In darkness it is legally obligatory to light vehicles, both to enable their operators to see where they are travelling and to make the vehicle visible and recognisable to other road users. Legal requirements differ between vehicle types and, for bicycle lighting, also between countries. So far, in spite of some indications, there is no definite evidence of a relationship between crashes involving cyclists and darkness, nor on the presence of bicycle lighting and crash risk in darkness, possibly due to a lack of prevalence data. There are no standards for scientific or consumer testing of bicycle lights, and visibility-related results are inconclusive. Three test methods were employed to evaluate generic bicycle light features such as beam shape, brightness, steady versus flashing beam, and mounting position for their effects on seeing and being seen, while investigating which method was best suited for which type of evaluation. In static indoor laboratory testing, brightness was measured and beam shape documented photographically. In static outdoor testing, visibility from 300 m was evaluated, subjective ratings were collected, and beam shape documented. Three front-light beam types were then evaluated for their effect on gap acceptance in a dynamic setting in an urban environment, and subjective ratings were collected. All 18 tested lamps fulfilled the legal requirements, with a bright, steady front beam receiving the highest subjective ratings. Gap acceptance was influenced only by cyclist speed, but neither by beam type nor mounting position. To capture the effect of generic bicycle light features on perception, behaviour, and acceptance, one single method is not enough, and standardised laboratory tests should be combined with research-question–specific dynamic testing. © 2021 The Authors