Read the explanation
This saved simulator models a thirty-two-hour threshold with a one-point-five overtime multiplier. For forty scheduled hours at twenty-five dollars per hour, regular hours are the minimum of forty and thirty-two, or thirty-two. Overtime hours are the maximum of zero and forty minus thirty-two, or eight. Regular pay is eight hundred dollars and modeled overtime is three hundred, totaling eleven hundred weekly. Bars show regular and overtime pay at point six pixels per dollar. This describes the program assumptions only and does not establish any current legal overtime rule, worker eligibility, contract provision or enacted policy. The comparison branch uses a forty-hour threshold and the same hourly wage and multiplier. At a forty-hour schedule all forty hours are regular in that branch, totaling one thousand dollars. The modeled thirty-two-hour branch gives eleven hundred, a one-hundred-dollar weekly difference. Bars use point four pixels per dollar. At fifty scheduled hours the forty-hour branch pays one thousand plus ten times thirty-seven point five, or thirteen hundred seventy-five. The thirty-two-hour branch pays eight hundred plus eighteen times thirty-seven point five, or fourteen hundred seventy-five. Its difference remains one hundred. This is a controlled arithmetic comparison, not a prediction about hours, employment or productivity. At the default forty hours, eleven hundred weekly times fifty-two gives fifty-seven thousand two hundred annual pay per person, versus fifty-two thousand in the comparison branch. The difference is five thousand two hundred. Ten workers make the company annual difference fifty-two thousand. The upper bar uses point zero zero zero five pixels per one-person annual pay and the lower uses point zero one pixels per company difference; their scales are explicitly different. The program assumes identical employees, constant schedules and fifty-two paid weeks. It omits payroll taxes, benefits, leave, eligibility, scheduling changes and output. The native JSON export preserves entered parameters and computed results, not proof of a policy outcome.