Tie-breaker designs are more efficient than regression discontinuity designs for estimating treatment effects. They combine aspects of randomized controlled trials and regression discontinuity designs. Tie-breaker designs are better when it's unfair to deny treatment to deserving recipients. They also provide statistical benefits over regression discontinuity designs. Using tie-breaker designs requires fewer subjects to achieve the same accuracy as regression discontinuity designs, especially when estimating treatment effects at multiple score values. Larger experimental radii choices in tie-breaker designs lead to even greater efficiency.