Motivation The transition towards an all-renewable electricity sector increases the need for flexibility [1]. Battery storage systems are one option to serve this need. According to [2], 25 GW of storage grid connection requests were approved in Germany in 2024. However, high volumes of storage also impact electricity prices and may cannibalise each other's revenues. This is often referred to as "avalanche effect" [3]. The question arises as to how the competition between storage units can be modelled such that their price repercussions are adequately reflected in their bidding strategies. Methods We present a new method [4] for simulating competing storage systems that solves the dispatch problem of competing storage agents. The method is implemented in the agent-based electricity market simulation model AMIRIS [5] and is based on dynamic programming. Here, storage content levels are discretised. An optimal path of storage levels is determined to either maximise storage profits or to minimise the total system cost. The value of state transitions along the path is based on electricity price forecasts provided by a dedicated forecast agent. Forecasts include estimates of a) the market clearing price without storage operation, b) the change of market price correlated with potential storage dispatch, and c) the correlation of each individual storage unit’s dispatch with respect to the total storage dispatch. The latter is estimated based on previous simulation results. This storage dispatch correlation is factored into an individual multiplier that each storage unit can employ to estimate the total price impact of storage dispatch. Results We test our method in a backtesting scenario that depicts the German electricity market in 2019. At first, storage capacities are aggregated into a single unit. Results from profit-maximising and system cost-minimising strategies are compared to the historical storage dispatch and electricity prices. We find the highest correlation of simulated prices with historical prices for a profit-maximising strategy employing market power. However, the highest correlation with historical dispatch is found with the cost-minimising strategy. In a second backtest, we disaggregate the storage capacities to 18 units and investigate their competition, again employing different dispatch strategies. Correlations with the historical dispatch improve significantly for both strategies. The system cost-minimising strategy very closely resembles the historical dispatch. This hints towards a competitive market environment. Employing the profit-maximising strategy, less than historical use of storage capacities can be observed resulting in higher storage profits in the simulation. This hints towards (currently unrealised) potential monetary gains for a collusive dispatch strategy. Last, we consider a situation with 20 GW of additional storage capacity in the system to assess possible revenue cannibalisation. We find a very clear cannibalisation of revenues. Further research should consider additional flexibility options and their impact onto revenue cannibalisation. Proper transformation scenarios should be employed for this analysis. An analytical comparison with game theory approaches also offers grant additional insights. Data [6] and model [7] are publicly available. References [1] https://doi.org/10.1016/j.rser.2021.111995 [2] https://www.bundesnetzagentur.de/1079644 [3] https://doi.org/10.1016/j.esr.2020.100608 [4] https://doi.org/10.5281/zenodo.17743531 [5] https://doi.org/10.21105/joss.05041 [6] https://doi.org/10.5281/zenodo.16978508 [7] https://doi.org/10.5281/zenodo.17105909