Despite intensive research in the field of mote-class Wireless Sensor Networks in recent years, real-life deployments are still challenging and systems are prone to failures. This can typically be attributed to fragile hardware or misbehaving software. Issues caused by software, often induced by the inherent constraints of resources, can be countered using simulations. However the simulation results often do not reflect those of the specific deployment.We suggest analyzing the actual environment conditions of a deployed network and map them to a simulator. Then, based on simulations, software and parameters can be tailored to the specific deployment.We developed two tool chains, RealSim and DryRun, and compared results from simulation runs to those acquired from two different testbeds using Tmote Sky nodes. This was done in two campaigns, each altering 2 configuration parameters from the hardware to the application layer. The presented data is based on over 1100 experiments, respectively over 270 h, on real hardware and almost 7000 simulations. The close relation of simulation and real measurements shows that our DrySim approach is feasible.