More seeds simply create more phenotypes of the same genotype. Plants produce seeds with slightly different variations. Even when all the seeds are from the exact same genetic stalk they frequently display different characteristics. This develops a variety of differently prioritized seeds capable of handling a greater fluctuation in environment. As an example, a phenotype designed to endure drought, will thrive when conditions are dry, and the oposite occurs as well.
Professional breeders normally maintain strict environmental control which influences the plant to create seeds with less genetic sway (phenotype difference). Still however, cannabis will almost always encourage phenotype diversity. This is why when we purchase 10 seeds of the same stalk, we almost always have slightly different versions of a specific strain and choose the one that best fits our grow methodology.
More seeds of the same strain can be beneficial but how we develop the seeds is important. Normally, less environmental differences equates to more standardized seed production.
When I produce seed, I first develop some goals. Things like more or less potency, flavor or trichrome production. I'll choose the stabile parent strains and breed them. I always shoot for a few hundred seeds (only a few buds). Then I'll start 20 or 30 of the healthiest and find the phenotypes that best suit my goals. This takes time, patience and consistant environmental attentiveness. Often, multiple phenos show promise and it can be difficult choosing the one that best suits your needs but culling them out, and prioritizing growth traits with your original goal in mind is important.
Creating stabile and desirable strains takes real commitment. Accidental breeding indeed brings the occasional gem but most properly developed stains have undergone an intense breeding process. I'm in that process now, working on a new cross of Reclining Budda/Chem Dawg/Jack Cleaner. I'm down the three females after a 35 seed run and I've been at it for over a year and counting. I expect to find the pheno once these three flower out.
Lastly, if you start 1000 seeds of the same genetic stalk, you will likely find the phenotype you desire. If you start 10, then yes, your chances of finding that same pheno diminish. More seeds equates to more phenotypes examples. Less stabile environments create conditions where more seeds will be needed to find the right one. Hope this helps you a bit.