Cosmic ray detectors like the $3000\,\text{km}^2$ surface array of the Pierre Auger Observatory are capable of observing high-energy photons in the range of $10^{18}$ to $10^{20}\,\text{eV}$ if the flux is sufficiently high.
However, no clear candidates for ultra-high-energy photons have been identified yet, so simulations must be used to study typical trigger patterns and observables for discriminating photons from hadrons, e.g., with neural networks.
Thinning algorithms are applied to keep the computation time and file sizes in a manageable range since the simulation of ultra-high-energy particle showers is computationally expensive.
In \textsc{Corsika}, particles with energies below a certain fraction of the primary energy, the thinning level, are exposed to thinning.
In the case of thinning, only one of the particles emerging from an interaction is tracked.
By assigning a corresponding weight, this particle then represents a number of its siblings.
However, the weights of particles that originate from electromagnetic interactions can be $100$ times larger than for hadronic interactions.
In contrast to hadronic showers, where a major part of the signal in a surface detector is produced by muons, photon showers are almost purely electromagnetic.
Using simulations of photon-induced showers with two different thinning levels, the influence on different observables used for photon-hadron discrimination is investigated.
Effects deriving from both statistical sampling and detector simulations are considered.
Possible influences on station-level as well as event-level observables are probed.
With this study, we are reassured that the optimal thinning parameters determined for hadron-induced showers are also sufficient for photon-induced showers.