Background: Electronic gaming machines (EGMs) are recognised as one of the most harmful forms of gambling at an individual level. However, it is unclear whether restriction of EGM functions and accessibility results in meaningful reductions in population-level gambling harm. Methods: This study employed a natural policy experiment to compare gambling problems between Australian residents in Western Australia (WA), where EGMs are restricted to a single venue and have different structural features, and residents in other Australian jurisdictions where EGMs are widely accessible in casinos, hotels and clubs. The accessibility of non-EGM forms of gambling is generally similar across all Australian jurisdictions. The study draws on a large (N = 15,000) nationally-representative dataset.Findings: Although gambling participation was slightly higher in WA, EGM participation was approximately half that found in the rest of Australia. Aggregate gambling problems and harm were about one-third lower in WA, and self-reported attribution of harm to EGMs by gamblers and affected others was 2·7x and 4x lower, respectively. Analyses using mediation and moderation determined that the attributable impact of EGMs in WA was 70% lower than elsewhere, and was responsible for the vast majority of the discrepancy in gambling problems (indirect path = -0·055, 95% CI -0·071; -0·038). Attributable problems to casino table games were slightly higher in WA, potentially due to redirection of demand, but this effect size was small relative to the reduction of problems due to EGMs. On an individual level, for a given level of participation in EGMs and other forms, the risk of problems were similar in WA and the rest of Australia. Interpretation: Reduced harm from gambling in WA is due to restricted accessibility of EGMs, and there is little transfer of problems to other gambling forms. These results suggest that restricting the accessibility of EGMs substantially reduces gambling harm.Funding: Gambling Research Australia (GRA) funded this research. GRA is a partnership between the Commonwealth, State and Territory Governments to initiate and manage a national program of gambling research.