Title: Submillimeter Galaxy Number Counts in a Semi-analytic Model: the "Count Matching" Approach Abstract: Fitting submillimeter galaxies (SMGs) into the current theory of galaxy formation has been a challenge since their discovery, even though they are the most luminous star-forming sources at the epoch where galaxy formation peaks. Recent ALMA observations of the Extended Chandra Deep Field South (for which LABOCA detections are available) show that the bright end of the LABOCA number counts is actually comprised by emission from multiple fainter sources, given the high sensitivity and resolution of ALMA maps compared to LABOCA. With the aim of exploring the properties of SMGs, and in analogy to the now- standard abundance matching approach, we perform a "Count Matching" approach through lightcones drawn from a semi-analytic model. We choose various physical galaxy properties given by the model as proxies for their submillimeter fluxes, assuming a monotonic relationship so that the combined LABOCA plus bright-end ALMA observed number counts are reproduced. After turning the catalogs of galaxy positions and fluxes given by the different proxies into submillimeter maps that include a modeling of the observational process, we perform a source extraction as done for maps obtained through observations. With this we study the effects of the observational process in the recovered counts, as well as the galaxy properties derived from the detected sources for each proxy. The difference between the counts with and without the beam of LABOCA can be used to find the best proxy: sources with the highest submm fluxes will have different clustering depending on the assumed proxy, since the sources with the highest value of a given property will be clustered in a particular way; the best proxy then will be such that the clustering measured from the model galaxies is the same that the one given by the ALMA sources, while the recovered clustering from the simulated maps is the same that the one given by the LABOCA SMGs. This can also be confirmed exploring other statistics like redshift, morphology, stellar mass and host halo mass distributions.