysis rests on a direct read of GLI-31 v1.1 chapters 1 to 6, with GLI-14 and GLI-15 consulted for the product comparison. Operators should consult qualified legal counsel on how their regulator applies the standard.

Frequently asked questions

What is GLI-31?

GLI-31 is Gaming Laboratories International’s technical standard for Electronic Raffle Systems. Version 1.1 was released on 24 July 2015 and covers raffle management, Raffle Sales Units, random number generation, servers and communications.

How is an electronic raffle system certified under GLI-31?

Section 1.4.1 provides for two phases. Laboratory testing comes first, with the system and Raffle Sales Units assembled. On-site testing after installation follows where the regulatory body requires it.

Does GLI-31 require a random number generator?

Section 4.1.1 requires the winning number selection process to be random and allows an approved RNG to achieve this. Where a manual draw is used instead, section 2.4.1 sets rules for identical counterfoils, and a printed counterfoil is not required when an electronic RNG selects the winner.

How does GLI-31 differ from GLI-14 and GLI-15?

GLI-31 covers a pooled raffle draw from sold numbers. GLI-15 covers electronic bingo and keno, and GLI-14 covers finite scratch ticket and pull-tab systems where a Central System selects a predetermined ticket from a game set.

Key Resources

Gaming Laboratories International, GLI Standard 31: Electronic Raffle Systems, Version 1.1 (24 July 2015). Gaming Laboratories International, GLI Standard 15: Electronic Bingo and Keno Systems, Version 1.3 (6 September 2011). Gaming Laboratories International, GLI Standard 14: Finite Scratch Ticket and Pull-Tab Systems, Version 2.2 (6 September 2011). The adopting regulator’s own raffle, charitable gaming or lottery rules should be read alongside each standard.