Event classifiers based on charged-particle multiplicity have been extensively used in pp collisions at the LHC. However, a drawback of the multiplicity-based event classifiers is that selected samples at high charged-particle multiplicity are biased towards hard processes. These biases blur the effects of multi-parton interactions (MPI) and make it difficult to pinpoint the origins of fluid-like effects in small systems.

This proceedings contribution exploits a new event classifier, the charged-particle flattenicity, defined in ALICE using the charged-particle multiplicity estimated in the intervals $2.8 <\eta< 5.1$ and $−3.7 <\eta< −1.7$. Final results on the production of identified and unidentified charged particles as a function of flattenicity in pp collisions at $\sqrt{s}=13~\mathrm{TeV}$ are discussed. It is shown how flattenicity can be used to select events in a way that is more sensitive to MPI. All the results are compared with predictions from QCD-inspired Monte Carlo event generators.