Positron emission tomography (PET) detectors with the ability to encode depth-of-interaction (DOI) information allow us to simultaneously improve the spatial resolution and sensitivity of PET scanners. In this study, we propose a DOI PET detector based on a stair-pattern reflector arrangement inserted between pixelated crystals and a single-ended scintillation light readout. The main advantage of the proposed method is its simplicity; DOI information is decoded from a flood map and the data can be simply acquired by using a single-ended readout system. Another potential advantage is that the two-step DOI detectors can provide the largest peak position distance in a flood map because two-dimensional peak positions can be evenly distributed. We conducted a Monte Carlo simulation and obtained flood maps. Then, we conducted experimental studies using two-step DOI arrays of 5 × 5 Lu1.9Y0.1SiO5:Ce crystals with a cross-section of 1.7 × 1.7 mm2 and different detector configurations: an unpolished single-layer (US) array, a polished single-layer (PS) array and a polished stacked two-layer (PT) array. For each detector configuration, both air gaps and room-temperature vulcanization (RTV) silicone gaps were tested. Detectors US and PT showed good peak separation in each scintillator with an average peak-to-valley ratio (PVR) and distance-to-width ratio (DWR) of 2.09 and 1.53, respectively. Detector PSRTV showed lower PVR and DWR (1.65 and 1.34, respectively). The configuration of detector PTAir is preferable for the construction of time-of-flight-DOI detectors because timing resolution was degraded by only about 40 ps compared with that of a non-DOI detector. The performance of detectors USAir and PSRTV was lower than that of a non-DOI detector, and thus these designs are favorable when the manufacturing cost is more important than timing performance. The results demonstrate that the proposed DOI-encoding method is a promising candidate for PET scanners that require high resolution and sensitivity and operate with conventional acquisition systems.